Editor in ChiefLinda Tapsell, PhD, MHPEd, DipNutrDiet, BSc, AdvAPD, FDAA, AM University of Wollongong, Australia

EditorJudi Porter, PhD, MHSc, MCFSM, GCHPE, GDipNutDiet, BAppSc, AdvAPD, AN Eastern Health/Monash University, Australia

Statistics EditorMarijka Batterham, PhD, MMedStat, MSc, BSc, GStat, AdvAPD, AN University of Wollongong, Australia

Systematic Literature Review EditorElizabeth Neale, PhD, BND (Hons), APD University of Wollongong, Australia

Clinical Trials EditorSharleen O’Reilly, PhD, GCHPE, BSc(Hons), APD University College Dublin, Ireland

Qualitative Research EditorClaire Palermo, PhD, MPH, MNutDiet, BSc, APD Monash University, Australia

Supplement EditorLynda Ross, PhD, BND (Hons Class 1), AdvAPD Griffith University, Australia

Editorial Board MembersEkta Agarwal, PhD, APD, RD Bond University, AustraliaLucinda Bell, PhD, APD, AN Flinders University, AustraliaAndrea Braakhuis, PhD, RD The University of Auckland, New ZealandKatrina Campbell, PhD, AdvAPD Bond University, AustraliaWei Chen, PhD, Zhejiang University, ChinaClare Corish, PhD, FINDI, RDUniversity College Dublin, IrelandKacie Dickinson, PhD, BNutrDiet (Hons), APD Flinders University, AustraliaJane Elmslie, PhD, GDipSci, DipHSc, NZRD Christchurch School of Medicine, New ZealandSuzie Ferrie, PhD, CNSC, AdvAPD Royal Prince Alfred Hospital, AustraliaVicki Flood, PhD, MPH, APD University of Sydney, AustraliaJanelle Gifford, PhD, MSc, BSc, BBus, Adv APD, Adv Sports Dietitian Core Nutrition, AustraliaRebecca Golley, PhD, BND, BSc(Hons), APD, AN University of South Australia, AustraliaKathryn Hart, PhD, BSc(Hons), DipADP, RD University of Surrey, United KingdomIngrid Hickman, PhD, BHSc, AdvAPD, AN Princess Alexandria Hospital, AustraliaVasant Hirani, PhD, DDiet, MSc, BSc(Hons), APD University of Sydney, Australia

Tilakavati Karupaiah, PhD, APD, AN Taylor’s University, MalaysiaNicole Kiss, PhD, MNutDiet, BSc, AdvAPD Deakin University, AustraliaJimmy Louie, PhD, MNutrDiet, BSc(Hons), APD, AN University of Hong Kong, Hong KongEvangeline Mantzioris, PhD, BND, BSc, Grad Cert High Educ, APD, AN, SDA University of South Australia, AustraliaAndrew McAinch, PhD, MNutrDiet, BApplSc, BSc(Hons), APD, AN Victoria University, AustraliaKirrilly Pursey, PhD, APD University of Newcastle, AustraliaAnna Rangan, PhD, GDipNutrDiet, BSc, APD University of Sydney, AustraliaDianne Reidlinger, PhD, RD, APD Bond University, AustraliaNerissa Soh, PhD, BMedSc, MNutrDiet, APD, AN University of Sydney, AustraliaSze-Yen Tan, PhD, MSc, APD, AN Deakin University, AustraliaHelen Truby, PhD, M.Hum Nutr, AdvAPD Monash University, AustraliaRobin M. Tucker, PhD, RD Michigan State University, United States of AmericaCarol Wham, PhD, MSc, BHSc, DipEd, NZRD Massey University, New ZealandSerene Yoong, PhD, BNutrDiet(Hons), APD, AN Hunter New England Local Health District, AustraliaJo Zhou, PhD, MNutDiet, BSc, DipMed, APD Women’s and Children’s Hospital Adelaide, Australia

Editorial AssistantClaire Wilkinson, BSc (Hons)

Journal ManagerPaul Wilkinson PhD, Bsc (Hons), APD

Journal Strategic Planning CommitteeRobyn Littlewood (DAA Director Responsible)Marijka BatterhamJudy BauerElizabeth NealeSharleen O’ReillyJudi PorterClaire PalermoGabrielle RyanLinda Tapsell (Chairperson)Claire WilkinsonPaul Wilkinson

Cover image courtesy of iStock (Credit: KatarzynaBialasiewicz)

Address for Editorial Correspondence:Editor, Nutrition & Dietetics1/8 Phipps CloseDeakin ACT 2600AustraliaEmail: journal@daa.asn.au

NDI.JEB.Jul17

Disclaimer: The Publisher, the Dietitians Association of Australia and Editors cannot be held responsible for errors or any consequences arising from the use of information contained in this journal; the views and opinions expressed do not necessarily reflect those of the Publisher, the Dietitians Association of Australia and Editors, neither does the publication of advertisements constitute any endorsement by the Publisher, the Dietitians Association of Australia and Editors of the products advertised.

For submission instructions, subscription and all other information visit http://wileyonlinelibrary.com/journal/ndi

COPYRIGHT AND COPYING

Copyright © 2018 Dietitians Association of Australia. All rights reserved. No part of this publication may be reproduced, stored or transmitted in any form

or by any means without the prior permission in writing from the copyright holder. Authorization to copy items for internal and personal use is granted by the copyright holder for libraries and other users registered with their local Reproduction Rights Organisation (RRO), e.g. Copyright Clearance Center (CCC), 222 Rosewood Drive, Danvers, MA 01923, USA (www.copyright.com), provided the appropriate fee is paid directly to the RRO. This consent does not extend to other kinds of copying such as copying for general distribution, for advertising or promotional purposes, for republication, for creating new collective works or for resale. Permissions for such reuse can be obtained using the RightsLink “Request Permissions” link on Wiley Online Library. Special requests should be addressed to: permissions@wiley.com

Volume 75 Number 5 November 2018 ISSN 1446-6368

EditorialHarnessing the power, know-how and capacity of the most vulnerable in designing and implementing policy 445

Julie Brimblecombe

Qualitative ResearchExploring the impact of Aboriginal health placement experiences on the preparation of dietetic graduates for practice with Aboriginal communities 448

Ruby Svarc, Corinne Davis, Helena McDonald, Julia Perruzza, Jennifer Browne, Robyn Delbridge, Keith Morgan, Sharleen O’Reilly, Claire Margerison and Claire Palermo

Nutritional EpidemiologyAnaemia in pregnancy among Aboriginal and Torres Strait Islander women of Far North Queensland: A retrospective cohort study 457

Dympna Leonard, Petra Buttner, Fintan Thompson, Maria Makrides and Robyn McDermott

Food security in a sample of clients attending HIV clinics in New South Wales: A cross-sectional study 468Angela Langton, Rosalind Moxham, Sarah Hirst, Louise Houtzager, Julie Coutelas, Amanda Rider and Christine Yusuf

Cross-sectional analysis of unhealthy foods, race/ethnicity, sex and cardiometabolic risk factors in U.S. adults 474Joan A. Vaccaro, Gustavo G. Zarini and Fatma G. Huffman

Association between adherence to the Dietary Approaches to Stop Hypertension diet with food security and weight status in adult women 481

Saeideh Tabibian, Elnaz Daneshzad, Nick Bellissimo, Neil R. Brett, Ahmad R. Dorosty-Motlagh and Leila Azadbakht

Associations between dietary behaviours and perceived physical and mental health status among Korean adolescents 488Subin Park, Soo J. Rim and Junghyun H. Lee

Health Services ResearchCharacteristics of older adults with diabetes: What does the current aged care resident look like? 494

Olivia Farrer, Alison Yaxley, Karen Walton and Michelle Miller

‘Breaking Barriers, Breaking Bread’: Pilot study to evaluate acceptability of a school breakfast program utilising donated food 500

Natika Deavin, Anne-Therese McMahon, Karen Walton and Karen Charlton

Communicating health—Optimising young adults’ engagement with health messages using social media: Study protocol 509

Catherine Lombard, Linda Brennan, Michael Reid, Karen M. Klassen, Claire Palermo, Troy Walker, Megan S.C. Lim, Moira Dean, Tracy A. McCaffrey and Helen Truby

Nutrition and Dietetics MethodologyUse of hand grip strength in nutrition risk screening of older patients admitted to general surgical wards 520

Angela Byrnes, Alison Mudge, Adrienne Young, Merrilyn Banks and Judy Bauer

Buchanania obovata: An Australian Indigenous food for diet diversification 527Selina A. Fyfe, Michael E. Netzel, Ujang Tinggi, Eva M. Biehl and Yasmina Sultanbawa

Health Services EvaluationFirst Nations students’ perceptions of school nutrition policy implementation: A mixed methods study 533

Christina Gillies, Alexander Research Committee, Anna Farmer, Katerina Maximova and Noreen D. Willows

Letter to the EditorWhat disease did Goethe witness during his journey through the Italian Alps? Was it pellagra or another disease of malnutrition? 541

Marta Licata and Silvia Iorio

Erratum 542

EDITORIAL

Harnessing the power, know-how and capacity ofthose considered most vulnerable in designing andimplementing policy

The differences in health that exist between different popu-lation groups are avoidable and unfair.1 They are largelydetermined by differences in access to economic resources,distribution of power, social policy and the values held bysociety.1 They impact many population subgroups includ-ing Indigenous people, young adults, the elderly and thoseliving with medical conditions.

Significant improvements in health could and must bemade by addressing the social determinants that are mostlyresponsible for these avoidable and unfair health inequities.These social determinants are the ‘conditions in which peo-ple are born, grow, live, work and age’ (p. 189).1 Green andKreuter2 and McLeroy et al.,3 introduced a socio-ecologicaltheory that helped shift the focus of health promotion in thelate 1980s/early 1990s on deficiencies in individual behav-iour, to the environmental variables that were seen as largelyshaping a person’s behaviour. This broad holistic perspectivedoes not deny individual agency. Indeed, an important con-struct of socio-ecological theory is the notion of reciprocaldeterminism where we as individuals can in turn influencethe broader systems of which we are a part.4

The power of people cannot be underestimated. Harnes-sing this power and supporting people’s capacity to advo-cate and be agents of change, can achieve environmentalchange that may otherwise not be possible. An example ofthis is Amata community in the Anangu PitjantjatjaraLands, South Australia, who in 2008 removed top-sellingsugary drinks from sale in their store when provided withevidence by a public health dietitian on the high sales ofsugary drinks and amount of sugar from drinks purchased.5

The entry of ‘systems thinking’ into dietetics practice in thelast decade, while still early in its application, stimulatespractitioners to not only identify the environmental influ-ences on individual behaviour but also consider the rela-tionship between these, the reinforcing and balancingfeedback loops, the key players, points of leverage and theconsequences of change on the system.6,7 Systems thinkingalso reminds us that the conditions that influence eatingbehaviour and diets are dynamic and therefore requiremonitoring and ongoing evaluation. The toolkit used bydietitians to tackle the environmental conditions that shapeand reinforce eating behaviour has had to extend farbeyond nutrition education to advocacy, facilitation, inter-sectoral action, capacity building and community

engagement, making the practice of dietetics both challeng-ing and rewarding.

This issue of Nutrition & Dietetics explores the theme ofnutrition in vulnerable populations including Indigenouspopulations in Australia and Canada, minority groups inUSA, youth in Korea and Australia, elderly populations inaged care facilities or undergoing surgery, school-aged chil-dren in Australia and Canada and people living with HIV.New knowledge and insights provide a forum to achievegreater understanding of the issues in addressing healthinequities for populations considered vulnerable. Authorslook beyond the individual to the conditions in which theirstudy participants were born, grow, live, work and age.

Anaemia in pregnancy is common worldwide with par-ticularly high prevalence in vulnerable populations.8 It canlead to detrimental outcomes for both the mother andchild.9 In Australia, Aboriginal and Torres Strait Islanderpeople are almost twice as likely as non-Indigenous peopleto be at risk of anaemia, and 1 in 10 Aboriginal and TorresStrait Islander women are at-risk.10 Leonard et al.,11 in thisissue of Nutrition & Dietetics, examine anaemia in preg-nancy among Aboriginal and Torres Strait Islander womenof Far North Queensland (QLD). Through combining avail-able datasets, authors reveal an alarming rate of over halfthe mothers of Far North Queensland having anaemia in preg-nancy; a rate similar to that reported for two remote NorthernTerritory communities. Six of every 100 Australian birthsare to an Aboriginal or Torres Strait Islander parent.12

Authors look beyond individual behaviour to recommend‘policy commitment to improving food security, promotinggood nutrition and providing targeted iron supplementationand/or fortification for pregnant women’.

To provide optimal health care and close the health gapbetween Aboriginal and Torres Strait Islander peoples andnon-Indigenous Australians, a culturally competent healthworkforce is essential. Building a culturally competent work-force starts at university level. Svarc et al.13 consider thepreparation of dietetics graduates for practice with Aboriginalcommunities. The authors challenge dietetic curriculum leadsto incorporate elements of placement experiences withAboriginal and Torres Strait Islander populations into curric-ulum content where placement experiences may not be feasi-ble, an exciting area of research and practice.

The dietary importance of the extremely nutritious nativefoods, most familiar to many Aboriginal and Torres StraitIslander people, has largely been overlooked in health pre-vention strategies and policy.14 Indigenous Australians haveAccepted September 2018

© 2018 Dietitians Association of Australia 445

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 445–447 DOI: 10.1111/1747-0080.12496

an intricate relationship with and detailed knowledge ofthese foods. Fyfe et al.15 in this issue of Nutrition & Dietetics,reveal the superb nutritional properties of the ‘Green Plum’,the fruit of a native plant that grows prolifically in thesub-tropical woodlands of northern Australia and is con-sumed by Indigenous peoples of the Northern Territoryand Western Australia. The plant has medicinal properties,reminding me of a time when I experienced tooth-achewhen staying in a north east Arnhem community, NorthernTerritory, and senior women shared their knowledge of theanalgesic properties of the ‘munydjutj’ (Green Plum) plant.Opportunity for Indigenous Australians to establish wildharvesting and processing enterprises of such native foodsin a way that is culturally appropriate and in the control ofAboriginal and Torres Strait Islander peoples, could con-tribute importantly to improved livelihoods, food security,nutrition and health.

Food security exists when ‘all people, at all times, havephysical, social and economic access to sufficient, safe andnutritious food, which meets their dietary needs and foodpreferences for an active and healthy life’.16 Australia is a foodsecure country, however, food is not equitably distributed.Twenty-two per cent of Aboriginal and Torres Strait Islanderpeoples, 16% of Australians living in rental accommodationand 11% of Australians with no employment were living in ahousehold that in the previous 12 months had run out offood and had not been able to afford to buy more.17 Many ofthe people living with HIV in Australia are members of thesepopulation groups but there is little known about the preva-lence of food insecurity among people living with HIV.18 Thearticle by Langton et al.18 provides an up to date report offood insecurity among this population group.

Persistent racial/ethnic differences in health are a majorpublic health problem.19 The physical, social and socio-economic characteristics of our neighbourhoods can impactour health and may partly explain health disparities.20 Vac-caro et al.21 show differences in unhealthy food consumptionby race/ethnicity among Mexican American, other Hispanic,non-Hispanic black and non-Hispanic white adults usingdata from the US National Health and Nutrition Survey. Theneighbourhood food environment likely contributes to thesedifferences. In USA, there is evidence that lower incomecommunities and those with higher percentages of ethnicminorities have both greater access to outlets that sellunhealthy food and lower access to those selling healthyfood.22,23 More research is needed to guide effective policyand policy implementation and evaluation in this complexarea of environments, eating behaviour, health and equity,where there are powerful interests at play.

School breakfast programs offer a public health approachto addressing food insecurity for children that minimisesstigma when made available for all. Deavin et al.24 explorethe acceptability and perceived benefits of the ‘BreakingBread, Breaking Barriers’ program in a public primaryschool in the Illawarra region of New South Wales,Australia, that provided a free breakfast to school-aged chil-dren using donated food. Operating over two terms, theprogram saved an impressive 14.4 t of food from landfill

through conversion into 44 000 meals. It will be useful toconsider how this model can be applied more broadly andwith the provision of only healthy donor food.

There has been less research among adolescent andyoung adult populations than that of adults and children.25

Adolescents and young adults are establishing behavioursthat can track in to adult life and are at a stage when theyare very receptive to their environment and can readilyadopt high risk behaviours. Park et al.25 investigate theassociation between dietary habits and perceived mentaland physical health among a representative sample ofKorean youths with data from an online self-report surveyconducted by the Korea Centres for Disease Control andPrevention. The study protocol of Lombard et al.26 detailsan innovative 4-phase study that seeks to understand howyoung adults use social media in relation to their healthissues. This cross-disciplinary team is combining the toolsand techniques of social marketing with social media plat-forms to reach young adults including AboriginalAustralians, to motivate them to engage in healthy eatingbehaviours to reduce the risk of obesity.

Our elderly are our national treasures and yet their needscan be too easily overlooked. By 2050, up to 9% of theAustralian population will be aged 85 years or more.27 Thisissue highlights the importance of providing adequate,enjoyable and culturally appropriate nutrition for older peo-ple in aged care facilities. Farrer et al.27 forewarn of howgenerational differences in the health-care needs of our age-ing population will present new challenges to our health-care services.

We need public health approaches that help create phys-ical and social environments conducive to healthy eating tohave the greatest potential to improve the health of ournations.28 Over the years, we have come to understand thatcertain populations are at risk of disparate health outcomesbecause of inequities in the distribution of wealth andpower and/or the prejudices and intolerances that existwithin society. In tackling issues upstream we need to harnessthe power, know-how and capacity of those considered themost vulnerable in designing and implementing policy.

Funding source

The author received no funding to write this Editorial.

Conflict of interest

The author has no conflict of interest to report.

Authorship

JB is the sole author of this manuscript.

Associate Professor Julie Brimblecombe, BSc,PostGradDipNutrDiet, MPH, PhD, Department of Nutrition,

Dietetics and Food, Monash University; Honorary Fellow,Menzies School of Health Research

J. Brimblecombe

446 © 2018 Dietitians Association of Australia

References

1 World Health Organisation. Social Determinants Approaches toPublic Health: From Concept to Practice. Geneva: WHO, 2011.

2 Green LW, Kreuter MW. Educational and ecological assess-ment of factors affecting health-related behaviour and environ-ments. In: Green LW, Kreuter MW, eds. Health PromotionPlanning: An Educational and Ecological Approach, 3rd edn.California: Mayfield Publishing Company, 1991; 152–87.

3 McLeroy KR, Bibeau D, Steckler A, Glanz K. An ecological per-spective on health promotion programs. Health Educ Q 1988;15: 351–77.

4 Wallerstein N, Bernstein E. Empowerment education: Freire’sideas adapted to health education. Health Educ Q 1988; 15:379–94.

5 Butler R, Tapsell L, Lyons-Wall P. Trends in purchasing pat-terns of sugar-sweetened water-based beverages in a remoteAboriginal community store following the implementation of acommunity-developed store nutrition policy. Nutr Diet 2011;68: 115–9.

6 Naaldenberg J, Vaandrager L, Koelen M, Wagemakers AM,Saan H, de HK. Elaborating on systems thinking in health pro-motion practice. Glob Health Promot 2009; 16: 39–47.

7 Hawe P, Shiell A, Riley T. Theorising interventions as events insystems. Am J Community Psychol 2009; 43: 267–76.

8 World Health Organisation. The Global Prevalence of Anaemiain 2011. Geneva: WHO, 2015.

9 Stevens GA, Finucane MM, De-Regil LM et al. Global, regional, andnational trends in haemoglobin concentration and prevalence oftotal and severe anaemia in children and pregnant andnon-pregnantwomen for 1995–2011: a systematic analysis of population-representative data. Lancet GlobHealth 2013;1: e16–25.

10 Australian Bureau of Statistics. 427.0.55.003—Australian Aborig-inal and Torres Strait Islander Health Survey, Biomedical Results,2012–2013. Canberra: ABS, 2014.

11 Leonard D, Buttner P, Thompson F, Makrides M, McDermott R.Anaemia in pregnancy among Aboriginal and Torres StraitIslander women of Far North Queensland: a retrospectivecohort study. Nutr Diet 2018; 75: 457–67.

12 Clarke M, Boyle J. Antenatal care for Aboriginal and TorresStrait Islander women. Aust Fam Physician 2014; 43: 20–4.

13 Svarc R, Davis C, McDonald H et al. Exploring the impact ofAboriginal health placement experiences on the preparation ofdietetic graduates for practice with Aboriginal communities.Nutr Diet 2018; 75: 448–56.

14 Brimblecombe J, Maypilama E, Colles S et al. Factors influenc-ing food choice in an Australian Aboriginal community. QualHealth Res 2014; 24: 387–400.

15 Fyfe SA, Netzel ME, Tinggi U, Biehl EM, Sultanbawa Y. Bucha-nania obovata: an Australian Indigenous food for diet diversifi-cation. Nutr Diet 2018; 75: 527–32.

16 Food and Agriculture Organisation of the United Nations(FAO). Voluntary Guidelines to Support the Progressive Realisationof the Right to Adequate Food in the Context of National FoodSecurity. Rome: FAO Council, 2005.

17 Australian Bureau of Statistics. 4727.0.55.005—AustralianAboriginal and Torres Strait Islander Health Survey, NutritionResults—Food and Nutrients, 2012–13. Canberra: ABS, 2015.

18 Langton A, Moxham R, Hirst S et al. Food security in a sampleof clients attending HIV clinics in New South Wales: a cross-sectional study. Nutr Diet 2018; 75: 468–73.

19 Institute of Medicine (US) Committee on Understanding andEliminating Racial and Ethnic Disparities in Health Care. In:Smedley BD, Stith AY, Nelson AR, eds. Unequal Treatment: Con-fronting Racial and Ethnic Disparities in Health Care. Washington,DC: National Academies Press (US), 2003.

20 Mujahid MS, Moore LV, Petito LC, Kershaw KN, Watson K,Diez Roux AV. Neighborhoods and racial/ethnic differences inideal cardiovascular health (the Multi-Ethnic Study of Athero-sclerosis). Health Place 2017; 44: 61–9.

21 Vaccaro JA, Zarini GG, Huffman FG. Cross-sectional analysisof unhealthy foods, race/ethnicity, sex and cardiometabolic riskfactors in U.S. adults. Nutr Diet 2018; 75: 474–80.

22 Lamichhane AP, Warren J, Puett R et al. Spatial patterningof supermarkets and fast food outlets with respect toneighborhood characteristics. Health Place 2013; 23:157–64.

23 Lamichhane AP, Warren JL, Peterson M, Rummo P, Gordon-Larsen P. Spatial-temporal modeling of neighborhood sociode-mographic characteristics and food stores. Am J Epidemiol2015; 181: 137–50.

24 Deavin N, McMahon A-T, Walton K, Charlton K. ‘BreakingBarriers, Breaking Bread’: pilot study to evaluate acceptabilityof a school breakfast program utilising donated food. Nutr Diet2018; 75: 500–8.

25 Park S, Rim SJ, Lee JH. Associations between dietary behav-iours and perceived physical and mental health status amongKorean adolescents. Nutr Diet 2018; 75: 488–93.

26 Lombard C, Brennan L, Reid M et al. Communicating health—optimising young adults’ engagement with health messages usingsocial media: study protocol. Nutr Diet 2018; 75: 509–19.

27 Farrer O, Yaxley A, Walton K, Miller M. Characteristics ofolder adults with diabetes: what does the current aged care res-ident look like? Nutr Diet 2018; 75: 494–9.

28 World Health Organisation. Equity, Social Determinants andPublic Health Programs. Geneva: WHO, 2010.

EDITORIAL

© 2018 Dietitians Association of Australia 447

ORIGINAL RESEARCH

Exploring the impact of Aboriginal health placementexperiences on the preparation of dietetic graduatesfor practice with Aboriginal communities

Ruby SVARC,1* Corinne DAVIS ,1* Helena MCDONALD,1 Julia PERRUZZA,1 Jennifer BROWNE,2

Robyn DELBRIDGE,2 Keith MORGAN,2 Sharleen O’REILLY ,3 Claire MARGERISON 4

and Claire PALERMO 1

1Department of Nutrition and Dietetics, Monash University, 2Public Health and Research Unit, The VictorianAboriginal Community Controlled Health Organisation and 4School of Exercise and Nutrition Sciences, DeakinUniversity, Melbourne, Victoria, Australia; and 3UCD Agriculture and Food Science Centre, University College Dublin,Dublin, Ireland

AbstractAim: A health workforce with the ability to practice with Aboriginal communities is crucial to bridge the health gapbetween Aboriginal and non-Aboriginal Australians. This study aimed to explore the impact of university Aboriginalhealth placements on preparing dietetic graduates for practice with Aboriginal communities.Methods: A mixed methods sequential explanatory design was used. A sample of 594 dietetic graduates was invitedto complete a survey that identified Aboriginal health experiences and measured attitudes and self-confidencetowards working in Aboriginal health using a five-point Likert scale. Participants were divided into placement versusno-placement groups and compared using chi-squared tests. Sixteen of 33 participants who had completed anAboriginal health placement were invited to participate in a semi-structured interview to explore how placementinfluenced practice with Aboriginal communities. Interviews were analysed using content analysis.Results: A final sample of 120 participants showed that placement participants reported significantly higher self-confidence towards working in Aboriginal health compared with no-placement participants (No-placement = 35%agree, 36% neutral, 29% disagree; Placement = 74% agree, 11% neutral, 16% disagree; χ2 (2, 88) = 9.4; P = 0.01). Fif-teen participants were interviewed. Interview data indicated that situated learning experiences, breaking down ste-reotypes, empathy through learning from Aboriginal people, and Aboriginal health role-models were keycomponents of Aboriginal health placements in preparing dietetic graduates for practice with Aboriginalcommunities.Conclusions: The results suggest that Aboriginal health placements may be an effective strategy for preparing die-tetic graduates for practice with Aboriginal communities. The feasibility of placement or alternative curriculum con-tent needs to be explored.

Key words: competency standards, cultural competence, dietitian, education, Indigenous, university placement.

Introduction

Aboriginal and Torres Strait Islander Peoples (hereafterreferred to as Aboriginal people) account for approximately3% of the Australian population but are twice as likely torate their health as fair or poor, and have a life expectancyaround 10 years lower compared with non-Aboriginal peo-ple.1,2 The reasons behind poorer health outcomes arecomplex,2,3 but poor nutrition is responsible for 15% of thegap in health outcomes between Aboriginal and non-Aboriginal people.4–6

As food and nutrition experts, dietitians have an impor-tant role in working with Aboriginal people and communi-ties to promote health and prevent diet-related diseases.7,8

A lack of relevant training and awareness about Aboriginalissues for non-Aboriginal nutrition professionals has been

*Joint first authors.R. Svarc, BNutDiet, Student DietitianC. Davis, BNutDiet, APD, Student DietitianH. McDonald, BNutDiet, APD, Student DietitianJ. Perruzza, BNutDiet, Student DietitianJ. Browne, MPH, APD, Public Health NutritionistR. Delbridge, BNutDiet, APD, Public Health NutritionistK. Morgan, Aboriginal Nutrition Promotion RecruitS. O’Reilly, PhD, AdvAPD, Assistant ProfessorC. Margerison, PhD, APD, Senior LecturerC. Palermo, PhD, APD, Associate ProfessorCorrespondence: C. Davis, Level 1, 264 Ferntree Gully Road, NottingHill, VIC 3168, Australia.Email: crndvs@gmail.com

Accepted January 2018

© 2018 Dietitians Association of Australia448

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 448–456 DOI: 10.1111/1747-0080.12407

identified as a barrier to effective practice.9 A dedicated andculturally competent dietetic workforce is considered essen-tial to support action on nutrition and food security amongAboriginal communities.8–11

Cultural competence, defined as ‘a set of attitudes, skills,behaviours, and policies that enable organisations and staffto work effectively in cross-cultural situations’, has beenidentified as a competency standard for the dietetic work-force.12,13 Research in other health professions includingmedical, nursing, dental and pharmacy, indicates that uni-versity Aboriginal health placements increase culturalknowledge, positive attitudes towards Aboriginal people,and perceived confidence and preparedness in workingwith Aboriginal people.14–21 Little is known about whetherAboriginal health placements positively influence the atti-tudes, skills and behaviours of dietetic graduates and howthis occurs.

Universities are required to incorporate learning strate-gies about cultural competence into course curriculum, butthe approach has differed for strategies specific to workingwith Aboriginal people.13,22 Examples of strategies includediscrete subjects, learning throughout the curriculum, andopportunities for Aboriginal health placements.22 There hasnot been a formal assessment on the effectiveness of theseapproaches to prepare dietetic graduates for practice withAboriginal communities.

This study aimed to explore the impact of Aboriginalhealth placements on dietetic graduates’ attitudes and self-confidence towards working in Aboriginal health, and toexplore how Aboriginal health placements support thepreparation of dietetic graduates for practice with Aborigi-nal communities.

Methods

A mixed methods sequential explanatory design was usedto explore the impact of Aboriginal health placements onthe preparation of dietetic graduates for practice withAboriginal communities. This design consisted of two dis-tinct phases: an initial quantitative phase that testedwhether an Aboriginal health placement influences dieteticgraduates’ attitudes and self-confidence towards working inAboriginal health; followed by a qualitative phase thathelped explain how this change occurred.23,24 The studyteam included 10 members: four student researchers (RS,CD, HMD, JP), two members with extensive experienceworking in Aboriginal health (JB, RD), one member anAboriginal person (KM) and three researchers with experi-ence in dietetic teaching and learning (SOR, CM, CP).Compliance with STROBE and COREQ guidelines has beenaddressed. Monash University and Deakin UniversityHuman Research Ethics Committees (MUHREC CF16/909-2016000473; DUREC 2016-123) provided ethics approval.

Graduates from two Victorian universities (Monash Uni-versity and Deakin University) that offer Dietitians Associa-tion of Australia accredited dietetic programs and haveproduced graduates for more than one year at the time ofthe study were invited to participate. A sample of

594 dietetic graduates who completed their course at Dea-kin University from 2008 to 2016 and Monash Universityfrom 2010 to 2015 (95% of graduates) were invited to par-ticipate in a survey via email. Upon survey completion, par-ticipants were sent a follow-up email encouraging them toforward the link to additional eligible graduates. The surveylink was displayed on the Monash University website. Thesurvey was open for 182 days from April to October 2016.

A 21-question survey was developed, as a reliable andvalidated tool for measuring attitudes and self-confidence ofhealth professionals working with Aboriginal Australiansdoes not exist.25 The survey comprised three sections thatcollected cross-sectional data. The initial section containedan explanatory statement and five demographic questions(age, gender, graduating year and university, and Aboriginaland/or Torres Strait Islander status). The secondsection contained seven questions regarding Aboriginalhealth experiences pre- and post-graduation. Participantswere asked to specify their Aboriginal health experiences atuniversity including the type and length of experience.They were also asked whether any post-graduation workroles had a focus on Aboriginal health. The finalsection included nine questions measuring attitudes andself-confidence towards working in Aboriginal health pre-sented as statements that were rated on a five-point Likertscale (strongly agree to strongly disagree), in line with mea-sures used in other professions and adapted from studieswith similar outcomes of interest.17,21,26,27 Qualtrics soft-ware (Provo, UT, USA) was used to deliver the onlinesurvey.

Following the downloading of survey results, non-identified data were analysed in Microsoft Excel 2010(Redmond, WA, USA). Participants were divided into twogroups; those who had experienced an Aboriginal healthplacement and those who had not. An eligible Aboriginalhealth placement experience was defined as a field trip,clinical placement or public/community health placement,of at least one week duration, that involved interaction withAboriginal people. The eligibility criteria are consistent withthe Aboriginal health placements studied in different healthprofessions.14–20 Interactions within the context of lectures,tutorials, or similar were not classified as ‘placement’. Asubset of participants was created that included participantswho had never worked in a role with a focus on Aboriginalhealth post-graduation. This sub-set was divided into place-ment and no-placement groups (Figure 1).

Descriptive statistics (number and proportions) wereused to describe responses, with chi-squared used to com-pare groups.28 To meet chi-squared test assumptions,28

Likert-scale categories for attitude statements receivingresponses from less than 5% of participants were omittedand analysis was only undertaken on remaining categories(e.g. comparing only ‘strongly agree’, ‘agree’ and ‘neutral’).Likert scale responses were grouped for the self-confidencestatements to meet chi-squared test assumptions (e.g.‘strongly agree’ and ‘agree’ responses were combined toform an overarching ‘agree’ category). A P value <0.05 wasconsidered statistically significant.

Aboriginal health dietetic university placements

© 2018 Dietitians Association of Australia 449

Sixteen of 33 survey participants who had partaken inan Aboriginal health placement experience and agreed to becontacted, were purposively sampled for semi-structuredinterviews lasting approximately 30 minutes, to include arange of Aboriginal health experiences in terms of place-ment length, type and location. Interview questions weredesigned by the research team based on a ‘qualitativedescriptive’ approach to obtain rich information exploringAboriginal health placement experiences in relation to thepreparation for practice with Aboriginal communities(Table 1).24,29,30 Interviews were conducted either by tele-phone or face-to-face in an office meeting room. Otherqualitative research has shown that the depth of data col-lected does not differ significantly between interviewmodes.31 Verbal consent for audiotaping and note-takingwas obtained at the beginning of each interview.

The interviews were not fully transcribed due to resourceconstraints. An alternative six-step approach was used,which has been demonstrated as theoretically sound inhealth services research.32 The approach was undertaken bydietetic student, novice researchers while on an Aboriginalhealth/public health placement (RS, CD, HM, JP; all female)and included: audio-taping of interview and concurrentnote-taking by at least two researchers; verbal reflectionimmediately after the interview to revise notes; listening tothe audiotape at least twice and amending notes and tran-scribing poignant quotes; preliminary coding and contentanalysis, secondary content analysis, and the developmentof themes. A preliminary content analysis by the four stu-dent researchers was initially conducted whereby data werecoded in relation to the research question, and the fre-quency with which each code appeared in the data was

120 participants completed survey

87 participants did not have a

placement

33 participants had a placement

88 participants had not ever worked in a role with a focus on

Aboriginal health post-graduation

69 participants did not have a

placement

19 participants had a placement

Figure 1 Grouping of participants for survey analysis.

Table 1 Semi-structured interview questions

Question Probes

1. Can you tell me about your career so far?2. What influenced the types of jobs you applied for after

you graduated?What stimulated your interest? Placement? Would have

applied for job without placement?3. Could you please tell me about any Aboriginal health

experiences you had while on placement during yourtime studying nutrition and dietetics at university?

Organisation? Projects? Duration? Location? Health setting?Type of caseloads/patient conditions? Activitiesundertaken? Relationship with supervisor? Choice ofwhere to go and reason for going?

4. Are there any additional thoughts, feelings or memoriesthat come to mind when you think about yourAboriginal health placement?

Positive/negative experiences? Problems you came across.Give specific examples/stories - Limitations in certainskills, resources, supervision etc. Coping mechanisms:how did you manage problems?

5. Can you describe any ways that your Aboriginal healthplacement changed your perspectives of Aboriginalhealth and/or Aboriginal people?

How prepared did you feel when you first began workingwith an Aboriginal patient/person or health issue? Whatis the most significant change that you have made toyour work practices that you can attribute to yourAboriginal health placement experience?

6. Can you describe any way in which you think Aboriginalhealth placement experience influenced the way youwork (or would work) with Aboriginal people?

Do you think your placement contributed to this/thesedevelopments?

7. What key attributes do you think you have that assistyou in working with the Aboriginal community?

How did you develop these attributes?

8. If you could design the ideal Aboriginal health placementexperience for dietetics/nutrition students what wouldthis be/look like?

R. Svarc et al.

450 © 2018 Dietitians Association of Australia

recorded. Codes were grouped into similar categories in asecondary content analysis undertaken by RS to elicit com-mon themes. Data from survey and interviews were ana-lysed separately and then key findings drawn together anddiscussed with reference to the literature as is common insequential design.23,24

Results

A total of 571 emails were successfully sent inviting dieteticgraduates to participate. Of these, 121 participants com-pleted the survey. One participant was excluded as they didnot attend either participating university, yielding a finalsample size of 120 (response rate 20%), which is in linewith response rates for internet surveys.33

Most participants were female (96%) and aged25–34 years (81%), reflective of the dietetic workforceprofile,34 and 54% Monash University graduates comparedwith 46% Deakin University graduates (Table 2). Approxi-mately a quarter of participants had an Aboriginal healthplacement experience (28%), with most placements (53%)seven weeks or greater. There were no demographic differ-ences between placement and no-placement groups(Table 2). Of the 120 participants, 27% had worked in awork role which includes a focus on Aboriginal health.

Responses to the attitude statements are provided inTable 3. A higher proportion of participants had more posi-tive attitudes towards Aboriginal people/health in the place-ment group compared with the no-placement group.

However, the difference between the two groups was onlystatistically significant for two attitude statements, whichconcerned the importance of learning about Aboriginal his-tory, culture and health.

The placement group showed a greater proportion ofpositive self-confidence statements towards Aboriginal peo-ple/health (Table 3). Responses to the statements indicatedparticipants who had a placement experience felt signifi-cantly more confident working with Aboriginal people in aculturally safe manner and significantly less apprehensiveabout interactions with Aboriginal people compared withparticipants who had no placement experience.

Placement participants were more likely to have workedin a role, post-graduation, which included an Aboriginalhealth focus compared with no-placement participants(42% vs 21%; χ2 (1, 120) = 5.8; P = 0.02). Differencesbetween groups did not remain significant for the attitudestatements and for one self-confidence statement, when par-ticipants who had worked in a role, post-graduation, whichincluded an Aboriginal health focus, were omitted fromanalysis (data not shown). The only statement to remainsignificant was the self-confidence statement regarding con-fidence in working with Aboriginal people in a culturallysafe manner (Question 8; No-placement = 35% agree, 36%neutral, 29% disagree; Placement = 74% agree, 11% neu-tral, 16% disagree; χ2 (2, 88) = 9.4; P = 0.01).

Sixteen female participants agreed to be interviewed. Fif-teen interviews (12 telephone and 3 face-to-face) were com-pleted (12 Monash University graduates and 3 Deakin

Table 2 Demographic details of participants

Total (n = 120)No Aboriginal healthplacement (n = 87)

Aboriginal healthplacement (n = 33)1

Characteristic n % n % n %P value between

groups2

Identify as Aboriginal person 1 1 1 1 0 0 0.54Do not identify as Aboriginal person 119 99 86 99 33 100Gender

Transgender 1 1 1 1 0 0 0.82Female 115 96 83 95 32 97Male 4 3 3 3 1 3

Age (years)18–24 28 23 23 26 5 15 0.4125–34 81 68 56 64 25 7635–44 11 9 8 9 3 9

UniversityMonash University 65 54 47 54 18 55 0.96Deakin University 55 46 40 46 15 45

Graduation year2000–2004 11 9 9 10 2 6 0.252005–2009 7 6 3 3 4 122010–2014 74 62 53 61 21 642015–2016 28 23 22 25 6 18

1 For the Aboriginal health placement group, 24% had a placement length between 1 week but ≤3 weeks, 9% >3 weeks but ≤5 weeks, 15%>5 weeks but ≤7 weeks and 53% 7 weeks or greater.

2 χ2 analysis assumptions for identify as Aboriginal person, gender and graduation year characteristics were not met. However, it can beassumed that there are no differences between groups given the small numbers of participants in subgroups of the categories.

Aboriginal health dietetic university placements

© 2018 Dietitians Association of Australia 451

University graduates), as one participant withdrew due totime constraints. Twelve of the 15 graduates have workedin a role with an Aboriginal health focus. Two graduateshad a placement experience greater than eight weeks, eightgraduates greater than four weeks but less than or equal toeight weeks, and the remaining five graduates had a place-ment experience less than or equal to four weeks.

Four major themes relating to how Aboriginal healthplacements helped prepare dietetic graduates for practicewith Aboriginal communities were identified:

Theme 1: Situated learning experiences: Situated learningexperiences were mentioned by all graduates. These werediscussed when reflecting on their placement experience orwhen asked about their ideal placement experience. Situ-ated learning experiences on Aboriginal health placementsincluded: visiting Aboriginal communities, talking toAboriginal people, listening to Aboriginal people’s stories

and learning about Aboriginal culture first-hand. Suchexperiences were deemed as an important part of Aboriginalhealth placements by providing a learning experience in apractical way that theory cannot emulate.

‘It actually gave me a perspective on people as opposedto books, I understood the theory and knew the statsbut until I went on placement… I actually hadn’t spokento any Aboriginal people and I was really overwhelmedbecause suddenly…there was a real person experiencinga real life in front of me’—Graduate 11, 2-week remoteplacement.

Situated learning allowed students to connect withAboriginal people by providing opportunities for first-handinteraction and observation. Students could attach compel-ling emotions and memories to the once theoretical people

Table 3 Responses to attitude and self-confidence statements by placement experience group

Stronglyagree Agree Neutral Disagree

Stronglydisagree

Total χ2

n % n % n % n % n % n (df)

P valuebetweengroups

Attitudes1. I feel a personal/professional responsibility to

contribute to improving the health of Aboriginaland/or Torres Strait Islander people1

No plc 27 31 44 51 15 17 119 2.1 0.34Plc 15 45 14 42 4 12 (2)

2. It is important for dietitians to know aboutAboriginal and/or Torres Strait Islander historyand cultures2

No plc 56 68 26 32 114 4.4 0.04Plc 28 88 4 13 (2)

3. I have the responsibility to seek out opportunitiesto learn about Aboriginal and/or Torres StraitIslander health and cultures1

No plc 24 29 43 51 17 20 115 7.0 0.03Plc 17 55 11 35 3 10 (2)

4. Aboriginal and/or Torres Strait Islander peopleface significant barriers to achieving optimal healthoutcomes2

No plc 68 80 17 20 118 1.0 0.60Plc 29 88 4 12 (1)

5. Aboriginal and/or Torres Strait islander peoplemay define the concept of health in different waysto mainstream Australia2

No plc 46 55 38 45 117 3.2 0.07Plc 24 73 9 27 (1)

6. You can tell who is an Aboriginal and/or TorresStrait Islander person by looking at them3

No plc 40 52 37 48 109 2.7 0.44Plc 14 44 18 56 (1)

7. Aboriginal and/or Torres Strait Islander health is a‘niche’ area of practice that is not relevant to mosthealth professionals4

No plc 18 22 46 56 18 22 112 2.0 0.37Plc 4 13 16 53 10 33 (2)

Self-confidence5

8. I am confident that I can work with Aboriginaland/or Torres Strait Islander people in a culturallysafe manner

No plc 40 47 26 30 20 23 120 13.3 0.001Plc 28 82 2 6 4 12 (2)

9. I am apprehensive about interactions withAboriginal and/or Torres Strait Islander people

No plc 21 24 19 22 47 54 120 7.7 0.02Plc 1 3 7 21 25 76 (2)

1 Chi-square test was undertaken using the ‘strongly agree’, ‘agree’ and ‘neutral’ categories.2 Chi-square test was undertaken using ‘strongly agree’ and ‘agree’ categories’.3 Chi-square test was undertaken using the ‘disagree’ and ‘strongly disagree’ categories’4 Chi-square test was undertaken using the ‘neutral’, ‘disagree’ and ‘strongly disagree’ categories5 For the self-confidence statements, ‘strongly agree’ and ‘agree’ responses were combined to form an overarching ‘agree’ category and‘strongly disagree’ and ‘disagree’ responses were combined to form an overarching ‘disagree’ category

Plc, Aboriginal health placement experience.

R. Svarc et al.

452 © 2018 Dietitians Association of Australia

and communities they had learned about in the classroom.Graduates reported developing positive sentiments towardsAboriginal people such as compassion and respect. Situatedlearning experiences acted as a tangible platform for thedevelopment of greater awareness and understanding, andwere perceived by graduates as a foundational step towardsdeveloping cultural competence.

Theme 2: Breaking down stereotypes: Aboriginal healthplacements were emphasised by most graduates as powerfulcatalysts for breaking down stereotypes regarding Aborigi-nal people. For some students, it was their first opportunityto meet an Aboriginal person. Many graduates mentionedthat they realised they had previously been perpetuatingstereotypes about Aboriginal people’s physical appearance,character and health behaviours.

‘It was just so different to what [I thought] at the timebecause I grew up [with] very little exposure to Aborigi-nal people… so it was very different to what I had in mythoughts about what Aboriginal people were like andwhat their lives were like’—Graduate 12, 1-day ruralplacement.

Through breaking down stereotypes, Aboriginal healthplacements were reported to increase their confidence wheninteracting with Aboriginal people. Graduates identifiedthat, as students, they felt ignorant and apprehensive aboutinteracting with Aboriginal people and were afraid of caus-ing offence. Following placement experiences, graduatescommonly reported feeling more comfortable engaging withAboriginal people.

‘I think it put me a bit more at ease, like I wasn’t worriedso much anymore about conducting myself in a certainway’—Graduate 1, 8-week urban placement.

‘It definitely… helped me as a clinical dietitian when I’vehad Aboriginal clients or patients, [to] not be as afraidabout approaching them’.—Graduate 15, 8-week urbanplacement.

Aboriginal health placements assisted graduates’ confi-dence in their ability to interact with Aboriginal people.This occurred through opportunities for students to engagein conversation with Aboriginal people about their culturein a supported environment and to reflect upon their ownculture and beliefs. Graduates with longer placements,which included a situated learning experience often feltmore prepared to work in Aboriginal health compared withthose who had shorter placements or minimal contact withAboriginal people.

Theme 3: Empathy through learning from Aboriginal people:Learning about Aboriginal history and culture was a keyaspect of placement that participants believed contributedtowards increasing their empathy and cultural understand-ing. This learning was particularly effective when it camedirectly from Aboriginal people, such as Aboriginal healthprofessionals and other members of the community. Other

areas of learning included food and nutrition issues, effec-tive communication, and the existence and role of Aborigi-nal health services and professionals.

‘Definitely meeting all the different [Aboriginal] peopleand hearing everybody’s stories… did really help withbeing able to put myself into their shoes and think fromtheir perspectives about issues around food and foodaccess’—Graduate 1, 8-week urban placement.

Graduates also articulated how Aboriginal people onplacement taught them to interpret non-verbal communica-tion such as silences in conversation, and how they havetaken this information forward into their work as practi-tioners in Aboriginal health.

‘A few of the cultural things that the health promotionworker… shared with us have sort of stuck with melike… allowing those long silences in conversation orensuring that I come across as equal not authoritativeand learning from each other, rather than being theteacher’—Graduate 14, 8-week urban placement.

Most graduates also reported cultural awareness training,delivered by Aboriginal staff, as being an essential introduc-tion towards cultural competence. Graduates noted itsimportance in raising awareness not just of the perspectivesof Aboriginal people, but also the diversity that existswithin the Aboriginal population and betweencommunities.

‘[In cultural awareness training] there was a lot of talkabout… the variety there is in Aboriginal culture interms of location…like that Aboriginal community inNorthern Territory is a lot different to the Aboriginalcommunity in Victoria’—Graduate 7, 8-week urbanplacement.

Theme 4: Aboriginal health role-models: Exposure to rolemodels was identified as an important feature of Aboriginalhealth placements to facilitate cultural competency. Dieteticsupervisors were reported to be inspiring role models,whose examples were followed regarding how to communi-cate and build rapport with Aboriginal people. Throughdiscussion with and observation of their supervisors, gradu-ates could appreciate the importance of listening, and thesignificance of developing trust and respect as the gatewayto building strong and positive relationships with Aborigi-nal people.

‘But I remember [my supervisor] just walking up to hercasually and... just talking to her like another humanbeing, like another woman; showing the community thatshe was approachable and wasn’t judging them’—

Graduate 4, 1-day rural placement plus 1 week urbanplacement.

Graduates expressed how positive role models andteachers on placement increased their understanding of theimpact that discrimination and racism has upon cross-

Aboriginal health dietetic university placements

© 2018 Dietitians Association of Australia 453

cultural relationships both within the health profession andbroader Australian society.

‘…So it’s like you [a non-Aboriginal health worker] haveto prove that you’re not racist… it’s a different kind oftrust that needs to be built’—Graduate 4, 1-day ruralplacement plus 1 week urban placement.

Graduates also articulated how their supervisortransformed their view of the relative roles of the die-titian in facilitating better health outcomes for Aborigi-nal communities.

‘One of the first things they [my supervisors] relayed tomyself and the other student was how we’re here to lis-ten… if you’re working alongside Aboriginal people, lis-ten; take what they have on board, see how you cansupport their views, because they’re the experts….’—Graduate 6, 8-week urban placement.

Discussion

The present study aimed to explore the impact of Aborigi-nal health placement experiences on the preparation of die-tetic graduates for practice with Aboriginal communities.The quantitative findings indicate an association betweenAboriginal health placement experiences and positive atti-tudes towards Aboriginal people/health, and self-confidencein working in Aboriginal health. The qualitative findingssuggest that situated learning experiences, breaking downstereotypes, empathy through learning from Aboriginal peo-ple and exposure to role models are the mechanismsbehind the effect of Aboriginal health placement experi-ences. These findings are significant in considering thedevelopment of the non-Aboriginal dietetic workforce intothe future.

Graduates who had an Aboriginal health placementexperience tended to have more positive attitudes towardsAboriginal people/health. Situated learning experiences andlearning directly from Aboriginal people may have increasedgraduates’ empathy for Aboriginal people by providingfirst-hand opportunities for graduates to learn about thecomplex socioeconomic factors affecting Aboriginalhealth.35 The literature supports the efficacy of situatedlearning.36–39 Some authors suggest that no amount ofreading or theory can adequately prepare graduates to pro-vide appropriate health care for Aboriginal people.37 Wepropose that greater insight into Aboriginal health throughsituated learning experiences facilitated graduates’ compas-sion for, and personal connection with, Aboriginal people.This is supported by the findings of previous studies aboutAboriginal health placement experiences.19 Evidence sug-gests that situated learning experiences may also initiategreater and longer lasting changes than university-basedlearning experiences.39

Previous research suggests that students are often initiallyapprehensive about interactions with Aboriginal people, butcan overcome this through increased cultural familiarity via

situated learning experiences.17,37 The quantitative findingsof the present study support this by indicating an associa-tion between Aboriginal health placement experiences andincreased confidence in working with Aboriginal people,while the qualitative findings suggest that situated learningexperiences may be a factor in this change. Situated learn-ing experiences may therefore be a trigger for positivechanges in attitude and increasing confidence in workingwith Aboriginal people.

Our data also suggest that placement assisted graduatesto identify biased attitudes they unconsciously held towardsAboriginal people, through learning from Aboriginal peopleand dietetic supervisors. Many interviewed graduates identi-fied that before their placement experience, they had littleexposure to Aboriginal people. This is not surprisinggiven it is estimated that less than 1 in 10 non-AboriginalAustralians know and mix regularly with Aboriginal peo-ple.40 Experiences during Aboriginal health placementcould be described as a ‘disorientating dilemma’, where anexperience challenges an individuals’ current beliefs,assumptions and expectations about the world.41 Throughcritical reflection encouraged by Aboriginal people and die-tetic supervisors, graduates in the present study reportedbeing able to deepen their cultural understanding in abroader attempt to make sense of their altered worldview.This notion is supported by previous research that hasexplored the role of both Aboriginal mentors and non-Aboriginal mentors that are experienced in cross-culturalsituations, in developing a health professional’s culturalcompetence.42,43 In addition, study results suggest that bylearning how to build trust and respect with Aboriginalpeople through observation of role models in Aboriginalhealth, dietetic graduates may be better equipped todevelop positive relationships with Aboriginal people andhave increased confidence to work in the field. We believethese are crucial steps towards developing cultural compe-tence, as a lack of confidence or fear of working withAboriginal people has been identified to affect health pro-fessionals’ ability to work with Aboriginal people in a cul-turally appropriate manner.44,45

While the present study provides evidence of the abilityof Aboriginal health placements to increase dietetic gradu-ates’ positive attitudes and self-confidence towards workingin Aboriginal health, it may not be feasible to provide theseexperiences to all students. Rather, it may be more feasibleto incorporate elements of placement experiences into cur-riculum content. Further research is required to determinewhat feasible learning experiences could be developed thatincorporate the important elements of placement experi-ences and to test their effectiveness.

The low response rate to the survey may be considered alimitation to this study as the views reported may representa bias (e.g. social desirability), limiting the generalisabilityof the survey findings. Also, low response rate meant thatdifferences between placement settings (i.e. communitycompared with organisation setting) could not be explored.In addition, the survey was a non-validated and self-reported measure of attitudes and self-confidence towards

R. Svarc et al.

454 © 2018 Dietitians Association of Australia

working in Aboriginal health, which may not be a true indi-cation of effective practice with Aboriginal communities.However, the mixed method approach which used in-depthexploration of placement experiences in addition to the sur-vey results, provide credibility and support to the surveyfindings.23 Future research is required to explore the per-ceptions of Aboriginal people in relation to dietetic gradu-ates’ effectiveness of practice with Aboriginal communitiesto provide further credibility to the findings.

Aboriginal health placement experiences may providedietetic graduates with more positive attitudes and self-confidence in working in Aboriginal health. In addition,Aboriginal health placements which incorporate situatedlearning experiences, breaking down stereotypes, learningfrom Aboriginal people, and exposure to role models mayprovide an effective strategy for preparing dietetic graduatesfor practice with Aboriginal communities. This in turn hasthe potential to improve health outcomes of AboriginalAustralians into the future.

Funding source

The authors declare that no funding was received for thisstudy.

Conflict of interest

The authors have no conflicts of interest to declare.

Authorship

Study conception: CP, JB, RD, KM. Study design: CP, JB,RD, KM, RS, CD, HM, JP, SOR, CM. Collection of data: RS,CD, HM, JP. Analysis and interpretation of data: RS, CD,HM, JP, JB, RD, CP, KM, SOR, CM. Paper preparation: RS,CD. Revision of manuscript: RS, CD, CP, JB, RD, CM, SOR,HM, JP, KM. All authors have actively contributed, readand approved the final manuscript.

References

1 Australian Bureau of Statistics. Aboriginal and Torres StraitIslander Life Expectancy Increases. Canberra: Australian Govern-ment, 2013.

2 Trewin D, Madden R. The Health and Welfare of Australia’sAboriginal and Torres Strait Islander Peoples. Canberra:Australian Bureau of Statistics, 2015.

3 Burns J, Thomson N. Review of Nutrition and Growth amongIndigenous Peoples. Australian Indigenous HealthInfoNet: Perth,WA, 2008.

4 Australian Institute of Health and Welfare (AIHW). AustralianBurden of Disease Study: Impact and Causes of Illness and Deathin Aboriginal and Torres Strait Islander People 2011. Canberra:The Institute, 2016.

5 Australian Institute of Health and Welfare (AIHW). Cardiovas-cular Disease, Diabetes and Chronic Kidney Disease – AustralianFacts: Aboriginal and Torres Strait Islander People. Canberra: TheInstitute, 2015.

6 Australian Bureau of Statistics. Australian Aboriginal and TorresStrait Islander Health Survey: First Results, Australia, 2012-13.Canberra: Australian Government, 2013.

7 Rae K, Bohringer E, Ashman A, Brown L, Collins C. Culturalexperiences of student and new-graduate dietitians in theGomeroi gaaynggal ArtsHealth program: a quality assuranceproject. Health Promot J Austr 2016; 27: 162–6.

8 Lawerence S. Development of an Australian Aboriginal and TorresStrait Islander Nutrition Workforce: Information Paper. AustralianIndigenous HealthInfoNet: Perth, WA, 2009.

9 National Aboriginal and Torres Strait Islander Nutrition Work-ing Party. National Aboriginal and Torres Strait Islander NutritionStrategy and Action Plan. Canberra: National Public Health Part-nership, 2001.

10 Baillie E, Bjarnholt C, Gruber M, Hughes R. A capacity-building conceptual framework for public health nutritionpractice. Public Health Nutr 2009; 12: 1031–8.

11 Hughes R. A socioecological analysis of the determinants ofnational public health nutrition work force capacity: Australiaas a case study. Fam Community Health 2006; 29: 55–67.

12 National Health and Medical Research Council (NHMRC).Strengthening Cardiac Rehabilitation and Secondary Prevention forAboriginal and Torres Strait Islander Peoples: A Guide for HealthProfessionals. Canberra: The Council, 2005.

13 Dietitians Association of Australia (DAA). National CompetencyStandards for Dietitians in Australia. Canberra: The Associa-tion, 2016.

14 Newbury JW, Shannon S, Ryan V, Whitrow M. Developmentof ’rural week’ for medical students: impact and quality report.Rural Remote Health 2005; 5: 1–9.

15 Abuzar MA, Burrow MF, Morgan M. Development of a ruraloutplacement programme for dental undergraduates: students’perceptions. Eur J Dent Educ 2009; 13: 233–9.

16 Abuzar MA, Owen J. A community engaged dental curriculum:a rural indigenous outplacement programme. J Public HealthRes 2016; 5: 668.

17 Bennett P, Jones D, Brown J, Barlow V. Supporting rural/remote primary health care placement experiences increasesundergraduate nurse confidence. Nurse Educ Today 2013; 33:166–72.

18 Jamrozik K. Going bush - helping medical students learn fromAboriginal people. Med J Aust 1995; 163: 591–4.

19 Morrissey H, Ball P. First year pharmacy students’ health ser-vices experience at the top-end of Australia. Aust J Pharm2014; 95: 66–71.

20 Warren JM, Irish GL, Purbrick B et al. Developing the futureindigenous health workforce: the feasibility and impact of astudent-led placement programme in remote indigenous com-munities. Aust J Rural Health 2016; 24: 306–11.

21 Wright JR, Bourke L, Waite CJ et al. A short-term ruralplacement can change metropolitan medical students’

knowledge of, and attitudes to, rural practice. Med J Aust

2014; 201: 103–5.22 Edwards C. The indigenous content of Australian nutrition &

dietetics degrees. Aborig Isl Health Work J 2005; 29: 11.23 Creswell JW, Plano Clark VL. Designing and Conducting Mixed

Methods Research. Thousand Oaks, CA: SAGE Publications,Inc., 2006.

24 Doyle L, Brady A-M, Byrne G. An overview of mixed methodsresearch. J Res Nurs 2009; 14: 175–85.

25 Bainbridge R, McCalman J, Clifford A, Tsey K. Cultural Compe-tency in the Delivery of Health Services for Indigenous People.

Canberra: Closing the Gap Clearing House, 2015.

Aboriginal health dietetic university placements

© 2018 Dietitians Association of Australia 455

26 Paez KA, Allen JK, Carson KA, Cooper LA. Provider and cliniccultural competence in a primary care setting. Soc Sci Med2008; 66: 1204–16.

27 Doorenbos AZ, Schim SM, Benkert R, Borse NN. Psychometricevaluation of the cultural competence assessment instrumentamong healthcare providers. Nurs Res 2005; 54: 324–31.

28 McHugh ML. The chi-square test of independence. BiochemMed 2013; 23: 143–9.

29 Sandelowski M. What’s in a name? Qualitative descriptionrevisited. Res Nurs Health 2010; 33: 77–84.

30 Liamputtong P. Research Methods in Health: Foundations for Evi-dence Based Practice. South Melbourne: Oxford UniversityPress, 2010.

31 Sturges JE, Hanrahan KJ. Comparing telephone and face-to-face qualitative interviewing: a research note. Qual Res 2004; 4:107–18.

32 Halcomb EJ, Davidson PM. Is verbatim transcription of inter-view data always necessary? Appl Nurs Res 2006; 19: 38–42.

33 Hoonakker P, Carayon P. Questionnaire survey nonresponse: acomparison of postal mail and internet surveys. Int J Hum Com-put Interact 2009; 25: 348–73.

34 Brown L, CAPRA S, Williams L. Profile of the Australian die-tetic workforce: 1991–2005. Nutr Diet 2006; 63: 166–78.

35 Dharamsi S, Espinoza N, Cramer C, Amin M, Bainbridge L,Poole G. Nurturing social responsibility through communityservice-learning: lessons learned from a pilot project. MedTeach 2010; 32: 905–11.

36 Hart B, Cavanagh M, Douglas D. The "strengthening nursingculture project" – An exploratory evaluation study of nursingstudents’ placements within Aboriginal Medical Services. Con-temp Nurse 2015; 51: 245–56.

37 Benson J, Ryder C, Gill M, Balabanski A. A brief experience formedical students in a remote Aboriginal community. Aust FamPhysician 2015; 44: 752–9.

38 Kent-Wilkinson A, Starr L, Dumanski S, Fleck J, LeFebvre A,Child A. International nursing student exchange: rural andremote clinical experiences in Australia. J Agromedicine 2010;15: 58–65.

39 McAllister L, McEwen E, Williams V, Frost N. Rural attach-ments for students in the health professions: are they worth-while? Aust J Rural Health 1998; 6: 194–201.

40 Walter M. Keeping our distance: non-Indigenous/Aboriginalrelations in Australian society. In: Pietsch J, Aarons H, eds.Australia: Identity, Fear and Governance in the 21st Century.Canberra: ANU Press, 2013; 15–32.

41 Mezirow J. Learning as Transformation: Critical Perspectives on aTheory in Progress. San Francisco, CA: Jossey-Bass Pub-lishers, 2000.

42 Browne J, Thorpe S, Tunny N, Adams K, Palermo C. A qualita-tive evaluation of a mentoring program for Aboriginal healthworkers and allied health professionals. Aust N Z J Public Health2013; 37: 457–62.

43 Palermo C, Hughes R, McCall L. An evaluation of a publichealth nutrition workforce development intervention for thenutrition and dietetics workforce. J Hum Nutr Diet 2010; 23:244–53.

44 Bennett B, Zubrzycki J, Bacon V. What do we know? The expe-riences of social workers working alongside Aboriginal people.Aust Social Work 2011; 64: 20–37.

45 Wilson A, Magarey A, Jones M, O’Donnell K, Kelly J. Attitudesand characteristics of health professionals working in Aborigi-nal health. Rural Remote Health 2015; 15: 2739.

R. Svarc et al.

456 © 2018 Dietitians Association of Australia

ORIGINAL RESEARCH

Anaemia in pregnancy among Aboriginal and TorresStrait Islander women of Far North Queensland:A retrospective cohort study

Dympna LEONARD ,1 Petra BUTTNER,1 Fintan THOMPSON,1 Maria MAKRIDES2 andRobyn MCDERMOTT1

1Centre for Chronic Disease Prevention, Australian Institute of Tropical Health and Medicine, College of Public Health,Medical and Veterinary Sciences, James Cook University, Cairns, Queensland and 2Healthy Mothers, Babies andChildren, South Australian Health and Medical Research Institute, Adelaide, South Australia, Australia

AbstractAim: Anaemia during pregnancy is common worldwide. In Australia between 7.1% and 11% of mothers have beenreported to have anaemia in pregnancy. Higher rates are reported for Aboriginal and Torres Strait Islander women(Townsville: 34.2%, remote Northern Territory: 50%). The present study describes anaemia in pregnancy amongAboriginal and Torres Strait Islander women of Far North Queensland.Methods: Health service information was analysed for 2076 Aboriginal and Torres Strait Islander women who gavebirth between 2006 and 2010. The prevalence of anaemia in pregnancy, characteristics of the mothers and preg-nancy outcomes were described. Logistic regression for bivariate analyses and multivariable linear modelling withand without imputed data were used to compare those mothers who had anaemia in pregnancy with those whodid not.Results: More than half of Aboriginal and Torres Strait Islander women (54.5% (95% CI: 52.4%, 56.7%)) had anaemiain pregnancy. For mothers who gave birth in 2009 and 2010 (n = 1796) with more complete data, those who wereiron deficient during pregnancy were more likely to be anaemic (RR: 1.40, P = <0.001). Mothers (29.0%) from locali-ties of relative socioeconomic advantage had lower risk of anaemia in pregnancy (RR: 0.86, P = 0.003), as didmothers (31.9%) who were obese (RR: 0.87, P = 0.013).Conclusions: The prevalence of anaemia in pregnancy among Aboriginal and Torres Strait Islander women of FarNorth Queensland is high. Prevention and treatment of anaemia will improve the health of these mothers, and possi-bly the health and early development of their children.

Key words: Aboriginal, anaemia, mother, pregnancy, Torres Strait Islander.

Introduction

The ‘First Thousand Days’ from conception to around age2 years is a time of rapid growth and neurological develop-ment.1,2 Anaemia in pregnancy—defined as blood haemo-globin levels below 110 g/L—is a concern because ofpoorer health and pregnancy outcomes of mothers, and

also the potential detrimental effects on the health anddevelopment of their children.1,3

Infections, inflammation and genetic conditions(e.g. thalassaemia) can cause anaemia, as well as iron defi-ciency and other nutritional deficiencies.3 Although anessential nutrient, iron can have negative metabolic effects.4

To prevent damage, iron absorption is tightly regulated butincreases when iron requirements are high, as in preg-nancy.4 The pregnant mother requires iron not only for herimmediate demands—increased blood, tissue growth—butto provide iron stores to her baby.5 The main source of ironfor a baby in the first months of life is not breast milk orinfant formula but these iron stores acquired before birth.5

The high maternal iron requirements mean that anaemia inpregnancy is usually due to iron deficiency and is a strongpredictor of early onset anaemia in the child.3,6 High ratesof early childhood anaemia are a continuing concern inremote Aboriginal and Torres Strait Islander communitiesof northern Australia.7,8

D. Leonard, MPH, APD, PhD CandidateP. Buttner, PhD, Professorial Research FellowF. Thompson, MEpi, Data ManagerM. Makrides, PhD, ProfessorR. McDermott, PhD, ProfessorCorrespondence: D. Leonard, Centre for Chronic Disease Prevention,Australian Institute of Tropical Health and Medicine, College of PublicHealth, Medical and Veterinary Sciences, James Cook University, POBox 6811, Cairns, QLD 4870, Australia.Email: dympna.leonard@jcu.edu.au

Accepted September 2018

© 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of AustraliaThis is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution andreproduction in any medium, provided the original work is properly cited.

457

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 457–467 DOI: 10.1111/1747-0080.12481

Australian studies reporting anaemia in pregnancy includestudies from South Australia (unsupplemented control group:11% anaemic, births 1997–1999, including 3.3% Aboriginalmothers), Western Australia (6.2% anaemic >18 years, births2005–2006, 7.3% Aboriginal and Torres Strait Islandermothers) and for all births in South Australia 1999–2005(7.1% anaemic, 2.5% Aboriginal and Torres Strait Islandermothers).9–11 Higher rates of anaemia in pregnancy have beenreported for Aboriginal and Torres Strait Islander womenaccessing antenatal care in Townsville (34.2% anaemic, births2001–2003) and in two remote Northern Territory communi-ties (50.0% anaemic, births 2004–2006).12,13

Anecdotal reports by health service providers in FarNorth Queensland (Figure 1) indicate that anaemia amongAboriginal and Torres Strait Islander mothers and their chil-dren is also prevalent but published information is lacking.Consequently, research has been undertaken to investigateanaemia among Aboriginal and Torres Strait Islandermothers and their children in Far North Queensland froman intergenerational perspective. Here we describe anaemiain pregnancy, and investigate associations between anaemiaand various maternal characteristics, health indicators, andpregnancy outcomes of these mothers for a pregnancy andbirth between 2006 and 2010.

Methods

This is a retrospective cohort study using linked informa-tion extracted from three existing health service data

collections for mothers resident in Far North Queensland,in respect of pregnancies and births of their babies bornbetween 2006 and 2010.

Data sources: Electronic data systems used by health ser-vice providers store confidential client information withstrict provisions for data security and confidentiality. How-ever de-identified information may be released for researchpurposes, subject to stringent processes to ensure data secu-rity and confidentiality. The process of securing the neces-sary approvals and release of a linked, de-identified datasethas been described elsewhere.14 Briefly, data collectionsaccessed were the Queensland Perinatal Data Collection(PDC); the Queensland Health Pathology Services Data Col-lection (Auslab); and the community health services elec-tronic record system, Ferret, used mainly in remotelocations of Far North Queensland (Supporting informationTable S1). Information extracted from Auslab had beenrecorded from 2000 up to 2010, from Ferret from date ofrollout (see Figure 1), up to 2010 and from PDC from2006 to 2010. Individual records were linked and de-identified by the Queensland Health Statistical ServicesBranch for release to the research group in May 2017.

Participants: Study data were extracted from these datacollections for two cohorts—the Cape York cohort and the2009–2010 cohort.

The Cape York cohort includes mothers of Aboriginaland Torres Strait Islander children born between 2006 and2008, where the child had previously been included in anunpublished health service review of childhood growth inremote Cape York communities.

Figure 1 Far North Queensland: Hospital and Health Service boundaries, and location and year of rollout of the Ferret elec-tronic health records system. Based on the information provided by the Ferret Support team, Queensland Health, May 2017.Reproduced with permission of the Australian New Zealand Journal of Public Health.14

D. Leonard et al.

458 © 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

The 2009–2010 cohort includes all Aboriginal and/orTorres Strait Islander mothers with a PDC record for a birthin 2009 or 2010 in Far North Queensland.

As children were recruited to the Cape York childgrowth research after the neonatal period, information onperinatal mortality is not available for the Cape Yorkcohort. Perinatal mortality information is available forbabies of the 2009–2010 cohort.

Ethics approval was granted by Queensland Health FarNorth Queensland Human Research Ethics Committee(HREC/15/QCH/50-980) in June 2015. Subsequent toapplications to the respective Data Custodians for datarelease, approval under the Queensland Public Health Act2005 was granted by the Director General of QueenslandHealth in February 2016.

Variables and definitions: Anaemia in pregnancy wasdefined as haemoglobin less than 110 g/L as used byQueensland Health.15,16 Measurements of haemoglobin usedhere are results of pathology laboratory measurements. Irondeficiency was defined as Ferritin levels below 15 ug/L.15

Information recorded on the PDC includes mothers’ eth-nicity, parity, pre-pregnancy weight, height, smoking inpregnancy; birth status of babies—live/still born, gestationalage at birth. Other information was derived from PDCrecords (maternal age, teenage mothers, body mass indexcategories, prematurity and birthweight category) as definedby the Australian Institute of Health and Welfare and theNational Health and Medical Research Council—seeTable S2.17–19 Pre-existing diabetes was defined as a fastingoral glucose tolerance test result ≥7.0 mmol/L and/or a gly-cated haemoglobin reading ≥6.5%. Gestational diabetes wasdefined as an oral glucose tolerance test result ≥5.1 (fasting)and/or 10.0 (1 hour) and/or 8.5 mmol/L (2 hours) amongwomen without pre-existing diabetes.16,20,21 For definitionsof hypertension, iron deficiency, low red cell folate (RCF)and vitamin B12 levels, see Table S2. Information on foodinsecurity, diet or nutrient supplements are not recorded inthese electronic data collections.

The Socio-Economic Index for Areas (SEIFA 2011) ranksAustralian Bureau of Statistics Statistical Local Areas (SLAs)by deciles of relative socioeconomic advantage and disad-vantage.22 A ranking of ‘1’ indicates a locality of greatest rel-ative disadvantage while a ranking of ‘10’ indicates alocality of greatest relative advantage.22 The appropriateSEIFA decile ranking was allocated to each mother basedon her usual place of residence. For the purpose of thisanalysis, SEIFA deciles (1–10) were reduced to two catego-ries: SEIFA deciles 1 and 2 (the 20% most disadvantagedSLAs in Australia) or SEIFA decile 3 or higher. These cate-gories were selected as most of the mothers in this study(71.1%) lived in SEIFA categories 1 and 2.

Statistical analysis: Categorical variables were describedusing absolute and relative frequencies. The distribution ofnumerical variables were assessed; symmetrically distributednumerical characteristics were described using mean values,SDs and ranges; numerical values with a skewed distribu-tion (parity, Ferritin levels, baby’s gestational age at birth)were described using median, interquartile ranges (IQRs)

and ranges. The prevalence of anaemia was presented with95% confidence intervals (95% CIs).

Bivariate analysis: Characteristics of the mothers and theirpregnancy outcomes were compared between thosemothers who had been anaemic in pregnancy and thosewho had not, using logistic regression.

Multivariable analysis: The following characteristics wereconsidered during multivariable analyses (Cohort1 ‘2009–2010 cohort’ n = 1796; Cohort 2 ‘Cape York cohort’n = 280). Variables with complete dataset were ethnicity ofmother, age of mother, SEIFA category for residence ofmother, five or more antenatal care visits, pregnancy inducedhypertension, birthweight of baby (Cohort 2: complete data-set). Variables with missing values were: BMI category ofmother, parity, smoking during pregnancy, mother with pre-existing diabetes, gestational diabetes, low RCF value before orduring pregnancy, low vitamin B12 value before or duringpregnancy, iron deficiency during pregnancy, birthweight ofbaby (missing values Cohort 1). The number of missing valuesfor variables used in multivariable analyses is shown inTables 1–2 and Tables S1-S4.

‘Missing-ness’: Examination of patterns of missing datashowed data missing for some key variables; year of birth ofbaby (that is, the cohort) was significantly associated withmissing body mass index (P < 0.001), missing parity(P = 0.042), and missing iron status (P < 0.001), resultingin more missing data for the Cape York cohort mothers.Consequently it was decided to conduct analysis stratifiedby cohort. Tables S3 and S4 provide more information onmissing values and patterns of ‘missing-ness’.

Multivariable general linear models for the binomial familyusing the log link to estimate relative risks (RRs) were used toidentify independent risk factors for anaemia during preg-nancy for the complete case analysis. Backward and forwardstepwise modelling procedures were initially conducted toestablish basic multivariable models for both cohorts. Charac-teristics that were not part of the basic models were assessedfor potential confounding effects. A confounder was assumedto be a variable that changed estimates of characteristics in thebasic model by 10% or more.23 Once a model was estab-lished, all possible two-way interactions involving variables inthe model were assessed for statistical significance.

Multiple imputation: Multivariate multiple imputation wasconducted using Stata’s MI commands for sequential imputa-tion using chained equations. Missing values were imputedfor BMI of mother; parity; smoking during pregnancy;mother with pre-existing or gestational diabetes; iron defi-ciency during pregnancy; and birthweight of baby. Low RCFand vitamin B12 values before or during pregnancy were notimputed because these characteristics were missing in closeto 80% of cases in both cohorts and they did not show sta-tistically significant associations during bivariate or multivari-able complete case analyses. Before imputation, patterns ofmissing values were investigated and assumed to be ‘missingat random’ in each cohort.24 Linear regression was used toimpute missing values of continuous characteristics; logisticregression was used to impute missing values of dichoto-mous characteristics. Imputation models were based on the

Anaemia in pregnancy

© 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

459

following variables: anaemia during pregnancy, pregnancyinduced hypertension, ethnicity, age, SEIFA index and ante-natal care received. Twenty imputed datasets were createdfor each cohort. Multivariable general linear models for thebinomial family using the log link to estimate RRs were usedto identify independent risk factors for anaemia during preg-nancy for imputed data.

Results of multivariable models for complete case andimputed data analyses are presented as RRs and 95% CIs.

P-values of less than 0.05 were considered statistically sig-nificant. Analysis was conducted using Stata version 13 (Sta-taCorp, Lakeway Drive, College Station, Texas).

Results

Data provided in May 2017 included information for 2332mothers who gave birth to 2548 Aboriginal and Torres

Table 1 Prevalence of anaemia during the cohort pregnancy by characteristics of mothers, 2009–2010 cohort

Characteristics of mothers(data available and data missing1)

Mothers(n)

Motherswith anaemiain cohortpregnancy (n)

Prevalence ofanaemia inpregnancy,% (95% CI)

P-values(logistic regression)

All 1796 976 54.3% (52.0%, 56.6%) —

Ethnicity1 (complete dataset)Aboriginal 910 493 54.2% (50.9%, 57.4%) BaseTorres Strait Islander 653 363 55.6% (51.8%, 59.4%) 0.58Both Aboriginal and Torres Strait Islander 233 120 51.5% (45.0%, 58.0%) 0.465

Usual residence (complete dataset)Cairns and Hinterland 1133 581 51.3% (48.4%, 54.2%) BaseCape York 220 120 54.5% (47.9%, 61.2%) 0.385Torres and Northern Peninsula Area 443 275 62.1% (57.5%, 66.6%) <0.001*

SEIFA category1 (complete dataset)SEIFA 1 or 2 1276 719 56.3% (53.6%, 59.1%) 0.008*SEIFA 3–10 520 257 49.4% (45.1%, 53.7%)

Teenage mother 376 230 61.2% (56.2%, 66.1%) 0.003*Antenatal Care 5 visits or more1 (complete dataset),

n = 1417 (78.9%)1417 770 54.3% (51.7%, 56.9%) 0.996

Smoking this pregnancy1 (1791 records—5 missing) 1022 567 55.5% (52.4%, 58.5%) 0.259Body mass index categories2 (all ages—1675 measures—121 missing)1

Under weight (5.4%) 90 59 65.5% (55.5%, 75.6%) 0.152Healthy weight (36.7%) 615 354 57.6% (53.6%, 61.5%) BaseOver weight (26.0%) 436 233 53.4% (48.7%, 58.1%) 0.185Obese (31.9%) 534 246 46.1% (41.8%, 50.3%) <0.001*

Glucose tolerancePre-existing diabetes1 (1239 pathology measures—

557 missing)75 45 60.0% (48.7%, 71.3%) 0.390

Gestational diabetes1 (793 pathology measures—1003 missing)

141 77 54.6% (46.3%, 62.9%) 0.595

Pregnancy-induced hypertension1 (complete dataset) 92 58 63.0% (53.0%, 73.1%) 0.087Nutrient status before/during cohort pregnancy

Ever iron deficient during cohort pregnancy1

(1133 pathology measures—663 missing)672 452 67.3% (63.7%, 70.8%) <0.001*

Ever iron deficient before cohort pregnancy (561pathology measures)

324 224 69.1% (64.1%, 74.2%) <0.001*

Low red cell folate before/during cohortpregnancy1 (376 pathology measures—1420missing), n = 66 (17.6%)

66 44 66.7% (55.0%, 78.3%) 0.785

Low B12 before/during cohort pregnancy1 (328pathology measures–1468 missing), n = 63(19.2%)

63 48 76.2% (65.4%, 87.0%) 0.202

1 Information on number of missing values provided for those variables used for multivariable analysis.2 Criteria for body mass index categories for adults applied for mothers aged 18 years and older, and age-based criteria for mothers youngerthan 18 years (where available, n = 1675).

* P-value less than 0.05.

D. Leonard et al.

460 © 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

Strait Islander babies between 1 January 2006 and31 December 2010 in Far North Queensland.

Exclusions: For the purpose of this report, non-Indigenous mothers (n = 15) and mothers normally resi-dent outside of Far North Queensland (n = 29) wereexcluded. Births that were not the first birth in the cohortyears were excluded (n = 289). Mothers with missing infor-mation for haemoglobin levels during pregnancy (n = 119)were also excluded. Following exclusions, the resultantdataset included information for 2076 Aboriginal andTorres Strait Islander mothers who gave birth to 2095

babies including 19 sets of twins, between 2006 and 2010(Figure 2).

The 2009–2010 cohort: After exclusions there were 1796mothers of whom half (50.7%) were Aboriginal, 36.4%Torres Strait Islander and 13.0% both Aboriginal andTorres Strait Islander (Table 1). More than half were nor-mally resident in the Cairns and Hinterland Health ServiceDistrict (63.1%), the remaining in Cape York (12.3%) orthe Torres Strait and Northern Peninsula Area (24.7%)(Figure 1). Most mothers (71.1%) lived in localities with aSEIFA ranking in the lowest or second lowest decile.22

Table 2 Prevalence of anaemia during the cohort pregnancy by characteristics of mothers, Cape York cohort

Characteristics of mothers (data available and datamissing1)

Mothers(n)

Mothers withanaemia incohortpregnancy (n)

Prevalence ofanaemia inpregnancy,% (95% CI)

P-values(logistic regression)

All 280 156 55.7% (49.9%, 61.6%) n/aEthnicity1 (complete dataset)

Aboriginal 249 140 56.2% (50.0%, 62.4%) BaseTorres Strait Islander, n = 16 (5.7%)2 16 — — 0.335Both Aboriginal and Torres Strait Islander, n = 15

(5.4%)215 — — 0.775

Usual residence (complete dataset)Cairns and Hinterland 18 9 50.0% (24.4%, 75.6%) BaseCape York 261 146 55.9% (49.9%, 62.0%) 0.625Torres and Northern Peninsula Area2 — — — n/a

SEIFA category1 (complete dataset)SEIFA 1 or 2 255 144 56.5% (50.3%, 62.6%) 0.81SEIFA 3–10 25 12 48.0% (27.0%, 69.0%)

Teenage mother 57 39 68.4% (56.0%, 80.9%) 0.032*Antenatal Care 5 visits or more1 (complete dataset) 251 137 54.6% (48.4%, 60.8%) 0.265Smoking this pregnancy1 (278 record—2 missing) 187 108 57.8% (50.6%, 64.9%) 0.257Body mass index categories3 (all ages 107 measures—173 missing)

Under weight (23.4%) 25 14 56.0% (35.1%, 76.9%) 0.721Healthy weight (35.5%) 38 23 60.5% (44.2%, 76.8%) BaseOver weight (25.2%)2 27 — # 0.033*Obese (15.9%)2 17 — # —

Glucose tolerance2

Pre-existing diabetes1 (212 pathology measures—68 missing)

10 — — 0.783

Gestational diabetes1 (140 pathology measures—140 missing)

28 — # 0.006*

Pregnancy-induced hypertension1 (complete dataset)2 2 — — 0.097Nutrient status (iron, folate, B12) before/during cohort pregnancy

Ever iron deficient during cohort1 pregnancy (136pathology measures—144 missing)

54 34 63.0% (49.6%, 76.3%) 0.178

Ever iron deficient before cohort pregnancy (77pathology measures—203 missing)

37 28 75.7% (61.2%, 90.2%) 0.022*

Low red cell folate before/during cohort pregnancy1

(59 pathology measures—221 missing)2— — — 0.443

Low B12 before/during cohort pregnancy1 (51pathology measures—229 missing)2

— — — n/a

1 Information on number of missing values provided for those variables used for multivariable analysis.2 Numbers too small to report are shown as —.3 Criteria for body mass index categories for adults applied for mothers aged 18 years and older, and age-based criteria for mothers youngerthan 18 years (where available, n = 107).

* P-value less than 0.05.

Anaemia in pregnancy

© 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

461

Table

4Riskfactorsforanaemiadu

ring

pregnancy—

CapeYo

rkmothers

(n=28

0):Multivariableanalysis—

completecase

analysisandanalysiswith

impu

teddata

Characteristic

Com

pletecasesanalysis(n

=79)

Imputeddata

analysis(n

=280)

Anaem

ia,

n=33

(41.8%

)

NoAnaem

ia,

n=46

(58.2%

)

Relativerisk

(95%

confidence

interval)

P-value

Num

berof

missing

values

(%)

Anaem

ia,

n=156

(55.7%

)

NoAnaem

ia,

n=124

(44.3%

)

Relativerisk

(95%

confidence

interval)

P-value

Body

massindex

173(61.8%

)Und

erweight

10(30.3%

)6(13.0%

)1.04

(0.63,

1.72

)0.86

955

(35.4%

)23

(18.8%

)1.07

(0.82,

1.39

)0.60

9Health

yweight

14(42.4%

)11

(23.9%

)1

—64

(40.8%

)37

(29.6%

)1

—

Overweight

8(24.2%

)16

(34.8%

)0.67

(0.35,

1.30

)0.23

827

(17.4%

)34

(27.7%

)0.74

(0.51,

1.07

)0.10

7Obese

1(3.0%)

13(28.3%

)0.13

(0.02,

0.85

)0.03

4*10

(6.4%)

30(23.9%

)0.40

(0.19,

0.86

)0.01

9*

Both

mod

elswereadjusted

fortheconfou

ndingeffectof

having

hadpre-existin

gdiabetes

orgestationaldiabetes

(132

missing

values

impu

ted).Im

puteddata

areaverages

of20

impu

tatio

ns.

*P-valueless

than

0.05

.

Table

3Riskfactorsforanaemiadu

ring

pregnancy—

2009

–20

10mothers

(n=17

96):Multivariableanalysis—

completecase

analysisandanalysiswith

impu

teddata

Characteristic

Com

pletecaseanalysis(n

=1052)

Imputeddata

analysis(n

=1796)

Anaem

ia,

n=619(58.8%

)NoAnaem

ia,

n=433(41.2%

)

Relativerisk

(95%

confidence

interval)

P-value

Num

berof

missing

values

(%)

Anaem

ia,

n=976

(54.3%

)

Noanaemia,

n=820

(45.7%

)

Relativerisk

(95%

confidence

interval)

P-value

Body

massindex

121(6.7%)

Und

erweight

43(7.0%)

20(4.6%)

1.06

(0.88,

1.27

)0.53

366

(6.8%)

33(4.0%)

1.10

(0.94,

1.29

)0.24

6Health

yweight

255(41.2%

)15

2(35.1%

)1

–39

1(40.1%

)27

5(33.5%

)1

–

Overweight

167(27.0%

)11

4(26.3%

)1.0(0.89,

1.12

)0.97

125

3(26.0%

)21

3(26.0%

)0.97

(0.87,

1.08

)0.59

7Obese

154(24.9%

)14

7(34.0%

)0.87

(0.76,

1.0)

0.04

9*26

6(27.2%

)29

9(36.5%

)0.87

(0.77,

0.97

)0.01

3*

SEIFAcategory

nil

SEIFA1and2

468(75.6%

)30

5(70.4%

)1

719(73.7%

)55

7(67.9%

)1

SEIFA3–

1015

1(24.4%

)12

8(29.6%

)0.88

(0.78,

0.99

)0.03

3*25

7(26.3%

)26

3(32.1%

)0.86

(0.78,

0.95

)0.00

3*

Motherwas

iron

deficientdu

ring

pregnancy

663(36.9%

)

No

210(33.9%

)22

2(51.3%

)1

321(32.9%

)42

3(51.6%

)1

Yes

409(66.1%

)21

1(48.7%

)1.34

(1.19,

1.50

)<0

.001

*65

5(67.1%

)39

7(48.4%

)1.40

(1.23,

1.60

)<0

.001

*

Both

mod

elswereadjusted

fortheconfou

ndingeffectof

ageof

mother(nomissing

values

impu

ted).Impu

teddata

areaverages

of20

impu

tatio

ns.

*P-valueless

than

0.05

.

D. Leonard et al.

462 © 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

The mean age of mothers was 25.2 years (SD = 6.4) rang-ing from 13 up to 48 years. One in five mothers (21.0%)were teenagers. Median parity was two, ranging from 0 to16. Most mothers (78.9%) had at least five antenatal health-care visits in pregnancy. More than half (57.1%) smoked inpregnancy. Mean body mass index of the mothers aged18 years and over (measurements available n = 1535) was27.7 (6.6) ranging from 16.0 to 56 kg/m2.

Among mothers with glucose tolerance results, 6.1%(n = 1239) had pre-existing diabetes and 17.8% (n = 794)had gestational diabetes. PDC records showed that 5.1%had pregnancy-induced hypertension. Among mothers withprior records of blood pressure 17.6% (n = 812), hadhypertension. More than half of the mothers with measures

of Ferritin had iron deficiency during (59.3%, n = 1133) orbefore (57.8%, n = 561) the cohort pregnancy.

The Cape York cohort: The majority (88.9%) of thesemothers (n = 280) were Aboriginal and the remaining wereTorres Strait Islander (5.7%) or both Aboriginal and TorresStrait Islander (5.4%) (Table 2). Nearly all (93.2%) wereusually resident in Cape York, 6.4% in Cairns and Hinter-land Health Service District and one mother in the TorresStrait and Northern Peninsula Area (Figure 1). Most(91.1%) mothers lived in localities with a SEIFA ranking inthe lowest or second lowest decile.22

The mean age of these mothers was 25.0 years (SD =6.4) ranging from 15 to 40 years. One in five (20.4%) wereteenagers. Median parity was two, ranging from nil to eight.

*Later births excluded so each mother included once only

•

Exclusions:

•

Non-Indigenous mothers n = 15

Mothers – baby born between 2006 and 2010 in Far North Queensland

n = 2,332* mothers (99.4% Aboriginal and/or Torres Strait Islander)

n = 2,548 babies

•

Mother not usually resident in far north Queensland n = 29

Baby’s birth episode NOT first birth episode in cohort years*:

Second birth episode in cohort years n = 197

Third birth episode in cohort years n = 1

Birth episode in BOTH cohorts - later births excluded n = 91

n = 2,195 unique mothers, n = 2,215 unique babies (including 20 sets of twins)

Describe mothers:

n = 2,076

Cape York Cohort (n = 280)

2009-2010 Cohort (n = 1,796)

Mothers n = 339

Describe pregnancy outcomes (babies at birth):

n = 2,095 (includes 19 sets of twins)

Cape York cohort (n = 283)

2009-2010 cohort (n = 1,812)

Cape York Child Growth cohort

information provided on:

Babies n = 381

(born 2006 to 2008)

Mothers n = 1,993

2009 & 2010 birth cohort

information provided on:

Babies n = 2,167

(born 2009 OR 2010)

Excluding n = 119 mothers with no haemoglobin measurement in pregnancy

Figure 2 Flow chart: mothers and pregnancy outcomes showing study inclusions and exclusions.

Anaemia in pregnancy

© 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

463

Most mothers (89.6%) had at least five antenatal health-carevisits in pregnancy. Many (67.3%) smoked in pregnancy.Mean body mass index of the mothers aged 18 years andover (measurements available n = 98) was 24.4 (5.8) rang-ing from 16.1 to 37.7 kg/m2.

Where glucose tolerance results were available 4.7%(n = 212) had pre-existing diabetes and 20.0% (n = 140)had gestational diabetes. PDC records showed that 7.5% hadpregnancy-induced hypertension. Among mothers with priorrecords of blood pressure, 20.6% (n = 272) had hyperten-sion. Of those with measures of Ferritin, 39.7% (n = 54) hadiron deficiency during the cohort pregnancy while 48.1%(n = 77) had iron deficiency before the cohort pregnancy.

2009–2010 cohort—Pregnancy outcomes: Among the 1812babies born to these 1796 mothers, 53.9% were boys. Sev-enteen (0.9%) were still born and there were 10 (0.6%)neonatal deaths (Table S7). Among the liveborn babies(n = 1795), median gestational age at birth was 39 weeks,ranging from 22 to 42 weeks. Mean birthweight was3240 g (SD = 649 g) ranging from 440 to 5430 g. About1 in 10 babies were low birthweight (10.6%), premature(11.4%) or macrosomic (9.6%) (birthweight ≥ 4000 g).

Cape York cohort—Pregnancy outcomes: Among the283 babies born to these 280 mothers, 51.9% were boys(Table S8). No information was available for perinatal mor-tality for this cohort. Median gestational age was 39 weeks,ranging from 27 to 42 weeks. Mean birthweight was 3097 g(SD = 591 g) ranging from 800 to 5320 g. About one inseven babies (13.8%) were low birthweight, 11.7% prema-ture and some (4.6%) macrosomic (birthweight ≥ 4000 g).

Overall, more than half of the mothers (54.5% (95% CI:52.4%, 56.7%)) had anaemia during pregnancy—2009 and2010 birth cohort mothers: 54.3% (95% CI: 52.0%, 56.6%);Cape York Child Growth mothers: 55.7% (95% CI: 49.9%,61.6%). There was no significant difference in prevalence ofanaemia by cohort (P = 0.668). Compared to those not anae-mic in pregnancy, mothers of the 2009-2010 cohort whohad anaemia in pregnancy were younger (mean age 24.8(24.4, 25.2) v 25.7 (25.3, 26.1) P =0.003, had lower meanBMI (27.0 (26.6, 27.4) v 28.4 (27.9, 28.9) P = <0.001),higher parity (median parity 2 (IQR 1, 4) v 2 (1, 3) P =0.02) and lower Ferritin levels (anaemic mothers medianFerritin 14 (IQR 7,28) ug/L v non-anaemic mothers; 18.7(10, 42) P = 0.039). Mothers with iron deficiency (P <0.001)and those living in socio-economically disadvantaged locali-ties (p = 0.008) were more likely to have anaemia in preg-nancy but mothers who were obese were less likely to haveanaemia in pregnancy (P <0.001) (Table 1 and Table S5).Mothers with anaemia in pregnancy had babies with higherbirth weights than babies of non-anaemic mothers (meangrams 3269 (95% CI 3230, 3308) v 3205 (3159, 3252) P =0.038) and had fewer low birth weight babies (8.5% lowbirth weight babies (95% CI 6.9%, 10.4%) v 13.0% (10.9%,15.5%) P = 0.002. No other differences were seen in preg-nancy outcomes (Table S7). Compared to those not anaemicin pregnancy, Cape York mothers who had anaemia in preg-nancy had lower mean BMI (22.3 (20.9, 23.7) v. 26.0 (24.3,27.7) P = 0.002) and tended to be younger though the

difference in age was not statistically significant (mean age24.4 years (95% CI 23.4, 25.4) v. 25.9 (24.8, 27.0) P =0.054). Mothers who were overweight (n = 9, P = 0.033)were less likely to be anaemic as were mothers with gesta-tional diabetes (n = 7, P = 0.006) but the number of thesemothers was small. No difference was seen in parity (anae-mic mothers median parity 2 (IQR 1, 3) v 2 (1, 3) [P =0.602). Similarly no difference was seen in Ferritin levels(anaemic mothers; median Ferritin 20.8 ug/L (IQR10.3,47.8) v. 25.5 (15.3, 57.3) P = 0.589) but mothers withprior iron deficiency had more anaemia in pregnancy (P =0.022) (Table 2 and Table S6). There were no differences inpregnancy outcomes for anaemic and non-anaemic CapeYork mothers (Table S8). Results of bivariate comparisons ofmothers with anaemia with those without, and their preg-nancy outcomes, are shown in Tables S5–S8.

Multivariable analysis 2009–2010 cohort: After controllingfor age, mothers who were iron deficient during pregnancywere more likely to be anaemic (RR: 1.40, P < 0.001)(Table 3; imputed data analysis). Mothers from relativelyadvantaged localities (SEIFA decile 3 and above) were lesslikely to be anaemic (RR: 0.86, P = 0.003), as were motherswho were obese (RR: 0.87, P = 0.013).

Cape York cohort: After controlling for existing or gesta-tional diabetes, only body mass index remained statisticallysignificant for these mothers (being obese: RR = 0.40,P = 0.019) (Table 4; imputed data analysis).

Discussion

Among the Aboriginal and Torres Strait Islander mothers ofFar North Queensland described here, over half had anae-mia in pregnancy (54.5%, n = 2076). This is much higherthan among pregnant women elsewhere in Australia butsimilar to findings from two remote Northern Territorycommunities, where 50% of mothers had anaemia in preg-nancy.13 These results reflect the higher rates of anaemiareported among Australian Indigenous people in recentnational health surveys.25

Although other conditions can cause anaemia, iron defi-ciency is the ‘usual suspect’ as the cause of anaemia in preg-nancy.15,26 Among the 2009 and 2010 birth mothers, forwhom data were more complete, analysis confirmed that irondeficiency was strongly associated with anaemia in pregnancy.

Compared to the average Australian mother in 2016, themothers described here were younger, more likely to smokeand they had less antenatal care.27 The prevalence of obe-sity was higher among the 2009–2010 birth mothers,although not the Cape York mothers. About one in four ofthese Far North Queensland mothers had diabetes in preg-nancy compared to about one in eight mothers Australia-wide in 2016.27 Previous reports of the poor health andnutrition of young Indigenous women in North Queens-land also flagged the potentially detrimental intergenera-tional effects.28 The pregnancy outcomes reported here,with more premature and low birthweight babies, reflectthe poor nutrition and health status of these mothers. Theassociation of anaemia of mothers with increased birth

D. Leonard et al.

464 © 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

weight of the babies may reflect their marginal nutrition sta-tus, with the requirements of bearing a healthy weight babydepleting the limited nutritional reserves of these mothers.

An unexpected finding was that obese mothers were lesslikely to be anaemic than other mothers, though their ratesof anaemia (46.1%) were still high. It may be that thehigher food intake resulting in obesity provides a highernutrient intake. However, maternal obesity is never recom-mended because of the negative health effects on themother and her baby.28 Instead, mothers need diets suffi-ciently nutritious to meet their requirements without exces-sive energy intakes.

There are limitations in this study, which used health ser-vice information recorded during provision of routine care.Some information of interest is not recorded on electronic datacollections such as indicators of food insecurity or nutrientsupplement use. Missing data for some key variables particu-larly in the early years was a limitation. Some measurementsused in this analysis may have been available only for selectedmothers if clinical protocols prescribed specific pathology testsfor ‘high risk’ mothers. Examples include measurements ofglucose tolerance and Ferritin levels. Because ‘missing at ran-dom’ is an assumption for multiple imputation where ‘missingnot at random’ was suspected, respective characteristics werecarefully analysed in alternative models.

Despite these limitations, the findings reported here areconsistent with the high rates of anaemia among youngAboriginal and Torres Strait Islander children and preg-nant women reported elsewhere in remote Australia.7,13 Arecent review has shown that food insecurity is associatedwith increased risk of anaemia among women and youngchildren in high and low income countries.29 Food insecu-rity and poor diet of Aboriginal and Torres Strait Islanderpeople have been documented in remote settings inAustralia.30,31

There is increasing evidence of the importance of anutrient-rich diet in pregnancy, not only for the mother butfor the physical health and cognitive development of herchild.1,32 A nutrient-rich diet helps prevents anaemia andprovides many nutrients needed for health. For mothers onlow incomes, high cost is a barrier to healthy eating espe-cially in remote settings.33 Mothers with low iron status inpregnancy can benefit from supplements—in addition to ahealthy diet—but reaching those with highest needs is chal-lenging.34 Fortification of flour with folate appears to havebeen especially effective at reaching vulnerable populationgroups.35 But iron is a nutrient with potential for harm soiron supplementation must be targeted to those with spe-cific needs.4

Complementary interventions to improve food security,promote good nutrition, and provide targeted iron supple-mentation and/or fortification are needed, designed and imple-mented in partnership with the Aboriginal and Torres StraitIslander communities.36 Also essential are the policy commit-ment and funding to develop, implement and evaluate theseinterventions.37 To ‘Close the Gap’38 in the health, educationand economic status of Aboriginal and Torres Strait Islanderpeople in Far North Queensland compared to their non-

Indigenous peers, these high rates of anaemia in pregnancyamong Aboriginal and Torres Strait Islander mothers must bereduced.

Funding source

DL was supported by a National Health and Medical ResearchCouncil post-graduate scholarship APP1092732. Other agen-cies provided non-financial support to this research. Wewould like to acknowledge and thank the Aboriginal andTorres Strait Islander leaders of the key community controlledhealth service organisations in Far North Queensland whoconsidered and endorsed the proposed research, providing thesupport that made this research possible. In addition, weacknowledge and thank the Queensland Health Data Custo-dians and their research and data management staff for theirassistance and support for this work.

Conflict of interest

The authors affirm that they have no conflict of interest todeclare.

Authorship

DL conceived the research, obtained the necessary approvalsto secure the data required, conducted preliminary statisticalanalysis and prepared the first draft of this manuscript. FTassisted with data management and preparation, and contrib-uted to statistical analysis. PB contributed to study designand guided and contributed to statistical analysis. RM andMM contributed to study design, and manuscript develop-ment and preparation.

References

1 Georgieff MK, Ramel SE, Cusick SE. Nutritional influences onbrain development. Acta Paediatr 2018; 107: 1310–21.

2 Larson LM, Phiri KS, Pasricha SR. Iron and cognitive develop-ment: what is the evidence? Ann Nutr Metab 2017; 71:S25–38.

3 Pasricha SR, Drakesmith H. Iron deficiency anemia: problemsin diagnosis and prevention at the population level. HematolOncol Clin North Am 2016; 30: 309–25.

4 Prentice AM. Clinical implications of new insights intohepcidin-mediated regulation of iron absorption and metabo-lism. Ann Nutr Metab 2017; 71: S40–8.

5 Dewey KG, Chaparro CM. Iron status of breast-fed infants.Proc Nutr Soc 2007; 66: 412–22.

6 Balarajan Y, Ramakrishnan U, Özaltin E, Shankar AH,Subramanian SV. Anaemia in low-income and middle-incomecountries. Lancet 2011; 378: 2123–35.

7 Aquino D, Leonard D, Hadgraft N, Marley J. High prevalenceof early onset anaemia amongst Aboriginal and Torres StraitIslander infants in remote northern Australia. Aust J RuralHealth 2018; 26: 245–50.

8 Bar-Zeev SJ, Kruske SG, Barclay LM, Bar-Zeev N, Kildea SV.Adherence to management guidelines for growth faltering andanaemia in remote dwelling Australian aboriginal infants and

Anaemia in pregnancy

© 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

465

barriers to health service delivery. BMC Health Serv Res 2013;13: 250.

9 Makrides M, Crowther CA, Gibson RA, Gibson RS, Skeaff CM.Efficacy and tolerability of low-dose iron supplements duringpregnancy: a randomised controlled trial. Am J Clin Nutr 2003;78: 145–53.

10 Smithers LG, Gialamas A, Scheil W, Brinkman S, Lynch JW.Anaemia of pregnancy, perinatal outcomes and children’sdevelopmental vulnerability: a whole-of-population study. Pae-diatr Perinat Epidemiol 2014; 28: 381–90.

11 Lewis LN, Hickey M, Doherty D, Skinner R. How do preg-nancy outcomes differ in teenage mothers? A WesternAustralian study. Med J Aust 2009; 190: 537–41.

12 Panaretto K, Lee H, Mitchell M et al. Risk factors for preterm,low birth weight and small for gestational age birth in urbanaboriginal and Torres Strait islander women in Townsville. AustN Z J Public Health 2006; 30: 163–70.

13 Bar-Zeev S, Barclay L, Kruske S, Kildea S. Factors affectingthe quality of antenatal care provided to remote dwellingaboriginal women in northern Australia. Midwifery 2014; 30:289–96.

14 Leonard D, Buttner P, Thompson F, Makrides M,McDermott R. Linking ‘data silos’ to investigate anaemia amongAboriginal and Torres Strait Islander mothers and children inFar North Queensland. Aust N Z J Public Health September2018. https://doi.org/10.1111/1753-6405.12821.

15 Pasricha S-R, Flecknoe-Brown SC, KJ A et al. Diagnosis andmanagement of iron deficiency anaemia: a clinical update. MedJ Aust 2010; 193: 525–32.

16 Queensland Health. Primary Clinical Care Manual: Sexual andReproductive Health, 9th edn. Cairns: Queensland Health, 2016.

17 National Health and Medical Research Council. Australian Die-tary Guidelines. Canberra: NH&MRC, 2013.

18 Australian Institute of Health and Welfare. Australia’s Mothersand Babies 2015—In Brief. Canberra: AIHW, 2017.

19 Australian Institute of Health and Welfare. Perinatal Deaths inAustralia 1993–2012. Canberra: AIHW, 2017.

20 Metzger BE, Gabbe SG, Persson B et al. International Associa-tion of Diabetes and Pregnancy study groups recommendationson the diagnosis and classification of hyperglycemia in preg-nancy. Diabetes Care 2010; 33: 676–82.

21 Queensland Health and the Royal Flying Doctor Service(Queensland Section) and Apunipima Cape York Health Coun-cil. Chronic Conditions Manual: Prevention and Management ofChronic Conditions in Australia. Cairns: The Rural and RemoteClinical Support Unit, Torres and Cape York Hospital andHealth Service, 2015.

22 Australian Bureau of Statistics. Australian Census of Populationand Housing 2011: Socio-Economic Indexes for Areas (SEIFA).Canberra: ABS, 2013.

23 Kleinbaum DG, Kupper LL, Morgenstern H. EpidemiologicResearch. Principles and Quantitative Methods. New York, NY:Van Nostrand Reinhold, 1982.

24 Sterne JAC, White IR, Carlin JB et al. Multiple imputation formissing data in epidemiological and clinical research: potentialand pitfalls. BMJ 2009; 338: b2393.

25 Australian Bureau of Statistics. Australian Aboriginal and TorresStrait Islander Health Survey: Biomedical Results 2012–13. Can-berra: ABS, 2014.

26 Pasricha SR. Anaemia in pregnancy—not just iron deficiency.Acta Haematol 2013; 130: 279–80.

27 Australian Institute of Health and Welfare. Australia’s Mothersand Babies 2016—In Brief. Canberra: AIHW, 2018.

28 McDermott R, Campbell S, Li M, McCulloch B. The health andnutrition of young indigenous women in North Queensland—intergenerational implications of poor food quality, obesity,diabetes, tobacco smoking and alcohol use. Public Health Nutr2009; 12: 2143–9.

29 Moradi S, Arghavani H, Issah A, Mohammadi H, Mirzaei K.Food insecurity and anaemia risk: a systematic review andmeta-analysis. Public Health Nutr July 2018: 1–13.

30 Lee A, Rainow S, Tregenza J et al. Nutrition in remote aboriginalcommunities: lessons from Mai Wiru and the Anangu PitjantjatjaraYankunytjatjara lands. Aust N Z J Public Health 2015; 40: S81–8.

31 Brimblecombe J, Ferguson M, Liberato SC, O’Dea K. Charac-teristics of the community-level diet of aboriginal people inremote northern Australia. Med J Aust 2013; 198: 380–4.

32 Ojha S, Robinson L, Symonds ME, Budge H. Suboptimalmaternal nutrition affects offspring health in adult life. EarlyHum Dev 2013; 89: 909–13.

33 Ferguson M, King A, Brimblecombe JK. Time for a shift infocus to improve food affordability for remote customers. MedJ Aust 2016; 204: 409.

34 Callander EJ, Schofield DJ. Is there a mismatch between whogets iron supplementation and who needs it? A cross-sectionalstudy of iron supplements, iron deficiency anaemia and socio-economic status in Australia. Br J Nutr 2016; 115: 703–8.

35 Hilder L. Neural Tubes Defects in Australia, 2007–2011. Can-berra: Australian Department of Health, 2016.

36 Lee A, Ride K. Review of Nutrition among Aboriginal and TorresStrait Islander People. Perth: Edith Cowan University, 2018.

37 Browne J, Hayes R, Gleeson D. Aboriginal health policy: isnutrition the ‘gap’ in ‘Closing the Gap’. Aust N Z J Public Health2014; 38: 362–9.

38 Department of Prime Minister and Cabinet. Closing the GapPrime Minister’s Report 2017. Canberra: Commonwealth ofAustralia, 2017.

Supporting information

Additional Supporting Information may be found in theonline version of this article at the publisher’s web-site:

Table S1 Data collections used as information sources todescribe rates of anaemia in pregnancy and health indica-tors among Aboriginal and Torres Strait Islander mothers inFar North Queensland

Table S2 Definitions of variables used to describe Aboriginaland Torres Strait Islander mothers and their pregnancy out-comes in Far North Queensland, between 2006 and 2010

Table S3 Number and percentage (%) of missing values forvariables used for multivariable analysis for risk factors foranaemia in pregnancy for each cohort

Table S4 Assessment of ‘Missing-ness’: P-values are resultsof statistical tests correlating missing (yes/no) with cohort,outcome anaemia, variables with no missing values andwith other missing variables

Table S5 Mothers 2009 – 2010 cohort - comparison ofmothers with anaemia during the cohort pregnancy withmothers who did not have anaemia during this pregnancy:Bivariate analysis – logistic regression

D. Leonard et al.

466 © 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

Table S6 Cape York mothers - comparing mothers withanaemia during the cohort pregnancy with mothers whodid not have anaemia during this pregnancy: Bivariate anal-ysis – logistic regression

Table S7 Pregnancy outcomes - 2009–2010 births: compar-ing those babies whose mother had anaemia in pregnancywith those babies whose mother did not have anaemia in

pregnancy - results based on 1812 babies of whom 1795were live born babies: Bivariate analysis – logistic regression

Table S8 Pregnancy outcomes - Cape York births: compar-ing those babies whose mother had anaemia in pregnancywith those babies whose mother did not have anaemia inpregnancy (n = 283 - all babies live-born): Bivariate analysis– logistic regression

Anaemia in pregnancy

© 2018 The Authors. Nutrition & Dietetics published by John Wiley & Sons Australia, Ltdon behalf of Dietitians Association of Australia

467

ORIGINAL RESEARCH

Food security in a sample of clients attending HIVclinics in New South Wales: A cross-sectional study

Angela LANGTON,1 Rosalind MOXHAM ,1 Sarah HIRST,2 Louise HOUTZAGER,3 Julie COUTELAS,4

Amanda RIDER4 and Christine YUSUF51Positive Central, Redfern Health Centre and 2Administration, Redfern Health Centre, Sydney Local Health District,3Albion St Centre, South East Sydney Local Health District, 4HIV Outreach Team, Sydney South East Local HealthDistrict and 5Liverpool HIV clinic, Sydney South West Local Health District, Sydney, Australia

AbstractAim: The aim of this study was to determine the prevalence, demographics and location of food insecurity(FI) among people living with HIV at six health sites in Sydney, Australia and to identify the factors relating to FI.Methods: This was a cross-sectional study recruiting people living with HIV receiving HIV care from six sites acrossthe Sydney metropolitan area. The United States Department of Agriculture abbreviated six-item Subset Food Insecu-rity Tool was used to assess FI and a demographic questionnaire was completed. Bivariate analysis was conducted toinvestigate differences between variables. Descriptive and frequency statistics were used to collate the demographicquestionnaire and determine the prevalence of FI. All tests performed were two sided with a P-value of less than0.05, or 95% confidence interval not overlapping, indicating a statistically significant association.Results: Of the 162 participants 47% (n = 76) reported FI. The percentage of FI was found to be higher among thefemales (61%, n = 8 out of 13), unemployed (65%), receiving a government pension (63%), with a lower perceivedhealth status (68%), a lower CD4 T cell count (60%), a detectable or unknown viral load (67%), and missed taking theirantiretroviral therapy either in the last week or month (64%). All of the six participants who were Australian Aborigi-nal were food insecure.Conclusions: The study finds evidence of associations between FI, employment, lower immune function and poorerhealth outcomes for people living with HIV in Sydney.

Key words: Australia, food security, health outcomes, HIV.

Introduction

Food security exists: ‘when all people, at all times, havephysical, social and economic access to sufficient, safe andnutritious food, which meets their dietary needs and foodpreferences for an active and healthy life’.1

For people living with HIV (PLHIV) in Australia the cur-rent prevalence of food insecurity (FI) is unknown. It isestimated that there were 25 313 PLHIV in Australia in

2015, however, data on the number of PLHIV in Sydney isnot available.2 In recent times, there has been an increasingnumber of referrals to HIV dietitians in Sydney to assistPLHIV to access healthy food, educate them on how to eathealthily on a budget and where to access food aid andfood distribution sites.

FI of PLHIV in Sydney was studied in 2007 and 2009.In 2007, they found of the 196 PLHIV, the prevalence of FIwith hunger was 21%, FI without hunger was 10% (total31%) and food security was 69%.3 In 2009, of 304 PLHIV53% reported some form of FI. It was also seen thatrespondents on a pension were more likely to have FI thanthose receiving other sources of income.4 The most recentAustralian Health Survey; Nutrition and Physical Activity(2011–2012) reported that around 3.3% of people were liv-ing in a household in New South Wales (NSW) that, in theprevious 12 months, had run out of food and had not beenable to afford to buy more and 1.1% went without foodwhen they could not afford to buy any more.5 Those whoare food insecure have been found to be unemployed6 andfrom marginalised communities such as asylum seekers7

and the Australian Indigenous population.8 Many PLHIV inAustralia are also members of these population groups.

A. Langton, BSc (Nutrition and Dietetics), MSc (Exercise and SportScience), Senior DietitianR. Moxham, MSc (Nutrition and Dietetics), Community DietitianS. Hirst, MSc in Medicine (Infection and Immunity), AdministrationOfficerL. Houtzager, MSc (Nutrition and Dietetics), Manager—NutritionDevelopment DivisionJ. Coutelas, Master Nutrition and Dietetics, Senior DietitianA. Rider, BSc (Nutrition and Food Science, Honours), DietitianC. Jusuf, BSc (Nutrition 1st Class Honours), HIV DietitianCorrespondence: A. Langton, Community Health, 103-105 Redfern StRedfern, Camperdown, New South Wales 2050, Australia. Tel:93950444.Email: angela.langton@health.nsw.gov.au

Accepted April 2018

© 2018 State of New South Wales. Nutrition and Dietetics © 2018 Dietitians Association of Australia468

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 468–473 DOI: 10.1111/1747-0080.12436

Australia is a resource rich setting much like Canada andthe United States of America who have reported high ratesof FI among PLHIV using a validated food security assess-ment tool from the US Department of Agriculture (USDA)that tested for occurrence and frequency of FI.9 In theCanadian studies, the prevalence of FI in two separatecohorts of PLHIV on antiretroviral therapy (ART) was71 and 73%. The factors that were found to be significantlyassociated with FI were younger age, unstable housing,lower incomes, and tobacco and illicit drug use.10,11 In SanFrancisco USA, a cohort study of PLHIV found that overhalf (53.6%) were food insecure. The factors associatedwith increased amounts of FI were being white, recent‘crack’ use, a lack of health insurance, poor physical andmental health and having a low CD4 T cell count.12 Thenegative impact of FI on CD4 T cell count is importantbecause the CD4 T cell count serves as the major laboratoryindicator of immune function in patients who have HIV. Alow CD4 count is a significant predictor for poor treatmentoutcome and prognosis.13,14

PLHIV are faced with challenges such as poor mentalhealth, stigma and discrimination, are at higher risk of co-morbidities such as; cancer, cardiovascular disease, diabetesmellitus and bone disease, and have an increased mortal-ity.15 Some PLHIV living in Sydney, Australia experiencehomelessness, live in boarding houses or crisis accommoda-tion due to social or economic circumstances. This cancompromise their ability to prepare and store food that isin accordance with the Australian Guide to Healthy Eating.

In Sydney, Australia during March 2013, a pilot studydetermined the internal validity, usability and sensitivity ofthe validated USDA six-Item Subset Food Insecurity Toolusing a convenience sample of 20 clients from a communityHIV service. The tool was considered suitable after the pilotstudy as it was validated, uncomplicated and has a low par-ticipation burden, taking approximately 3 minutes to com-plete via telephone call.

The aims of the study were to determine the prevalenceof FI among PLHIV, the geographical location of the areashighest in FI and to identify the factors associated with FI.

Methods

This was a cross-sectional study recruiting PLHIV receivingHIV care from six sites. These six sites were in four healthdistricts covering a large geographical area of Sydney fromthe inner city to the outer suburbs. The sites included twosexual health clinics, one hospital HIV clinic, one HIV out-patient clinic and two community-based multidisciplinaryteams. The time of recruitment varied from each site andtook approximately 3 months. All six sites were recruitedover a 3 year period (2013–2016). The eligibility criteriafor this study were: being older than 18 years of age, diag-nosed HIV positive and receiving HIV care from one of thesix sites. The exclusion criteria consisted of those who hadmoderate to severe level of cognitive impairment, residingin supportive accommodation where meals are provided ornewly diagnosed with HIV in the last 6 months. The

recruitment process varied depending on the type of nutri-tion service. The community dietitians sent out consentforms via mail to clients, they returned them and then weretelephoned. The outpatient clinics had consent forms in thewaiting rooms where clients could fill out and then the die-titian would telephone the client at a different time to com-plete the questionnaire.

The questionnaire was both a food security assessmenttool and a demographic questionnaire. The USDA abbrevi-ated six-item Subset Food Insecurity Tool (Appendix I)was used to assess FI.16 This validated tool classified foodsecurity as: food secure, food insecure without hunger, andfood insecure with hunger. For statistical analysis, bothfood insecure measures were used to define the outcome asbinary indicator (food secure vs food insecure). The demo-graphic questionnaire collected gender, age, postcode, HIVinfection duration (years), CD4 T cell count (cells/uL), HIVviral load, medication adherence, self-rated health status,ethnic background, level of education, income source andemployment status. Variables were grouped into twogroups to complete the significance statistical tests andodds ratio. Employment status was grouped as eitheremployed or unemployed, education level was classified ascompleting final year of secondary school and higher versuslower than completing final year of secondary school andCD4 t cell count was grouped as less than 500 and500 cells/mm3 and above.

The study was approved by the Human Research EthicsCommittees of the Sydney Local Health District, SydneySouth West Health District, South East Sydney Local HealthDistrict and North Sydney Local Health District.

Bivariate analysis was conducted to investigate differ-ences between variables for those who were food secureand food insecure. Analysis of descriptive and frequencystatistics from the demographic questionnaire and theprevalence of FI is shown in summary in Table 1. χ2 testswere used to compare categorical variables, in instanceswhere expected counts were small (<5), Fisher exact testwere used. To assist in developing the 2 x 2 matrix, vari-ables were grouped into two variables. These were foodinsecure versus food secure, employed versus unemployedand a CD4 T cell count of more than or equal to 500 cells/mm3 versus a CD4 T cell count of less than 500 cells/mm3.All tests performed were two sided with a p-value of lessthan 0.05, or 95% confidence interval (CI) and can befound in Table 2. Data were analysed using SPSS version23.0 (IBM Corp, 2013).

Results

Between April 2013 and July 2016, 162 clients completedboth the USDA abbreviated six-item Subset Food InsecurityTool and the demographic questionnaire. Of the 162 clients,47% (n = 76) reported FI. Of those that reported FI, 43%(n = 33) clients were food insecure without hunger and57% (n = 43) clients were food insecure with hunger. Theremaining 53% (n = 86) clients enrolled in this studyreported being food secure.

Food insecurity among people living with HIV

© 2018 State of New South Wales. Nutrition and Dietetics © 2018 Dietitians Association of Australia 469

Table 1 Baseline descriptive characteristics and their association with food security among HIV positive clients

Foodsecure (n = 86)

Foodinsecure (n = 76)

Total(n = 162)

Values are number of participants (%) or median (Q1, Q3)Sociodemographic

Age Years 51 (40, 60) 52 (45, 57) 51 (43, 59)Gender Male 81 94% 68 89% 149 92%

Female 5 6% 8 11% 13 8%Background Australian 48 56% 50 66% 98 60%

Australian Aboriginal/Torres Strait Islander

0 0% 6 8% 6 4%

Pacific Islander 1 1% 0 0% 1 1%Asian 11 13% 6 8% 17 10%European 16 19% 6 8% 22 14%American 2 2% 3 4% 5 3%African 1 1% 2 3% 3 2%Other 2 2% 1 1% 3 2%New Zealand 0 0% 1 1% 1 1%Unknown 5 6% 1 1% 6 4%

Site

Redfern Health Centre 17 20% 19 25% 36 22%HIV Outreach Darlinghurst 11 13% 25 33% 36 22%The Albion Centre 33 38% 17 22% 50 31%Liverpool HIV Team 11 13% 13 17% 24 15%St Vincent’s Network 0 0% 0 0% 0 0%Clinic 16 RNSH 14 16% 2 3% 16 10%

SocioeconomicEmployment Not currently employed 33 38% 61 80% 94 58%

Employed part time 1 1% 4 5% 5 3%Employed as casual/temporary 5 6% 3 4% 8 5%Employed full time 27 31% 1 1% 28 17%Retired 14 16% 3 4% 17 10%Student 3 3% 2 3% 5 3%Other 0 0% 0 0% 0 0%Volunteer 2 2% 2 3% 4 2%Unknown 1 1% 0 0% 1 1%

Income

Salary 6 7% 3 4% 9 6%Pension 37 43% 65 86% 102 63%Superannuation 0 0% 2 3% 2 1%Inheritance 9 10% 0 0% 9 6%No income 29 34% 5 7% 34 21%Other 2 2% 1 1% 3 2%Unknown 3 3% 0 0% 3 2%

Education <final yearsecondary school

18 21% 27 36% 45 28%

≥final year secondary school 68 79% 49 64% 117 72%Clinical

Perceived health Poor 7 8% 10 13% 17 10%Fair 15 17% 36 47% 51 31%Good 37 43% 20 26% 57 35%Excellent 25 29% 10 13% 35 22%Unknown 2 2% 0 0% 2 1%

CD4 (cells/uL) 0–300 10 12% 15 20% 25 15%301–500 12 14% 18 24% 30 19%501–1000 42 49% 28 37% 70 43%1001+ 9 10% 4 5% 13 8%Unknown/not answered 13 15% 11 14% 24 15%

HIV viral load(<20copies/mL)

Undetectable 75 87% 56 74% 131 81%Detectable 5 6% 10 13% 15 9%

A. Langton et al.

470 © 2018 State of New South Wales. Nutrition and Dietetics © 2018 Dietitians Association of Australia

Table 1 shows the prevalence of client characteristicssubdivided by food security status. Overall, clients whoidentified as being food insecure were likely to be female,Australian Aboriginal, reside in the inner city, be unem-ployed, receiving a government pension, have a lower per-ceived health status, a lower CD4 T cell count, a detectableor unknown viral load and missed taking their ART’s eitherin the last week or month.

Table 2 shows that employment status and CD4 countwere significantly associated with FI among PLHIV.

Figure 1 shows the highest FI areas in the Sydney Metro-politan area from the five sites. The local health districtswhich are closest to the Sydney Central Business District,South Eastern Sydney and Sydney Health District had thegreatest prevalence of FI South West Sydney, North Sydneyand ‘other’ health districts reported less incidence of FIamong the study participants.

Discussion

FI has been identified as an increasing issue in Sydney pre-viously and this study provided up to date information on

the current issue. The findings in this study were higherthan in 2007 and marginally lower than in 2009, being31% in 2007, 53% in 2009 and 47% in 2016.3,4 It was notunexpected to find that PLHIV in Sydney who were foodinsecure had higher rates of unemployment and perceivedtheir health as poor or fair. These findings are consistentwith similar studies which reported a relationship betweenFI unemployment and poor health for PLHIV.10–12

From 1997 to 2013 the Australian CommonwealthDepartment of Health and Ageing conducted six compre-hensive surveys on PLHIV in Australia.17 The latest pub-lished results from 2013 support our findings that PLHIVhave low rates of being in full time employment, often dueto the impact of HIV disease management rather than will-ingness or skill, and high rates of receiving a social securityincome. These factors result in many PLHIV living belowthe poverty line and facing a regular challenge to affordfood or to choose between food and HIV medication.17

A relationship between FI and a low CD4 T cell count wasobserved but a conclusive assessment cannot be made basedon the results. The literature to date has mixed findings onthis relationship. In a study conducted in the US on

Table 1 Continued

Foodsecure (n = 86)

Foodinsecure (n = 76)

Total(n = 162)

Unknown/not answered 6 7% 10 13% 16 10%Medications

Ever missed ARV yes 28 33% 37 49% 65 40%no 57 66% 38 50% 95 59%N/A 1 1% 1 1% 2 1%

How manytimes missedmedication inpast week?

0 20 23% 17 22% 37 23%1 4 5% 7 9% 11 7%2 3 3% 5 7% 8 5%3 1 1% 2 3% 3 2%>3 0 0% 4 5% 4 2%N/A 57 66% 37 49% 94 58%Unknown/not answered 1 1% 4 5% 5 3%

How manytimes missedmedication inpast month

0 19 22% 16 21% 35 22%1–4 8 9% 7 9% 15 9%5–8 0 0% 3 4% 3 2%9+ 1 1% 4 5% 5 3%4 0 0% 2 3% 2 1%N/A 57 66% 37 49% 94 58%Unknown/not answered 1 1% 7 9% 8 5%

ARV, antiretroviral therapy; N/A, not applicable; RNSH, Royal North Shore Hospital.

Table 2 Factors associated with food insecurity among HIV positive clients

Foodsecure (n = 86)

Foodinsecure (n = 76)

Total(n = 162)

OR(95% CI) P value

Employment Employed 33 8 41 5.39 0.00Unemployed 52 68 120 (2.30–12.65)

CD4(cells/uL) Greater than 500 51 32 83 2.04 0.026Lower than 500 35 44 79 (0.93–3.70)

Values are number of participants (%) or median (Q1, Q3).

Food insecurity among people living with HIV

© 2018 State of New South Wales. Nutrition and Dietetics © 2018 Dietitians Association of Australia 471

592 PLHIV who had completed ≥4 food security assessmentsover a 10-year period, increases in CD4 counts were on aver-age 99.5 cells less for individuals with a least one episode ofFI, compared with those who were consistently foodsecure.17 Another study found that FI was not independentlyassociated with low CD4 counts, but more likely afteradjusted analysis to be associated with an unsuppressed HIVviral load. Viral load is used to assess progression of theinfection and is an important indicator of antiretroviral treat-ment. One explanation that was provided was that in thesestudies that have shown a relationship with CD4 and FI par-ticipants may or may not have been on ART whereas in thisstudy all participants were on ART.17,18 The increase in CD4count could have been due to marked improvements in thevirological response because taking ART, and not necessarilylinked to being food secure or not. Overall, the participantsreported a fair compliance with ART medication in bothgroups (FI = 50%, food secure = 66%) and there was a dif-ference found between the groups (FI = 49% vs foodsecure = 33%) when asked if they had ever missed their ARTmedication.

FI and the geographical areas in which it appeared to bepredominant in the Sydney region are the inner city loca-tions which are known to be well populated by gay andsingle Caucasian males. These individuals may be morelikely isolated because of an absence of family, sexual dis-crimination and/or HIV stigmatisation.19 The geographicalarea that had the least incidence of FI was the north shoreof Sydney. The outer western suburbs, known for its cul-turally and linguistically diverse (CALD) and emerging cul-tures, also showed less incidence of FI. This result does notreflect the current literature that associates CALD as a highFI risk factor.20 This is possible, as there were only a mini-mal number of CALD HIV positive people recruited intothe study due to language barriers in the absence of quali-fied registered translators. This is a geographical area whichrequires further research.

This study had limitations; due to the nature of theHIV nutrition clinics and how the dietitians operated andthe recruitment methods were different, which couldhave affected numbers recruited at some sites. OnlyPLHIV who attended public health clinics were recruitedand those receiving their care from private medical ser-vices were not. According to ethics approval the ques-tionnaire could only be administered by telephone,which prevented the questionnaire from being completedface to face in the clinics which may have influencedrecruitment.

The questionnaire was only validated for Englishspeakers and language was not an assessment item in thedemographics. Both these factors would have affected thenumber of CALD clients recruited and this may explainwhy we had higher prevalence of food security among thisgroup compared to other research.

Other limitations of the study were that illicit drug usewas not assessed. In Sydney Australia, there is a high use ofthe illicit drugs in this sample HIV population which maystill contribute to FI which could have confounded theresults.21 Finally, even though females and AustralianAboriginal people were more likely to be FI the recruitmentnumbers were too small for analysis and comment.

Over the long recruitment period there has been a rise infood costs, drug use and restrictions to eligibility to the Dis-ability Support Pension that may have affected FI that werenot considered in this study.

This local research has identified FI as an issue forPLHIV in Sydney, Australia. It hopes to educate thoseworking in the HIV field to be more aware of this issuewhen seeing clients such as: medical officers when consid-ering an ART regime; other health professionals beingaware of free or inexpensive food services they can referclients to or to ask questions around this issue whenspeaking to clients; the non-government sector who workprimarily with advocacy and education to add FI to theagenda and hope to increase funding and partnershipswith cheap food outlets. This research has highlighted theneed for further research of FI in other clinical areas andpopulations and its implications on health and lifestyleoutcomes.

In conclusion, FI is highly prevalent among PLHIV inSydney who are female, Australian Aboriginal, unemployedand receiving a government pension. This study identifiesthose who are at risk of FI and suggests suitable serviceinterventions which aim to minimise the burden of the dis-ease trajectory by improving food access among this popu-lation group.

Funding source

No funding was provided for this research project or forthe contents of this manuscript.

Conflict of interest

The authors have no conflicts of interest to declare.

34%

14%

36%

13%

4%

Figure 1 Food insecurity according to local health districts.( ) SLHD, Sydney Local Health District; ( ) SWSLHD, SouthWest Sydney Local Health District; ( ) SESLHD, South EastSydney Local Health District; ( ) NSLHD, North SydneyLocal Health District; ( ) Other, outside the districts.

A. Langton et al.

472 © 2018 State of New South Wales. Nutrition and Dietetics © 2018 Dietitians Association of Australia

Authorship

AL developed the initial study concept and design; AL, RM,LH, JC, AR, CY carried out acquisition and analysis of thedata; all authors critically revised draft versions of themanuscript.

References

1 Food and Agriculture Organisation of the United States. FoodInsecurity Statistics. (Available from: http://www.fao.org/economic/ess/ess-fs/en/, accessed 20 March 2017).

2 The Kirby Institute. HIV, Viral Hepatitis and Sexually Transmis-sible Infections in Australia Annual Surveillance Report 2016. TheKirby Institute. Sydney, Australia: UNSW Sydney.

3 Aibibula W, Cox J, Hamelin AM, McLinden T, Klein MB,Brassard P. Association between food insecurity and HIV viralsuppression: a systematic review and meta-analysis. AIDS Behav2017; 21: 754–65.

4 Houtzager L, Purnomo J, Ackerie A et al. Contributing factorsto food insecurity in PLHIV in Sydney. Nutr Diet 2009; 66: A7.

5 Australian Bureau of Statistics. Australian Health Survey, Nutri-tion and Physical Activity, 2011–12. Canberra, Australian:Australian Bureau of Statistics, 2012.

6 Burns C. A Review of the Literature Describing the Link betweenPoverty, Food Insecurity, and Obesity with Specific Reference toAustralia. VicHealth: Melbourne, Australian, 2004.

7 Gallegos D, Ellies P, Wright J. Still there’s no food: food inse-curity in refugee populations in Perth, Western Australia. NutrDiet 2008; 65: 78–83.

8 Australian Bureau of Statistics. Australian Aboriginal and TorresStrait Islander Health Survey: Nutrition Results—Food and Nutri-ents, 2012–13. Canberra, Australian: Australian Bureau of Sta-tistics, 2015.

9 Anema A, Fielden SJ, Shurgold S et al. Association betweenfood insecurity and procurement methods among people livingwith HIV in a high resource setting. PLoS One 2016; 11:e0157630. https://doi.org/10.1371/journal.pone.0157630

10 Cox J, Hamelin AM, McLinden TM et al. Food insecurity inHIV—hepatitis C co-infected individuals in Canada: the impor-tance of co-morbidities. AIDS Behav 2017; 21: 792–802.

11 Vogenthaler S, Kushel B, Hadley C et al. Food insecurityamong homeless and marginally housed individuals living withHIV/AIDS in San Francisco. AIDS Behav 2009; 13: 841–8.

12 Mellors JW, Munoz A, Giorgi JV et al. Plasma viral load andCD4+ lymphocytes as prognostic markers of HIV-1 infection.Ann Intern Med 1997; 126: 946–54.

13 Egger M, May M, Chêne G et al. Prognosis of HIV-1-infectedpatients starting highly active antiretroviral therapy: a collabo-rative analysis of prospective studies. Lancet 2002; 360:119–29.

14 Rodriguez-Penney AT, Ludicello JE, Riggs PK et al. HIV Neuro-behavioral Research Program HNRP Group. Co-morbidities inpersons infected with HIV: increased burden with older ageand negative effects on health-related quality of life. AIDSPatient Care STDS 2013; 27: 5–16.

15 Weiss J, Osorio G, Ryan E et al. Prevalence and patient aware-ness of medical co-morbidities in an urban AIDS clinic. AIDSPatient Care STDS 2010; 24: 39–48.

16 United States Department of Agriculture (USDA). “Definitionsof Food Security.” United States Department of Agriculture,2014. (Available from: http://www.ers.usda.gov/topics/food-nutrition-assistance/food-security-in-the-us/definitions-of-food-security.aspx, accessed 22 May 2014).

17 McMahon J, Wanke C, Elliot J et al. Repeated assessments offood security predict CD4 change in the setting of antiretroviraltherapy. J Acquir Immune Defic Syndr 2011; 58: 60–3.

18 Wang E, McGinnis K, Fiellin A et al. Food insecurity is associ-ated with poor virologic response among HIV-infected patientsreceiving antiretroviral medications. J Gen Intern Med 2011; 26:1012–8.

19 Ackerie A, Houtzager L, George C et al. Food Security of peo-ple living with HIV in the Sydney Metropolitan area, Australia,Sydney IAS 2007 Abstract no MOPEBO94.

20 Rosier K. Food Insecurity in Australia What is it, Who Experiencesit and How Can Child and Family Services Support FamiliesExperiencing it? Australian Institute of Family Studies. AustralianGovernment, Victoria, Australia, 2011.

21 Grieson R, Pitts M, Koelmeyer R. HIV Futures Seven: TheHealth and Wellbeing of HIV Positive People in Australia, Mono-graph Series Number 88. Melbourne, Australia: The AustralianResearch Centre in Sex, Health and Society, La Trobe Univer-sity, 2013.

Food insecurity among people living with HIV

© 2018 State of New South Wales. Nutrition and Dietetics © 2018 Dietitians Association of Australia 473

ORIGINAL RESEARCH

Cross-sectional analysis of unhealthy foods, race/ethnicity, sex and cardiometabolic risk factors inU.S. adults

Joan A. VACCARO , Gustavo G. ZARINI and Fatma G. HUFFMANDepartment of Dietetics and Nutrition, Florida International University, Miami, Florida, USA

AbstractAim: The purpose of the present study was to determine the association of unhealthy dietary food items with cardio-metabolic risk factors with and without sociodemographic factors.Methods: This cross-sectional study used data available to the public from the National Health and Nutrition Survey(NHANES) 2009–2010 where unhealthy food consumption was based on responses to the Dietary Screener Question-naire (unique to this NHANES cycle), and cardiometabolic risk factors were based on laboratory results, anthropomet-ric measures, interview and examination questions for 2045 adults aged 20–69 and belonging to four racial/ethnicgroups: 473 Mexican Americans (MA); 267 Other Hispanics (OH); 389, non-Hispanic Blacks (NHB) and 916 non-Hispanic Whites (NHW) (characterised by NHANES).Results: A higher percent of MA, followed by OH and NHB, consumed soft drinks as compared to NHW. Consump-tion of fried potatoes was over 75% across groups and was associated with higher odds dyslipidaemia (high non-HDLcholesterol) in the reduced model: OR = 1.38 (1.10, 1.73), P = 0.009 and full model: OR = 1.50 (1.15, 1.96),P = 0.005. All unhealthy foods measured were consumed more often by males as compared to females.Conclusions: Dyslipidaemia was associated with fried potato consumption and marginally with processed meats.Dietary interventions, tailored to specific populations, are needed to determine if substituting healthy foods in placeof unhealthy ones will improve cardiometabolic outcomes.

Key words: cardiovascular disease, hypertension, non-HDL cholesterol, race/ethnicity, waist-to-height ratio.

Introduction

Cardiometabolic risk is a condition with high likelihood ofdeveloping atherosclerotic cardiovascular disease and diabe-tes.1 Dyslipidaemia, central obesity and high blood pressureare central cardiometabolic risk factors and can be con-trolled largely by diet. Changes in dietary behaviour areessential to counteract cardiometabolic dysfunction.2 Spe-cific foods, rather than nutrients, are determinants of cardi-ometabolic risk.3–5 Diets high in seeds, nuts, fruit and fishhave been found to lower blood pressure, preventing orcontrolling hypertension (HTN).2,6

Conversely, unhealthy foods (high calorie with lownutritional benefit, containing excess sodium, sugar or satu-rated fat) have been associated with cardiometabolic

dysfunction. Adults eating at fast food restaurants, wherefried potatoes and processed meats are prominently avail-able, had poorer cardiometabolic outcomes than their coun-terparts eating at sit-down restaurants for a cohort of Blackand White adults from the Coronary Artery Developmentin Young Adult Study.7 Diets high in sodium, indicative ofhigh processed food intake, and low in fruits and vegeta-bles, high systolic blood pressure and high body massindex, were among the main contributing factors to mortal-ity in a cohort of Brazilian adults.8 Processed red meat has ahigher cardiometabolic risk as compared to unprocessedred meat primarily due to the high sodium content.9 Meta-analysis of prospective cohort studies indicated that sugar-sweetened beverages increase the risk of type 2 diabetes.10

The proposed mechanism of increased diabetes risk forconsumption of sugar-sweetened beverages is an increase inbody weight which leads to cardiometabolic dysfunction;however, further studies are needed to confirm themechanism.11

Consumption of unhealthy foods may differ by sex andrace/ethnicity. Diet contributes to health disparities formany cardiometabolic conditions in the U.S. populations ofBlacks, Asians and Hispanics.12 Health-related dietary dis-parities may also be due to low socioeconomic status and

J.A. Vaccaro, PhD, DPD, Adjunct ProfessorG.G. Zarini, PhD, RD, Research ScientistF.G. Huffman, PhD, RD, ProfessorCorrespondence: J.A. Vaccaro, Department of Dietetics and Nutrition,Florida International University, Robert R. Stempel College of PublicHealth and Social Work, 11200 SW, 8th ST, AHC-5 Room324, Miami, FL 33199, USA.Email: jvaccaro@fiu.edu.

Accepted May 2018

© 2018 Dietitians Association of Australia474

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 474–480 DOI: 10.1111/1747-0080.12439

often present as diets high in saturated fat and salt, whilelow in fruits, vegetables and whole grains.12 Higher fre-quency of consumption of fast food (which is notably highin saturated fat and salt) was found for Blacks as comparedto Whites, males and individuals with the lowest educationlevel.13

Increasing fruits and vegetables may not negate the cardi-ometabolic damage from unhealthy foods.14 Given the obe-sity epidemic, designing interventions to improvecardiometabolic health should consider substitution ofunhealthy for healthier foods as opposed to adding foodsthat increase total calories. The aim of the present studywas to determine the association of unhealthy food itemsand cardiometabolic risk factors with and without sociode-mographic factors, adjusting for age and tobacco use. It washypothesised that (i) inadequate fibre intake (<14 g/1000kcal) was expected to be present in approximately 90% ofthe adult population15 and (ii) odds of having unhealthydietary items will be higher for those persons with a cardio-metabolic risk factor (dependent variable for each model:high waist-to-height ratio (WHtR) (a measure of central,visceral obesity); high non-HDL cholesterol (non-HDL-C)(a measure of dyslipidaemia) and HTN in the reducedmodel (adjusting for age and tobacco use) and in the fullmodel (adjusting for race/ethnicity, sex, education level, ageand tobacco use).

Methods

This cross-sectional study was reported in accordance withthe STrengthening the Reporting of OBservational studiesin Epidemiology statement16 and uses data acquired andgenerated from the National Health and Nutrition Examina-tion Survey (NHANES) 2009–2010. This cycle included aunique addition of a dietary screening questionnaire withconsumption of unhealthy foods. NHANES is a program ofstudies designed to assess the health and nutritional statusof adults and children in the United States by combininginterviews and physical examinations. NHANES uses acomplex, multistage probability sample design to obtain arepresentative sample of the civilian non-institutionalisedpopulation of the United States. Details about the surveysare available at the website.17 For respondants aged12–69 years, the dietary screening questions were asked atthe mobile examination centre (MEC) during the MECinterview, by trained interviewers using the computer-assisted personal interviewing with periodic quality controlduring the data collection.18 The participants in the presentstudy are from the fasted subsample of the MEC andinclude 2045 adults aged 20–69 years and belonging tofour racial/ethnic groups classified by NHANES: 473 Mexi-can Americans (MA); 267 Other Hispanics (OH); 389 non-Hispanic Blacks (NHB) and 916 non-Hispanic Whites(NHW). The fifth racial/ethnic category (Others) had insuf-ficient sample size and was filtered out. The sample weightfor the MEC fasted subsample was applied in the analysisplan to reduce bias. The data used for the present studywere publicly available. Prior to public release, the study

protocol (continuation of Protocol No. 2005–06) wasapproved by the NCHS Research Ethics Review Board.19

Separate informed consent forms were signed by partici-pants for the interview and health examination. Participantsin the present study read, understood and signed informedconsent forms for the interview and health examination.

Unhealthy food consumption was based on responses tothe Dietary Screener Questionnaire, and cardiometabolicrisk factors were based on the laboratory results, anthropo-metric measures, interview and examination questions. TheDietary Screener Questionnaire (DSQ) was developed incollaboration with the Risk Factor Monitoring and MethodsBranch of the National Cancer Institute, National Institutesof Health for inclusion in the NHANES 2009–2010 sur-vey.17 DSQ is composed of 26 questions asking about thefrequency of consumption of selected foods and drinks inthe past month to capture intakes of fruits and vegetables,dairy/calcium, whole grains/fibre, added sugars, red meatand processed meat.20 The following unhealthy dietaryitems included in the present study: soft drinks; sports/energy drinks; fried potatoes and processed meats were val-idated against multiple 24-hour recalls. Responses werecoded ‘yes’ for any frequency and ‘no’ for stating zero timesin the past month. Soft drinks were considered affirmativeby a yes response to drink regular soda or pop that containssugar over the past month and did not include diet soda.Sports/energy drinks were among sweetened fruit drinks,sports or energy drinks such as Kool-Aid, lemonade, Hi-C,cranberry drink, Gatorade, Red Bull or Vitamin Water. Fruitjuices made at home with added sugar were included, butdiet drinks or artificially sweetened drinks were notincluded. Consuming fried potatoes was affirmative if avalue other than ‘no’ was given for the past month and con-sidered eating any kind of fried potatoes, including Frenchfries, home fries or hash brown potatoes. An affirmativeresponse for processed meats was any value over the pastmonth other than ‘no’ by the aid of a picture card and con-sidered bacon, lunch meats or hot dogs. Inadequate fibreintake was considered below 14 g/1000 kcal21 and calcu-lated using grams of fibre and energy, provided as variablesby NHANES, from the first 24-hour dietary recall. Solubleand insoluble dietary fibres are non-digestible carbohy-drates and lignin that are fermented by bacteria in the colonand are protective against cardiometabolic diseases.22

Cardiometabolic variables (laboratory and anthropometricmeasures) used in the present study included measurementsof inadequate cholesterol and body weight. Non-HDL-C wascalculated by subtracting HDL from total cholesterol. Ade-quate cholesterol was considered Non-HDL-C of <130 mg/dLbased on the American Association of Clinical Endocrinolo-gists’ (AACE) guideline for moderate cardiovascular risk inpersons with type 2 diabetes.23 WHtR was used as a proxymeasurement for visceral obesity.24,25 Waist circumference(cm) was divided by height (cm). Based on the work of Ash-well24 and Swainson et al.,25 a cutoff point of 0.59 was usedfor visceral obesity. HTN was considered if measured first-reading of systolic blood pressure was 130 mmHg or greater,or diastolic blood pressure was 80 mmHg or greater, or the

Unhealthy food and cardiometabolic risk

© 2018 Dietitians Association of Australia 475

individual reported taking medication for blood pressure.Tobacco use was considered affirmative by answering yes tohaving used tobacco or nicotine within the past five days.

Quantitative variables constructed for sociodemographicsincluded race/ethnicity (discussed in the participant andsample number section of methods), age and education.Age groups were formed as follows: 20–24; 35–49 and50–69 years for participant characteristics; however, it wasleft as a continuous variable for the models. Education levelwas collapsed from four to five categories by combining lessthan 9th grade with 9th–11th grade without a high-schooldiploma or general equivalency diploma (GED). The neweducation categories were as follows: less than a high-school degree; high-school degree or GED; some college orassociate degree; college degree or above.

All data were analysed with the Statistical Package forSocial Sciences (SPSS, version 24, IBM, New York, USA)with the module for complex sample design. The sampleweights used were for the fasted subsample of the MEC. Allanalysis considered the differential probabilities of selectionand the complex sample design, with SPSS, using theTaylor series linearisation. A P-value of less than 0.05 (two-sided) was considered statistically significant. Listwise dele-tion was applied to missing data for the unweighted samplebecause less than 2% of the cases were missing at random.Participants’ characteristics were presented as percent, 95%CIs and performed with cross-tabulations by race/ethnicity,sex and age-group. Reduced and full logistic regressionmodels were conducted for unhealthy foods and sociode-mographics for major cardiometabolic risk factors: highnon-HDL-C, high WHtR and HTN.

Results

The present study used data from applying the inclusioncriteria for participants with fasting laboratory samples.Details of how this subsample was determined are availablefrom the NHANES website.17 The characteristics of thestudy population by sex and race/ethnicity are presented inTables 1 and 2, respectively. Males have a higher percentreporting consumption of all unhealthy foods measured:soft drinks sports/energy drinks, fried potatoes, and pro-cessed meats and higher cardiometabolic risk factors ascompared to females. MA (80%), followed by NHB (74%)

and OH (68%), had a higher percent reported consumptionof soft drinks as compared to NHW (55%). Sports/energydrinks were reported consumed more often by NHB (67%),MA (63%) and OH (61%), as compared to NHW. The con-sumption of fried potatoes was the highest for NHW ascompared to other groups. Processed meat intake wasreported over 80% of participants for all groups; however,OH had the lowest percent of the groups. MA had the low-est percent of inadequate fibre intake as compared to othergroups; however, all groups were had over 80% of inade-quate fibre intake. Of the cardiometabolic risk factors, HTNwas significantly higher in men as compared to women andin NHB as compared to other groups and MA and OH hadsignificantly lower tobacco use as compared to NHBand NHW.

There were significant differences in the consumption ofunhealthy foods across age-groups (data not shown).Approximately, 75% of 20–34 year olds, 60% of35–49 year olds and 50% of 50–69 year olds reported con-suming soft drinks. The trend was similar for sports/energydrink consumption with approximately 70% of 20–34 yearolds; 50% of 35–49 year olds and 35% of 50–69 year olds(P < 0.001). Higher percent from the oldest group, 22.8(20.1, 25.6) as compared to the youngest group, 11.7 (7.9,17.0) reported ‘no’ for consuming fried potatoes(P = 0.001). The middle group had the lowest percent: 9.7(7.9, 13.1), followed by the youngest group: 11.2 (8.3,15.0), as compared to the oldest group 15.5 (12.0, 19.7)for reporting no consumption of processed meats(P = 0.013). All groups reported over three-quarters eatingfried potatoes and over 80% reporting eating processedmeats. Higher percent of fibre consumption per energyintake was calculated for 50–69 year olds, (12%), as com-pared to 20–14 year olds (7%) (P = 0.046); albeit, allgroups had over 85% of inadequate fibre intake.

The hypotheses were partially supported. Inadequatefibre intake was present in over 90% of the adult popula-tion (Table 1). There was some association with unhealthyfoods and cardiometabolic dysfunction (Table 3). Modelsfor unhealthy diet items and cardiometabolic risk factorsare shown in Table 3. Fried potato consumption was asso-ciated with high non-HDL-C in the reduced and full model.Unhealthy foods were not associated with WHtR or HTN.Higher odds of WHtR, an indicator of visceral fat, were

Table 1 95% CI unhealthy foods and cardiometabolic risk factors by sex

Parameter Males Females Total P-value

Soft drinks (yes) 67.3 (63.3, 71.0) 53.5 (49.8, 57.1) 60.5 (57.4, 63.5) <0.001Sports/energy drinks (yes) 58.2 (54.3, 61.9) 45.0 (41.7, 48.4) 51.7 (48.6, 54.7) <0.001Fried potatoes (yes) 85.0 (81.3, 88.2 80.3 (77.2, 83.1) 82.7 (80.0, 85.1) 0.025Processed meats (yes) 89.7 (86.7, 92.1) 85.9 (82.7, 88.6) 87.8 (85.4, 89.9) 0.032Dietary fibre (<14 g/1000 kcal) 92.3 (89.1, 94.7) 88.8 (86.9, 90.5) 90.6 (88.6, 92.2) 0.032Non-HDL-C (>130 mg/dL) 63.0 (59.2, 66.7) 55.1 (49.6, 60.4) 59.0 (55.5, 62.4) 0.016WHtR (>0.59) 38.8 (33.7, 44.1) 47.2 (44.5, 49.9) 43.0 (39.8, 46.3) 0.004Hypertension (≥130/80 mmHg) 45.5 (37.5, 53.8) 35.7 (30.8, 41.0) 40.6 (35.0, 46.5) 0.003Tobacco/nicotine last five days 30.6 (25.2, 36.4) 23.1 (20.4, 26.1) 26.9 (24.2, 29.7) 0.034

J. A. Vaccaro et al.

476 © 2018 Dietitians Association of Australia

found for females. Lower education was associated withhigh WHtR and HTN. Non-Hispanic Blacks had higherodds of HTN, adjusting for diet, education, tobacco useand age.

Discussion

Cardiometabolic outcomes and reporting any consumptionof unhealthy foods varied by race/ethnicity, sex and age-group in the present study. Dyslipidaemia, as measured byhigh non-HDL-C, was marginally higher in Hispanics(MA and OH) as compared to NHW. Tobacco use andHTN were significantly higher in NHB compared to othergroups. All measured cardiometabolic risk factors: highWHtR, high, non-HDL-C and HTN were determinants ofcardiometabolic mortality for person with low education.26

Health behaviour intervention designs should consider edu-cation level, cost of and access to healthy foods, as well asrace/ethnicity, sex and age.

The present study was the first to examine the inclusionof specific unhealthy foods with cardiometabolic risk factorsconsidering race/ethnicity, sex and age-groups inU.S. adults. Unhealthy food consumption varied by race/ethnicity, sex and age-group. Greater percent of minorities(MA, OH, NHB) reported consuming sugary beverages: softdrinks (>66%) and sports/energy drinks (>60%) as com-pared to NHW (55%, 46%, respectively). Our findingsagreed with Han and Powell27 who found that, amongU.S. children and adults, racial minorities, persons of loweducation and income were more likely to drink sugar-sweetened beverages as compared to NHW. Implications ofthese finding can be used in forming dietary interventionstargeting specific groups. The current findings that approxi-mately 90% of U.S. adults have inadequate fibre intake cor-responded to those of Clemens et al.;15 however, we foundMA consumed higher fibre compared to other races/ethnici-ties. As soft drink intake was also highest for MA as com-pared to other groups, substitution of already consumedproduce, high in fibre and water for soft drinks may be fea-sible. While processed meats, such as cold-cuts, hot dogsand bacon, were consumed more often by NHW, effortsshould be made to make lean meat a convenient andappealing option for this population.

Over three-quarters of the population reported consum-ing fried potatoes which was associated with dyslipidaemiaas measured by high non-HDL-C (>130 mg/dL). A reviewof fast food consumption and cardiometabolic risk indi-cated that high frequency of fast food consumption wasassociated with dyslipidaemia.28 Their results corroboratewith our findings as fast foods generally contain processedmeats and fried potatoes. Acrylamide, a carcinogen, formswith fried potatoes, because of the availability of carbohy-drate, oil and the high temperature.29 Fried potato con-sumption, when classified with all fried foods, wasassociated with greater cardiometabolic risk.30

Over 80% of the population reported consuming andprocessed meats which was marginally associated with highHDL-C. Inflammatory markers that contribute toT

able

295

%CIun

healthyfood

sandcardiometabolicrisk

byrace/ethnicity

Parameter

MA

OH

NHB

NHW

Total

P-value

Softdrinks

(yes)

80.2

(75.4,

84.3)

68.4

(60.2,

75.6)

73.6

(68.6,

78.1)

54.8

(51.0,

58.5)

60.5

(57.4,

63.5)

<0.001

Sport/energy

drinks

(yes)

63.4

(59.1,

67.6)

61.3

(55.7,

66.6)

67.5

(61.1,

72.8)

46.5

(42.5,

50.6)

51.7

(48.6,

54.7)

<0.001

Friedpo

tatoes

(yes)

77.6

(72.3,

82.1)

69.0

(61.7,

75.5)

78.5

(73.4,

82.9)

85.3

(82.2,

87.9)

82.7

(80.0,

85.1)

<0.001

Processedmeats(yes)

84.2

(78.7,

88.5)

80.5

(74.0,

85.6)

85.0

(79.1,

89.4)

89.4

(86.5,

91.7)

87.8

(85.4,

89.9)

0.00

6Fibre(<14

g/10

00kcal)

84.4

(79.1,

88.5)

91.0

(86.9,

93.9)

94.0

(89.3,

96.7)

90.8

(88.1,

92.9)

90.6

(88.6,

92.2)

0.01

4Non

-HDL-C(>13

0mg/dL

)64

.7(58.6,

70.4)

65.7

(60.1,

70.9)

55.6

(49.8,

61,3)

58.2

(53.7,

62.6)

59.0

(55.5,

62.4)

0.05

0WHtR

(>0.59

)49

.2(45.5,

52.8)

40.1

(35.4,

45.1)

47.0

(40.3,

53.9)

41.7

(37.2,

46.3)

43.0

(39.8,

46.3)

0.06

9Hypertension(≥13

0/80

mmHg)

33.8

(27.4,

40.7)

32.5

(23.4,

43.2)

56.1

(50.3,

61.8)

39.7

(32.1,

47.8)

40.6

(35.0,

46.5)

0.00

2Tob

acco/nicotinelast5days

19.5

(15.8,

23.8)

19.4

(13.8,

26.7)

32.8

(27.0,

38.2)

27.5

(23.9,

31.3)

26.9

(24.2,

29.7)

0.01

1

MA,Mexican

Americans;NHB,

non-HispanicBlacks;NHW,no

n-HispanicWhites;no

n-HDL-C,no

n-HDLcholesterol(sum

ofallothercompo

nentsof

serum

cholesterol);OH,Other

Hispanics;

WTtR,waist-to-height

ratio

.

Unhealthy food and cardiometabolic risk

© 2018 Dietitians Association of Australia 477

Table 3 Logistic regression models for unhealthy foods and cardiometabolic risk factors

Parameters Reduced model Full model

OR (95% CI) P-value OR (95% CI) P-value

Model 1: High non-HDL-CSoft drinks (yes) 2.36 (0.86, 1.83) 0.219 1.10 (0.76, 1.59) 0.609Sports/energy drinks (yes) 1.04 (0.90, 1.20) 0.566 0.97 (0.83, 1.14) 0.737Fried potatoes (yes) 1.38 (1.10, 1.73) 0.009 1.50 (1.15, 1.96) 0.005Processed meats (yes) 1.49 (0.99, 2.26) 0.057 1.48 (0.97, 2.27) 0.069Age (years) 1.03 (1.02, 1.04) <0.001 1.03 (1.02, 1.04) <0.001Tobacco use (yes) 0.96 (0.77, 1.21) 0.725 0.89 (0.69, 1.16) 0.374Race/ethnicity — 0.063MA 1.48 (0.99, 2.20) ns 0.054OH 1.56 (1.09, 2.24) ns 0.018NHB 1.00 (0.75. 1.35) ns 0.981NHW (reference) 1.00

Sex (male) 1.45 (1.09, 1.94) 0.015Education — 0.151<high school 1.55 (1.01, 2.39) ns 0.046High school diploma or GED 1.28 (0.79, 2.06) ns 0.292Some college or AA degree 1.12 (0.77, 2.62) ns 0.544College degree or more (reference) 1.00

Model 2: High WHtRSoft drinks (yes) 1.26 (0.97, 1.63) 0.079 1.15 (0.86, 1.54) 0.316Sports/energy drinks (yes) 1.03 (0.74, 1.42) 0.859 1.02 (0.72, 1.45) 0.898Fried potatoes (yes) 1.06 (0.76, 1.48) 0.725 1.09 (0.78, 1.53) 0.576Processed meats (yes) 1.24 (0.79, 1.97) 0.375 1.24 (0.78, 1.95) 0.337Age (years) 1.03 (1.03, 1.04) <0.001 1.03 (1.02, 1.04) <0.001Tobacco use (yes) 0.86 (0.62, 1.21) 0.360 0.72 (0.50, 1.05) 0.081Race/ethnicity — 0.057MA 1.11 (0.82, 1.49) 0.471OH 0.86 (0.64, 1.15) 0.284NHB 1.08 (0.78, 1.49) 0.633NHW (reference) 1.00 —

Sex (male) 0.74 (0.61, 0.91) 0.006Education — 0.001<High school 2.90 (1.91, 4.41) <0.001High school diploma or GED 2.29 (1.55, 3.39) <0.001Some college or AA degree 2.54 (1.73, 3.73) <0.001College degree or more (reference) 1.00 —

Model 3: HypertensionSoft drinks (yes) 1.20 (0.97, 1.48) 0.092 0.97 (0.76, 1.24) 0.804Sports/energy drinks (yes) 1.26 (0.84, 1.88) 0.253 1.10 (0.71, 1.70) 0.652Fried potatoes (yes) 0.98 (0.70, 1.37) 0.890 0.98 (0.72, 1.70) 0.984Processed meats (yes) 1.11 (0.82, 1.51) 0.890 1.07 (0.78, 1.47) 0.647Age (years) 1.07 (1.06, 1.08) <0.001 1.07 (1.06, 1.08) <0.001Tobacco use (yes) 0.93 (0.63, 1.38) 0.704 0.74 (0.51, 1.08) 0.113Race/ethnicity — 0.003MA 0.89 (0.57, 1.38) 0.573OH 0.68 (0.40, 1.83) 0.676NHB 2.34 (1.35 4.04) 0.005NHW (reference) 1.00 —

Sex (male) 1.91 (1.28, 2.85) 0.004Education — 0.036<High school 1.97 (1.26, 3.08) 0.006High school diploma or GED 2.02 (1.24, 3.30) 0.007Some college or AA degree 1.83 (1.15, 2.90) 0.014College degree or more (reference) 1.00 —

GED, general equivalency diploma; MA, Mexican American; NHB, non-Hispanic Blacks; NHW, non-Hispanic Whites; ns, notsignificant (as the P value of the variables race/ethnicity and education were P > 0.05); non-HDL-C, non-HDL cholesterol (sum of all othercomponents of serum cholesterol); OH, Other Hispanic; WHtR, waist-to-height ratio.

J. A. Vaccaro et al.

478 © 2018 Dietitians Association of Australia

cardiometabolic disorders, C-reactive protein, interlukin-6and homocysteine were higher for persons who ate pro-cessed meats from a large, multi-ethnic population study.31

Higher homocysteine and haemoglobin A1c were associatedwith higher processed meat consumption in Britishadults.32

As the addition of healthy foods, such as fruits and vege-tables, has not been shown to negate unhealthy diet on car-diometabolic risk outcomes,14 substitution of healthy foodsfor unhealthy ones or elimination may be a sounder healthstrategy; albeit, there are limited studies supporting thisclaim. Elimination of unhealthy foods by social marketingworked as a strategy for young adults to reduce their con-sumption of junk food.33 These strategies have yet to beapplied in Black and Hispanic neighbourhoods. In poorerneighbourhoods, fast food media targeted to Blacks andHispanics, together with lower prices than supermarkets,contribute to high consumption of soft drinks, fried andprocessed foods in these populations.34

The present study had several limitations. The result canbe generalised to U.S. adults, aged 21–69 years, of themajor racial/ethnic groups (MA, OH, NHB and NHW) inthe early twenty first century; albeit, results may differ forAsian Americans, multi-racial Americans and older adults.Limited types of unhealthy foods were considered; theabsolute quantities consumed of unhealthy foods were notmeasured. The magnitude of unhealthy food consumptionmay be overestimated as comparisons were none/little ver-sus moderate/high. Food consumption was measured as asingle time-point, may have recall bias and may not reflectyearly diet. The strength of the study was the use of objec-tively measured anthropometrics and laboratory data,which represented the U.S. population, allowing compari-sons among racial/ethnic groups.

The relationship of unhealthy foods with cardiometabolicrisk has yet to be established. In conclusion, different pro-portions of unhealthy food consumption were found byrace/ethnicity, sex and age-group. Persons of the lowesteducation level had the highest odds of all cardiometabolicrisk factors, adjusting for race/ethnicity, sex and age. Stud-ies establishing type and amount of unhealthy foods con-sumption with cardiometabolic risks are needed.

Funding source

The present study used publicly available data. The authorsdeclare that no funding was received for the present study.

Conflict of interest

All authors disclose that there are not any commercial orsimilar associations that might create a conflict of interest inconnection with the submitted manuscript.

Authorship

JAV and FGH developed the concept and performed the lit-erature search. All authors determined the aim and

hypotheses. JAV conducted the statistical analysis. GGZexamined the analysis and commented on the discussion.All authors contributed to manuscript preparation with crit-ical revisions, agree that no parts of the manuscript havebeen published elsewhere and approved the final version.

References

1 Ruilope LM, de la Sierra A, Segura J, Garcia-Donaire JA. Themeaning of cardiometabolic risk in hypertension patients. USEndocrinol 2007; 1: 60–3.

2 Tapsell LC. Dietary behaviour changes to improve nutritionalquality and health outcomes. Chronic Dis Transl Med 2017;3: 154–8.

3 Jacobs DR Jr, Tapsell LC. Food, not nutrients, is the fundamen-tal unit in nutrition. Nutr Rev 2007; 65: 439–50.

4 Mozaffarian D, Ludwig DS. Dietary guidelines in the 21st cen-tury – a time for food. JAMA 2010; 304: 681–2.

5 Mozaffarian D. Dietary and policy priorities for cardiovasculardisease, diabetes, and obesity – a comprehensive review. Circu-lation 2016; 133: 187–225.

6 Ndanuko RN, Tapsell LC, Charlton KE, Neale EP,Batterham MJ. Dietary patterns and blood pressure in adults: asystematic review and meta-analysis of randomized controlledtrials. Adv Nutr 2016; 7: 76–89.

7 Duffey JK, Gordon-Larsen P, Steffen LM, Jacobs DR,Popkin BM. Regular consumption from fast food establishmentsrelative to other restaurants is differentially associated with meta-bolic outcomes in young adults. J Nutr 2009; 139: 2113–8.

8 de Oliveira Otto MC, Afshin A et al. The impact of dietary andmetabolic risk factors on cardiovascular diseases and type 2 dia-betes mortality in Brazil. PLoS One 2016; 11: e0151503.

9 Micha R, Michas G, Mozaffarian D. Unprocessed red and pro-cessed meats and risk of coronary artery disease and type 2 dia-betes – an updated review of the evidence. Curr Atheroscler Rep2012; 14: 515–24.

10 Sonestedt E, �verby NC, Laaksonen DE, Birgisdottir BE. Doeshigh sugar consumption exacerbate cardiometabolic risk factorsand increase the risk of type 2 diabetes and cardiovascular dis-ease? Food Nutr Res 2012; 56: 19104. https://doi.org/10.3402/fnr.v56i0.19104.

11 Malik VS, Popkin BM, Bray GA, Després J-P, Hu FB. Sugarsweetened beverages, obesity, type 2 diabetes and cardiovascu-lar disease risk. Circulation 2010; 121: 1356–64.

12 Satia JA. Diet-related disparities: understanding the problemand accelerating solutions. J Am Diet Assoc 2009; 109: 610–5.

13 Anderson B, Lyon-Callo S, Fussman C, Imes G, Rafferty AP.Fast-food consumption and obesity among Michigan adults.Prev Chronic Dis 2011; 8: A71.

14 Smith KJ, Blizzard L, McNaughton SA, Gall SL, Dwyer T,Venn AJ. Takeaway food consumption and cardio-metabolicrisk factors in young adults. Eur J Clin Nutr 2012; 66: 577–84.

15 Clemens R, Kranz S, Mobley AR et al. Filling America’s fiberintake gap: summary of a roundtable to probe realistic solu-tions with a focus on grain-based foods. J Nutr 2012; 142:1390S–401S.

16 Institute of Social and Preventive Medicine. Clinical Epidemio-logical & Biostatistics. Berlin, Germany: University of Berline,2009. (Available from: https://www.strobe-statement.org/index.php?id=strobe-home, accessed 12 October 2017).

17 Center for Disease Control and Prevention/National Center forHealth Statistics CDC/NCHS. National Health and Nutrition

Unhealthy food and cardiometabolic risk

© 2018 Dietitians Association of Australia 479

Examination Surveys (NHANES). Atlanta, GA, USA: Center forDisease Control and Prevention, 2017. (Available from: https://wwwn.cdc.gov/nchs/nhanes/default.aspx, accessed 7 August2017).

18 Ahuja JKA, Montville JB, Omolewa-Tomobi G et al. USDA Foodand Nutrient Database for Dietary Studies, 5.0. Beltsville, MD: U.S. Department of Agriculture, Agricultural Research Service,Food Surveys Research Group, 2012.

19 Center for Disease Control and Prevention/National Center forHealth Statistics (CDC/NCHS). National Health and NutritionExamination Survey. NCHS Research Ethics Review Board (ERB)Approval. Atlanta, GA, USA: CDC/National Center for HealthStatistics, 2017. (Available from: https://www.cdc.gov/nchs/nhanes/irba98.htm, accessed 7 August 2017).

20 National Cancer Institute. Dietary Screener Questionnaire in theNHANES 2009–10. Bethesda, MD: NCI, 2018. (Available from:https://epi.grants.cancer.gov/nhanes/dietscreen/, accessed11August2018).

21 United States Department of Agriculture (USDA). The FoodSupply and Dietary Fiber: Its Availability and Effect On Health.Washington, DC: USDA, 2007, No. 36. (Available from:https://www.cnpp.usda.gov/sites/default/files/nutrition_insights_uploads/Insight36.pdf, accessed 18 August 2017).

22 Howlett JF, Betteridge VA, Champ M, Craig SAS, Meheust A,Jones JM. The definition of dietary fibre – discussions at theNinth Vahouny Fiber Symposium: building scientific agree-ment. Food Nutr Res 2010; 54: 5750. https://doi.org/10.3402/fnr.v54i0.5750.

23 American Association of Clinical Endocrinologists and AmericanCollege of Endocrinology (AACE). Comprehensive Type 2 DiabetesManagement Executive Summary. Jacksonville, FL: AACE, 2017.(Available from: https://www.aace.com/sites/all/files/diabetes-algorithm-executive-summary.pdf, accessed 10 August, 2017).

24 Ashwell M. Waist to height ratio and the Ashwell® shape chartcould predict the health risks of obesity in adults and childrenin all ethnic groups. Nutr Food Sci 2005; 35: 359–64.

25 Swainson MG, Batterham AM, Tsakirides C, Rutherford ZH,Hind K. Prediction of whole-body fat percentage and visceraladipose tissue mass from five anthropometric variables. PLoSOne 2017; 12: e0177175.

26 Kilander L, Berglund L, Boberg M, Vessby B, Lithell H. Educa-tion, lifestyle factors and mortality from cardiovascular diseaseand cancer. A 25-year follow-up of Swedish 50-year-old men.Int J Epidemiol 2001; 30: 1119–26.

27 Han E, Powell LM. Consumption patterns of sugar sweetenedbeverages in the United States. J Acad Nutr Diet 2013;113: 43–53.

28 Bahadoran Z, Mirmiran P, Azizi F. Fast food pattern and cardi-ometabolic disorders: a review of current studies. Health Pro-mot Perspect 2015; 5: 231–40.

29 Tareke E, Rydberg P, Karlsson P, Eriksson S, Törnqvist M.Analysis of acrylamide, a carcinogen formed in heated food-stuffs. J Agric Food Chem 2002; 50: 4998–5006.

30 Gadiraju TV, Patel Y, Gaziano JM, Djoussé L. Fried food con-sumption and cardiovascular health: a review of current evi-dence. Nutrients 2015; 7: 8424–30.

31 Nettleton JA, Steffen LM, Mayer-Davis EJ et al. Dietary patternsare associated with biochemical markers of inflammation andendothelial activation in the multi-ethnic study of atherosclero-sis (MESA). Am J Clin Nutr 2006; 83: 1369–79.

32 Hobbs DA, Lovegrove JA, Givens DI. Associations between redand processed meat consumption and cardiometabolic riskmarkers among British adults. Proc Nutr Soc 2018; 77: E22.https://doi.org/10.1017/S0029665117004414.

33 Robinson E, Harris E, Thomas J, Aveyard P, Higgs S. Reducinghigh calorie snack food in young adults: a role for social normsand health based messages. Int J Behav Nutr Phys Act 2013;10: 73.

34 Freeman A. Fast Food: Oppression Through Poor Nutrition.95 Cal. L. Rev. 2221, 2007. (Available from: http://scholarship.law.berkeley.edu/californialawreview/vol95/iss6/8, accessed 18 August2017).

J. A. Vaccaro et al.

480 © 2018 Dietitians Association of Australia

ORIGINAL RESEARCH

Association between adherence to the DietaryApproaches to Stop Hypertension diet with foodsecurity and weight status in adult women

Saeideh TABIBIAN,1 Elnaz DANESHZAD,1 Nick BELLISSIMO,2 Neil R. BRETT,2

Ahmad R. DOROSTY-MOTLAGH1 and Leila AZADBAKHT 1,3

1Department of Community Nutrition, School of Nutritional Sciences and Dietetics and 3Diabetes Research Centre,Endocrinology and Metabolism Clinical Sciences Institute, Tehran University of Medical Sciences, Tehran, Iran; and2School of Nutrition, Ryerson University, Toronto, Canada

AbstractAim: To investigate the association between adherence to the Dietary Approaches to Stop Hypertension (DASH) dietwith food security and weight status in adult women.Methods: This cross-sectional study was carried out in 227 women—20–50 years of age—who were referred from10 health centres. Dietary intakes were assessed using validated food frequency questionnaires. The DASH scorewas calculated using the Fung method based on eight food and nutrient components (high intake of fruits, vegeta-bles, whole grains, nuts and legumes, low-fat dairy, low intakes of red and processed meats, sweetened beveragesand sodium). The United States Department of Agriculture (USDA) household food security questionnaire was usedto assess food security status.Results: The prevalence of food insecurity was 33.9%. A greater percentage of overweight and obese people werein the food insecure group (P = 0.006). In addition, the prevalence of overweight and obese people was lower withgreater adherence to the DASH dietary pattern (P = 0.017). After controlling for age and energy intake, participantsin the highest tertile of adherence to DASH diet had 66% lower odds of overweight and obesity than those in the low-est tertile (odds ratio (OR): 0.34; 95% confidence interval (CI): 0.15; 0.79). This relationship of DASH diet tertile andoverweight and obesity was significant for both food secure women (OR: 0.54, 95% CI: 0.23; 0.97) and food insecurewomen (OR: 0.35, 95% CI: 0.06; 0.42).Conclusions: Adherence to the DASH diet is associated with a reduced risk of overweight and obesity, based onbody mass index, in both food secure and insecure Iranian women.

Key words: DASH diet, food security, obesity, weight status, women.

Introduction

The Dietary Approaches to Stop Hypertension (DASH) is ahealthy dietary pattern1 rich in fruit and vegetables, low inmoderate-fat dairy products, legumes and nuts and low insaturated fatty acids, sodium, added sugars and refinedcarbohydrates.2,3 It is thought that this dietary pattern will

help to prevent cardiovascular disease by improving bloodpressure and weight status.4,5 The DASH dietary patternmay also improve blood glucose control and blood lipidprofiles in adults with type 2 diabetes.6

Food security is a factor that influences dietary patternsand adequacy and is defined as all people at all times hav-ing access to enough food in order to have a socially accept-able, diverse and healthy diet.7 Some variables thatassociate with food insecurity include loss of food aid, jobloss or lack of fixed occupation, age, ethnic minority andlow education of the head of household.8 As summarised ina recent review, food insecurity may lead to weight lossbacause of lack of food or may be obesogenic due tounhealthy diets higher in calorically dense foods.9 Althoughthe literature shows mixed results, food insecurity withouthunger may have a stronger effect to be obesogenic inwomen than in men.9 Furthermore, other risk factors,including hypertension, may be more prevalent in foodinsecure populations because of decreased nutritional

S. Tabibian, MSc StudentE. Daneshzad, MSc, PhD CandidateN. Bellissimo, PhD, Associate ProfessorN.R. Brett, PhD, Postdoctoral FellowA.R. Dorosty-Motlagh, PhD, ProfessorL. Azadbakht, PhD, ProfessorCorrespondence: L. Azadbakht, Department of Community Nutrition,School of Nutritional Sciences and Dietetics, Tehran University ofMedical Sciences, PO Box 1416643931, Tel: (+98) 2188955569;Fax: (+98) 2188984861, Tehran, Iran.Email: azadbakhtleila@gmail.com

Accepted May 2018

© 2018 Dietitians Association of Australia 481

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 481–487 DOI: 10.1111/1747-0080.12440

adequacy and dietary diversity.10 Specifically, several stud-ies have reported that food insecurity is linked to lowerHealthy Eating Index scores,11 because of diets being lowerin fruits and vegetables12,13 and higher in empty calories.9

Obesity is a global epidemic in both developed anddeveloping countries, and has a multitude of negativepsychological, physical and economic consequences.14,15 Asmild-to-moderate food insecurity likely associates withobesity because of unhealthy dietary patterns, it is impor-tant to understand how food insecurity relates to adherenceto the healthy dietary pattern of the DASH diet. Currently,no studies have examined the association of adherence tothe DASH diet, food security and weight status. Further-more, although the prevalence of food insecure householdsin the USA is 15%,16 the prevalence in urban centres inIran was shown to be 44%, with almost twice as manywomen as men being food insecure.17 Thus, the presentstudy was conducted to determine the relationship betweenadherence to the DASH diet with food security and weightstatus in women in Iran.

Methods

The present study adhered to the STrengthening the Report-ing of OBservational studies in Epidemiology’s principles forconducting and disseminating observational studies. Thiscross-sectional study was conducted in 227 women, whowere recruited from the 10 health centres of the TehranUniversity of Medical Sciences (TUMS). Multistage clustersampling was performed, and some centres related to TUMSwere selected randomly. Women referred to these selectedcentres, for any reason, could be included in the presentstudy. The questionnaires were collected with the informedconsent of all participants who had the criteria for enteringthe study. Inclusion criteria included women being20–50 years of age, being Iranian, being non-migrant, notpregnant or lactating, not being postmenopausal and nothaving diseases such as diabetes, cardiovascular disease,cancer, liver dysfunction or kidney dysfunction. All partici-pants completed the demographic, food security and foodfrequency questionnaires (FFQs). The present study wasapproved by the research council of the School of NutritionalSciences and Dietetics (number: 9411323010), TUMS.

The women’s weight was measured to the nearest 0.1 kg,wearing minimal clothing and without shoes, by a digitalscale (SECA, Hamburg, Germany). Height was measured tothe nearest millimetre with an inflexible measuring rod.Body mass index (BMI) calculated as weight in kilogramdivided by height in metres square. Weight status was cate-gorised based on BMI, with BMI less than 18.5 beingunderweight, BMI 18.5–24.9 being healthy weight andBMI ≥ 25 being overweight and obesity. The waist circum-ference was measured to the nearest millimetre using a tapemeasure, from the narrowest area between the last rib andthe navel.18 Waist circumference greater than 88 cm con-sidered as abdominal obesity. Physical activity was evalu-ated based on metabolic equivalents × h/d (Met.h/d) byrecording activity over 24 hours.19 Frequency and time

spent on activities of light, moderate, strong and intenseactivity levels were determined, and the level of physicalactivity of individuals was calculated as Met.h/d.19

The dietary intakes of women over the last year were eval-uated using a semi-quantitative FFQ including a list of168 food items that had been validated previously.20 Thevalues listed for each meal were converted to gram per dayby using a household scale guide.21 We assessed the DASHscore based on the Fung method that assesses food and nutri-ents consumption and emphasised to increase or decreasetheir consumption.22,23 The DASH dietary pattern focuses oneight food groups (fruits, vegetables, whole grains, nuts andlegumes, low-fat dairy products, red and processed meat,sodium and sweetened beverages),22 where the high con-sumption of the first five groups and the low consumption ofmeat, sweetened beverages and salt are desirable.22 In the pre-sent study, individuals were classified into deciles based onthe intakes from the food groups. In the epidemiologicalstudies, to find the relationship exposure with outcome, thesubjects are ranked and compare the highest rank to lowestrank. Therefore, according to sample size, each decile of thesample was assigned points. The lowest decile of intake forfruit, vegetables, whole grains, nuts and low-fat dairy prod-ucts was given 1 point, and the highest intake decile wasgiven 10 points. For sodium, red and processed meats, andsweetened beverages, low intake was desired. Therefore, thelowest intake decile was given a score of 10 points and thehighest intake decile was given 0 point. Finally, we summedthe resulting eight component scores to obtain an overallDASH adherence score of 8–80, with a higher score denotingbetter overall adherence to the diet.

The 18-item USDA questionnaire,24 which examineshousehold food security in the last 12 months, was used.This questionnaire has been validated in previous studies inIran25 Scoring of the questionnaire is as follow: the optionsincluding ‘often correct’, ‘sometimes correct’, ‘almost everymonth’, ‘some months’ and ‘yes’ were given one point. Theoptions including ‘not true’, ‘only 1 or 2 months’ and ‘no’were given zero points. A score of 0–2 points was categorisedas food secure, 3–7 points as food insecure without hunger,8–12 points as food insecure with moderate hunger and ascore of ≥13 points as food insecure with severe hunger.

The relationship between qualitative variables and house-holds’ food security classes was determined with the chi-square test. The relationship between quantitative variablesand household food security classes was assessed by one-wayanalysis of variance. The dietary intakes of individuals werecompared using covariance analysis and by adjusting for theeffect of age and energy intake among tertiles of DASH adher-ence. The prevalence of overweight and obesity and abdomi-nal obesity in the tertiles of DASH adherence was measuredusing the chi-square test. To determine the relationshipbetween DASH diet and overweight or obesity, logistic regres-sion was used by adjusting for confounding factors such asenergy intake, age, physical activity, oral contraceptive pillconsumption, economic situation and marital status in differ-ent models, based on the previous data in Iranian women.26

We evaluated the odds of having food insecurity among

S. Tabibian et al.

482 © 2018 Dietitians Association of Australia

tertiles of DASH adherence with binary logistic regression.Statistical analyses were performed using SPSS version 16(SPSS Inc, Chicago, IL, USA). P < 0.05 was considered statis-tically significant. DASH dietary pattern and food securitywere considered as an independent and a dependent variable,respectively.

Results

The mean age, weight, waist circumference and BMI were33.5 � 8.3 years, 68 � 12.5 kg, 87 � 11 cm and25.9 � 4 kg/m2, respectively. The prevalence of food insecu-rity in the women was 33.9%, of which 25.1% were foodinsecure without hunger and 8.8% were food insecure withmoderate and severe hunger. Therefore, 66% of the popula-tion were food secure. Because of the low percentage of foodinsecurity with severe hunger, two groups of food insecuritywere combined with moderate and severe hunger. Table 1shows the distribution of demographic and anthropometriccharacteristics by food security status among participants.

Findings from chi-square tests showed that food insecuritywas associated with age, education and occupation of partici-pants, head of household education and occupation, havinga child under 18-year-old, and number of people employedin the household and BMI category (P < 0.05).

The dietary intakes of participants are shown based onthe eight nutritional components of DASH dietary pattern(Table 2). Intake of fruits, vegetables, red meat and low-fatdairy products was significantly lower in the food insecuregroup (P < 0.05). There was no significant difference interms of macronutrient intake among DASH dietary scores.However, participants following the DASH dietary patternhad significantly greater intakes of potassium, calcium,magnesium, vitamin C, B1, B2, beta–carotene and dietaryfibre, with lower intakes of sodium.

A greater adherence to DASH was associated with alower BMI in the women (P = 0.017) (Table S1, Supportinginformation). Association between food insecurity andDASH adherence was investigated by calculating the preva-lence of food insecurity in the tertiles of DASH adherence.In addition, this association was evaluated in unadjustedbinary logistic regression models (Table 3). The odds offood insecurity were negatively related to DASH tertile(P = 0.017) with the highest tertile of DASH adherence,having 60% lower odds of food insecurity than the lowesttertile (95% confidence interval (CI): 0.19–0.8).

There was no correlation of DASH adherence with over-weight and obesity in the crude model (P = 0.18) (Table 3).However, after adjustment for age and energy intake, theodds of overweight and obesity among women in the high-est tertile of DASH adherence were 66% lower than thelowest tertile (model 1, 95% CI: 0.15–0.79). Similarly, afteradjusting for the age and energy intake, smoking, physicalactivity, oestrogen consumption and familial history of dia-betes and stroke (model 2), there was a significant inverserelationship between DASH diet adherence and overweightand obesity (P < 0.038). After adjusting for confoundingvariable in model 3, the odds of obesity decreased in twosecure and insecure groups (food secure women: odds ratio(OR): 0.54, 95% CI: 0.23–0.97; food insecure women: OR:0.35, 95% CI: 0.06–0.42). Finally, increased adherence tothe DASH diet was not significantly associated with abdom-inal obesity in the crude model or the models adjusted forconfounding variables (Table 3).

Discussion

In the present study, there was a greater prevalence of over-weight and obesity in the food insecure group of Iranianwomen. Furthermore, DASH dietary score in both foodsecure and food insecure women negatively associated withthe BMI category. However, DASH dietary score did notassociate with waist circumference. The findings of thiscross-sectional study suggest a relationship between adher-ence to DASH diet and the risk of overweight and obesity,based on the BMI category, but not waist circumference, infood insecure and food secure women.

Table 1 Comparison of demographic and anthropometriccharacteristics of Iranian women by food security statuscategory (n = 227)

Characteristics

Foodinsecure(n = 77)

Foodsecure

(n = 150) P-value(a)

Body mass index (kg/m2)<18.5 4 (5.2) 1 (0.7) 0.00618.5–25 24 (31.2) 74 (49.3)>25 49 (63.6) 75 (50)

Education of womenDiploma and less 76 (98.7) 38 (25.3) 0.0001College education 1 (1.3) 112 (74.7)

Job of womenHousewife/other 12 (15.6) 78 (52) 0.0001Free job/employee 65 (84.4) 72 (48)

Children under 18-year-oldYes 29 (37.7) 35 (23.3) 0.018No 48 (62.3) 115 (76.7)

Household economic levelWeak 76 (98.7) 41 (27.3) 0.001Average 1 (1.3) 54 (36)Desirable 0 (0) 55 (36.7)

Head of the household occupationFree job/administrative

manage/employee34 (44.2) 140 (93.3) 0.0001

Worker/other 43 (55.8) 10 (6.7)Head of household educationDiploma and less 77 (51.3) 77 (100) 0.0001College education 0 (0) 73 (48.7)

Employed member (s) of the household0–1 55 (71.4) 59 (39.3) 0.00012–3 84 (56) 21 (27.3)≥4 7 (4.7) 1 (1.3)

Data are: n (%).(a) χ2 test was used for all variables to compare food security statuscategories.

DASH diet and food security

© 2018 Dietitians Association of Australia 483

The present study showed that there was greater preva-lence of overweight and obesity, based on the BMI cate-gory, in the food insecure women. Likely leading to foodinsecurity, was the lower education status and the lowerhousehold economic level in food insecure compared tofood secure women. The observed food insecurity andweight status relationship agrees with previous cross-sectional data in women from the USA.27 Furthermore, inthe present study, consumption of fruits, vegetables, low-fat dairy products and meat was lower in the food inse-cure group compared to the food secure group. Similarly,2003–2010 National Health and Nutrition ExaminationSurvey data,27 and Canadian cross-sectional data reportedthat food insecurity associated with unhealthy dietary pat-terns because of low consumption of fruits, vegetables,and meat or higher intakes of added sugars.27 This patternof intake may partially reflect the higher cost of healthierfoods, such as lean meat, fish and some fresh vegetablesand fruits. The food insecure group in the present studyhad a higher prevalence of women with children <18 yearsof age (37.7%) compared to the food secure group(23.3%). Interestingly, an American study of ~8000women reported that food insecure mothers are at a

higher risk of obesity compared to food insecure womenwithout children,28 because of the gendered nature ofchildcare where mothers are prioritising the diet of chil-dren ahead of their own diet.

Both food secure and food insecure women in the pre-sent study had a lower risk of overweight and obesity inthe highest tertile of the DASH diet. Although the DASHdiet was designed to lower cardiometabolic risk, high con-cordance with the DASH diet may also decrease the risk ofobesity, as shown in previous studies in women.26,29 As itis likely that only a high concordance with the DASH dietwill lower the risk of obesity, it is logical that the over-weight and obesity risk of the middle DASH tertile in thepresent study was not different from the lowest DASHtertile. In the present study, women in the highest DASHtertile consumed greater amounts of fruits, vegetables, low-fat dairy products, whole grains, legumes, nuts and multi-ple micronutrients compared to lower tertiles. The highestDASH tertile also had a lower sodium intake compared toother tertiles. In observational studies, the consumption ofvegetables and fruits has associated with lower risk of obe-sity.30 However, based on a 2014 meta-analysis of rando-mised trials, fruit and vegetable intake may not relate to

Table 2 The dietary intakes of participants across food security status and tertiles of the Dietary Approaches to Stop Hyper-tension (DASH) eating pattern (n = 227)

Food security Tertiles of DASH score

Food groupsFood insecure(n = 77)

Food secure(n = 150) P-value

1 (n = 80)Cut-point<40.90

2 (n = 76)Cut-point41–47

3 (n = 71)Cut-point>47.01 P-value(a)

Fruit 145 � 144 376 � 203 0.0001 227 � 2.1 328 � 2.2 453 � 2.4 0.0001Vegetables 329 � 222 352 � 170 0.030 253 � 2.1 344 � 2.2 448 � 2.3 0.0001Meat 93 � 50 140 � 76 0.0001 89 � 0.71 95 � 0.74 102 � 0.79 0.330Nuts/legume 38.5 � 25 45 � 32 0.132 34 � 0.34 41 � 0.36 53 � 0.38 0.0001Fats and oils 16 � 10.46 16 � 10.38 0.798 16 � 0.12 15 � 0.13 16 � 0.14 0.896Whole grains 121 � 77 123 � 80 0.873 93 � 0.8 117 � 0.94 161 � 0.9 0.0001Low-fat dairy 368 � 231 555 � 293 0.0001 354 � 3.1 481 � 3.3 658 � 3.5 0.0001Sweet beverages 52 � 84 55 � 66 0.785 65 � 0.8 56 � 0.9 42 � 1 0.189NutrientsProtein (g) 74 � 22 79 � 35 0.266 78 � 0.21 76 � 0.22 78 � 0.24 0.770Fat (g) 63 � 26 66 � 28 0.401 65 � 0.19 65 � 0.2 64 � 0.2 0.897Carbohydrate (g) 339 � 108 325 � 117 0.380 328 � 0.4 329 � 0.4 333 � 0.4 0.627Cholesterol (mg) 170 � 102 221 � 158 0.010 187 � 1.5 205 � 1.6 222 � 1.7 0.225Sodium (mg) 5002 � 2531 5046 � 2983 0.912 6400 � 28 4683 � 29 3856 � 31 0.0001Potassium (mg) 3031 � 1210 3300 � 1622 0.201 2888 � 11 3194 � 11 3587 � 12 0.0001Magnesium (mg) 246 � 89 274 � 131 0.099 251 � 0.88 263 � 0.93 281 � 0.99 0.034Calcium (mg) 946 � 389 1021 � 446 0.212 904 � 3 997 � 3 1098 � 4 0.0001Fibre (g) 16 � 7 17 � 11 0.646 15 � 0.09 17 � 0.1 19 � 0.1 0.012Folic acid (mg) 284 � 109 313 � 204 0.236 307 � 1.6 291 � 1.7 312 � 1.8 0.569Vitamin C (mg) 125 � 6.6 131 � 4.7 0.045 100 � 6.02 131 � 6.17 157 � 6.36 0.0001Vitamin B1 (mg) 1.94 � 0.41 1.87 � 0.29 0.139 13.5 � 0.62 15.3 � 0.64 16.4 � 0.66 0.006Vitamin B2 (mg) 1.84 � 0.05 2.04 � 0.03 0.002 1.86 � 0.05 1.96 � 0.05 2.1 � 0.05 0.005Beta–Carotene (μg) 633 � 87 718 � 63 0.0439 464 � 85 674 � 87 957 � 89 0.0001Vitamin B6 (mg) 1.5 � 0.45 1.67 � 0.32 0.074 1.58 � 0.04 1.62 � 0.04 1.71 � 0.04 0.122Vitamin B12 (mg) 3.44 � 0.1 4.25 � 0.141 0.001 3.8 � 0.19 3.9 � 0.2 4.15 � 0.21 0.500

(a) Adjusted for energy intake. Analysis of covariance (ANCOVA) test was used to test for differences among tertiles or between food securitystatus groups. The bold P-values represent significant P-values (P < 0.05).

S. Tabibian et al.

484 © 2018 Dietitians Association of Australia

weight loss, because the most important factor for weightloss is energy restriction. Simlarly, low-fat dairyproducts,31–33 legumes34,35 and whole grains36 have nega-tively associated with obesity in observational studies.Effects of these foods on weight maintenance may be dueto low energy density, high fibre intake,34,35 protein37 con-tent or micronutrients like calcium.38 Lastly, a lowersodium intake paired with a higher potassium intake, as inthe highest tertile of the DASH diet, was related to lowerodds of being hypertensive, especially in women.39

Although the DASH diet score negatively associated withoverweight and obesity based on the BMI category, thisrelationship was not seen when using waist circumferenceas a marker of abdominal obesity. It is possible that thislack of statistical relationship may be due to the influenceof lifestyle or environmental factors, including that foodinsecurity and low household income may be temporary, orcould change seasonally. However, it is also possible thatthe DASH diet score associated differently with BMI tertilesand waist circumference categories in women because ofthe differences between these measures of weight status.Women with higher BMI may have more adiposity thanthose in the healthy BMI category, but accumulation of adi-pose tissue could happen in the gluteal-femoral region,which is associated with lower cardiovascular disease riskthan abdominal obesity.40 Gluteal-femoral adipose accumu-lation is commonly referred to as the pear-shaped bodytype and is more common in women than men.40 As BMIis not able to differentiate among body types, when

associating diet and food security status with obesity inwomen, waist circumference may be a better marker of obe-sity and disease risk than BMI.

The present study was novel in that it is the first study,to our knowledge, that has associated food security with aparticular food pattern (DASH pattern) and overweight andobesity risk. However, the present study had a number oflimitations. First, by design, cross-sectional work is not ableto elucidate cause and effect relationships. Also, due thecross-sectional design, it cannot be ascertained whetherfood insecurity in households was temporary or chronic.Furthermore, although both BMI and waist circumferencewere measured, more accurate assessments of adipositymay help to further clarify how obesity risk relates to theinteraction of diet pattern and food security status. Lastly,future work would be strengthened by the measurement ofcardiometabolic parameters including blood pressure, bloodlipids and fasting blood glucose.

In conclusion, the findings of the present study showedthat in Iranian women, food insecurity and low DASH die-tary score both increase the risk of overweight and obesity,based on BMI categories. Longitudinal studies are neededto confirm these findings.

Funding source

The present study was supported by Tehran University ofMedical Sciences (grant number: 9411323010).

Table 3 Odds ratios of overweight and obesity based on body mass index categories and waist circumference across the ter-tiles of adherence to the Dietary Approaches to Stop Hypertension (DASH) eating pattern

Tertiles of DASH score

Model

1 (n = 80)Cut-point<40.90

2 (n = 76)Cut-point41–47

3 (n = 71)Cut-point>47.01 P value(a)

Food insecurity 1 0.978 (0.5–1.8) 0.4 (0.19–0.8) 0.017Obesity total population

Crude 1 1.65 (0.86–3.1) 0.63 (0.3–1.19) 0.180Model 1(b) 1 1.2 (0.6–2.6) 0.34 (0.15–0.79) 0.019Model 2(c) 1 1.2 (0.57–2.6) 0.35 (0.14–0.82) 0.038

Secure womenCrude 1 1.5 (0.68–3.57) 0.69 (0.32–1.5) 0.32Model 1(b) 1 0.86 (0.32–2.2) 0.3 (0.11–0.88) 0.2Model 2(c) 1 1.96 (1.02–5.5) 0.54 (0.23–0.97) 0.03

Insecure womenCrude 1 1.8 (0.64–5.5) 0.59 (0.17–2.06) 0.66Model 1(b) 1 1.12 (0.6–6.66) 0.29 (0.06–1.32) 0.31Model 2(c) 1 1.58 (0.7–5.6) 0.35 (0.06–0.92) 0.02

Abdominal obesityCrude 1 1.4 (0.7–2.7) 0.8 (0.4–1.5) 0.570Model 1(b) 1 1.18 (0.54–2.3) 0.53 (0.24–1.1) 0.140Model 2(c) 1 1.07 (0.51–2.2) 0.54 (0.23–1.2) 0.200

(a) Binary logistic regression was used in all models and to produce the resulting P values. The bold P-values represent significant P-values(P < 0.05).(b) Adjusted for energy intake and age.(c) Adjusted for energy intake, age, physical activity, oral contraceptive pill consumption, economic situation and marital status.

DASH diet and food security

© 2018 Dietitians Association of Australia 485

Conflict of interest

The authors declare that they have no conflict of interest.

Authorship

ST, ARD and LA proposed, conducted and designed thestudy. LA supervised the study. ST and LA performed thestatistical analyses. ST prepared a first draft of the manu-script and ED finalised it. ED helped in conducting thestudy and drafting the paper and providing the suitablediscussion on the results. NRB and NB contributed to datainterpretation, statistical analyses and revised the manu-script. All authors approved the final version of themanuscript.

References

1 US Department of Health and Human Services, US Departmentof Agriculture. Dietary Guidelines for Americans 2005.Washington, DC: US Government Printing Office, 2005.

2 Fung TT, Chiuve SE, McCullough ML, Rexrode KM,Logroscino G, Hu FB. Adherence to a DASH-style diet and riskof coronary heart disease and stroke in women. Arch InternMed 2008; 168: 713–20.

3 Department of Health and Human Services. Your Guide to Low-ering Your Blood Pressure with DASH. Washington, DC: USDepartment of Health and Human Services, 2006. NIH Publi-cation No. 06-4082.

4 Van Meijl LE, Mensink RP. Effects of low-fat dairy consump-tion on markers of low-grade systemic inflammation and endo-thelial function in overweight and obese subjects: aninterventionstudy. Br J Nutr 2010; 28: 1–5.

5 Moore TJ, Alsabeeh N, Apovian CM et al. Weight, blood pres-sure, and dietary benefits after 12 months of a web-basednutrition education program (DASH for health): longitudinalobservational study. J Med Internet Res 2008; 10: e52.

6 Azadbakht L, Surkan PJ, Esmaillzadeh A, Willett WC. The die-tary approaches to stop hypertension eating plan affectsC-reactive protein, coagulation abnormalities, and hepaticfunction tests among type 2 diabetic patients. J Nutr 2011;141: 1083–8.

7 Furness BW, Simon PA, Wold CM, Anderson J. Prevalence andpredictors of food insecurity among low-income households inLos Angeles Country. Public Health Nutr 2004; 7: 791–4.

8 Jayna M, Dave AE, Evans RP, Saunders KW, Watkins KA.Associations among food insecurity, acculturation, demo-graphic factors, and fruit and vegetable intake at home in His-panic children. J Am Diet Assoc 2009; 109: 697–701.

9 Morales ME, Berkowitz SA. The relationship between foodinsecurity, dietary patterns, and obesity. Current Nutr Rep2016; 5: 54–60.

10 Holben DH, Pheley AM. Diabetes risk and obesity in food-insecure households in rural Appalachian Ohio. Prev ChronicDis 2006; 3: A82.

11 Drewnowski A, Specter SE. Poverty and obesity: the role ofenergy density and energy costs. Am J Clin Nutr 2004;79: 6–16.

12 Dixon LB, Winkleby M, Radimer K. Dietary intakes and serumnutrients differ between adults from food insufficient and

food-sufficient families: Third National Health and NutritionExamination Survey, 1988–1994. J Nutr 2001; 131: 1232–46.

13 Kendell A, Olson CM, Frongillo EA Jr. Relationship of hungerand food insecurity to food availability and consumption. J AmDiet Assoc 1996; 96: 1019–24.

14 Obesity: preventing and managing the global epidemic. Reportof a WHO consultation. World Health Organ Tech Rep Ser2000; 894: 1–253.

15 Kolotkin RL, Meter K, Williams GR. Quality of life and obesity.Obes Rev 2001; 2: 219–29.

16 Franklin B, Jones A, Love D, Puckett S, Macklin J, White-Means S. Exploring mediators of food insecurity and obesity: areview of recent literature. J Community Health 2012; 37:253–64.

17 Ramesh T, Dorosty-Motlagh AR, Abdollahi M. Prevalence ofhousehold food insecurity in the City of Shiraz and its associa-tion with socio-economic and demographic factors, 2008. IranJ Nutr Sci Food Technol 2010; 4: 53–64.

18 Wang Y, Lim H, Caballero B. Use and interpretation of anthro-pometry. In: Ross C, Caballero B, Cousins R, Tucker K,Ziegler T, eds. Modern Nutrition in Health and Disease. China:Troy B, 2014; 649–59.

19 Seyed Ghalaeh R, Gholi Z, Saraf Bank S, Azadbakht L. Fruitand vegetable intake associated with body mass index andwaist circumference in young female students. Health Syst Res2011; 7: 11.

20 Asghari G, Rezazadeh A, Hosseini-Esfahani F, Mehrabi Y,Mirmiran P, Azizi F. Reliability, comparative validity and stabil-ity of dietary patterns derived from an FFQ in the Tehran lipidand glucose study. Br J Nutr 2012; 108: 1109–17.

21 Ghaffarpour M, Houshiar-Rad A, Kianfar H. The Manual forHousehold Measures, Cooking Yields Factors and Edible Portion ofFood. Tehran, Iran: Keshaverzi Press, 1999.

22 Fung TT, Hu FB, Wu K, Chiuve SE, Fuchs CS, Giovannucci E.The Mediterranean and Dietary Approaches to Stop Hyperten-sion (DASH) diets and colorectal cancer. Am J Clin Nutr 2010;92: 1429–35.

23 Farhadnejad H, Asghari G, Mirmiran P, Azizi F. Association ofDietary Approach to Stop Hypertension (DASH) diet with3-year changes in body mass index and risk of obesity in ado-lescents: Tehran lipid and glucose study. Iran J EndocrinolMetab 2017; 18: 325–33.

24 Bickel G, Nord M, Price C, Hamilton W, Cook J. Measuringfood security in the United States. Guide to measuring house-hold food security, revised 2000 food and nutrition service, officeof analysis, nutrition, and evaluation, US Department of Agricul-ture, Alexandria, VA, 2000 (Available from: https://www.fns.usda.gov/guide-measuring-household-food-security-revised-2000).

25 Hakim S, Dorosty AR, Eshraghian M. Association of food inse-curity and household socio-economic status with the bodymass index among urban women in Dezful. J Sch Public HealthInst Public Health Res 2010; 8: 55–66.

26 Barak F, Falahi E, Keshteli AH, Yazdannik A, Esmaillzadeh A.Adherence to the dietary approaches to stop hypertension(DASH) diet in relation to obesity among Iranian femalenurses. Public Health Nutr 2014; 8: 1–8.

27 Nguyen BT, Shuval K, Bertmann F, Yaroch AL. The supple-mental nutrition assistance program, food insecurity, dietaryquality, and obesity among U.S. adults. Am J Public Health2015; 105: 1453–9.

28 Martin MA, Lippert AM. Feeding her children, but risking herhealth: the intersection of gender, household food insecurityand obesity. Soc Sci Med 2012; 74: 1754–64.

S. Tabibian et al.

486 © 2018 Dietitians Association of Australia

29 Jones N, Ray S, Monsivais P. The DASH diet, cardiovasculardisease risk and obesity in the United Kingdom. FASEB J 2015;29: 119.3.

30 Bes-Rastrollo M, Martinez-Gonzalez MA, Sanchez-Villegas A,de la Fuente Arrillaga C, Martinez JA. Association of fiberintake and fruit/vegetable consumption with weight gain in aMediterranean population. Nutrition 2006; 22: 504–11.

31 Comuzzie AG, Allison DB. The search for human obesitygenes. Science 1998; 280: 1374–7.

32 Xue B, Moustaid-Moussa N, Wilkinson WD, Zemel MB. Theagouti gene product inhibits lipolysis in human adipocytes viaa ca2+−dependent mechanism. FASEB J 1998; 12: 1391–6.

33 Slimani H, Chajès N, May V et al. Nut intake and 5-yearchanges in body weight and obesity risk in adults: results fromthe EPIC-PANACEA study. Euro J Nutr 2017 [Epub ahead ofprint]. https://doi.org/10.1007/s00394-017-1513-0.

34 Grooms KN, Ommerborn MJ, Pham DQ, Djoussé L, Clark CR.Dietary fiber intake and cardiometabolic risks among USadults, NHANES 1999–2010. Am J Med 2013; 126: 1059–67.

35 Raben A, Christensen NJ, Madsen J, Holst JJ, Astrup A.Decreased postprandial thermogenesis and fat oxidation butincreased fullness after a high-fiber meal compared with a low-fiber meal. Am J Clin Nutr 1994; 59: 1386–94.

36 McRae MP. Health benefits of dietary whole grains: anumbrella review of meta-analyses. J Chiropr Med 2017;16: 10–8.

37 Weigle DS, Breen PA, Matthys CC et al. A high-protein dietinduces sustained reductions in appetite, ad libitum caloricintake, and body weight despite compensatory changes in diur-nal plasma leptin and ghrelin concentrations. Am J Clin Nutr2005; 82: 41–8.

38 Zemel MB. Proposed role of calcium and dairy food compo-nents in weight management and metabolic health. PhysSportsmed 2009; 37: 29–39.

39 Murao S, Takata Y, Yasuda M, Osawa H, Kohi F. The influenceof sodium and potassium intake and insulin resistance onblood pressure in normotensive individuals is more evident inwomen. Am J Hypertens 2018 [Epub ahead of print] https://doi.org/10.1093/ajh/hpy041.

40 Zaadstra BM, Seidell JC, Van Noord P et al. Fat and femalefecundity: prospective study of effect of body fat distributionon conception rates. BMJ 1993; 306: 484–7.

Supporting information

Additional Supporting Information may be found in theonline version of this article at the publisher’s web-site:

Table S1 Body mass index across tertiles of DietaryApproaches to Stop Hypertension score and food securitystatus among Iranian women (n = 227)

DASH diet and food security

© 2018 Dietitians Association of Australia 487

ORIGINAL RESEARCH

Associations between dietary behaviours andperceived physical and mental health status amongKorean adolescents

Subin PARK ,1 Soo J. RIM1 and Junghyun H. LEE21Mental Health Research Institute and 2Department of Psychiatry, National Centre for Mental Health, Seoul, SouthKorea

AbstractAim: To investigate the associations between a wide range of dietary behaviours and self-reported physical andmental health status in a representative sample of Korean adolescents.Methods: Data from the 2016 Korea Youth Risk Behaviour Web-based Survey were used. Participants weremiddle- and high-school students (n = 65 528) in grades 7–12. Participants’ dietary behaviours were assessedusing questionnaires on three encouraged dietary behaviours (consumption of fruits, vegetables, and milk) andthree discouraged behaviours (skipping breakfast and consumption of fast food and soft drinks). Participants werealso asked to rate their perceived general and oral health, happiness, sleep satisfaction, stress, depressed mood,and suicidal ideation.Results: After adjusting for sex, school grade, residential area, socioeconomic status, and other dietary behaviours,a high intake of fast food and soft drink and frequent skipping of breakfast were all associated with worse physicaland mental health status. Moreover, a high intake of fruits, vegetables, and milk were associated with better per-ceived general health, oral health, happiness, and sleep satisfaction.Conclusions: Our data suggest that existing encouraged dietary habits mostly have beneficial effects on perceivedphysical and mental health in Korean adolescents. However, the cross-sectional study design prevents our ability toassess causal relationships.

Key words: adolescent, behaviour, diet, mental health.

Introduction

During adolescence, important physical, neurological, cog-nitive, emotional, social, and behavioural developmentoccurs.1 As a result, adolescents become more vulnerable toenvironmental changes. Numerous health-related behav-iours that can affect later life, such as tobacco and alcoholuse, obesity, and physical inactivity, begin during thisstage,2 as do various mental disorders.3 In addition, adoles-cents with a high-risk lifestyle are more likely to have worsehealth in adolescence and adulthood compared to thosewith a low-risk lifestyle.4 Given that behaviours during ado-lescence can affect health in later life, research on adoles-cents seems crucial.

Dietary behaviour is a particularly important factorinfluencing adolescents’ development and health. However,comparatively less research has been done on adolescents’health compared to that of children or elders.5 Recently,there have been several studies on adolescents’ dietarybehaviour and its association with health, but these haveconcentrated more on how dietary behaviour relates toexternalising behaviours (e.g. hyperactivity) than to interna-lising ones (e.g. depressive symptoms, anxiety).6

Increasing attention is now being given to the associationbetween dietary behaviours and internalising behaviours,6–8

but so far the results are inconsistent and further research isrequired. While a meta-analysis6 showed consistently thatunhealthy foods are associated with worse mental health, itwas not able to find consistent support for the notion thathealthy foods boost adolescents’ mental health.

South Korea has gone through rapid economic growth inthe past few decades, leading to marked changes in Koreanadolescents’ lifestyles. One major change concerns their die-tary behaviour. Korean adolescents have adopted more wes-ternised diets, which has led to an increase in weight andheight and a decrease in puberty onset.9 Adolescents whoconsume a western-style diet tend to have a lower desire to

S. Park, MD, PhD, Director of Department of Research PlanningS.J. Rim, BA, Research AssistantJ.H. Lee, MD, PhD, PsychiatristCorrespondence: S. Park, Department of Research Planning, MentalHealth Research Institute, National Centre for Mental Health,127 Yongmasan-ro, Gwangin-gu, Seoul 04933, South Korea.Email: subin-21@hanmail.net

Accepted May 2018

© 2018 Dietitians Association of Australia488

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 488–493 DOI: 10.1111/1747-0080.12444

consume fruits, vegetables, dairy products, and wholegrains, and prefer consuming soft drinks and fast foods.10

Several studies have examined dietary patterns in Koreanadolescents and their association with specific areas ofhealth, such as obesity-related risk factors,11 depression,12

and body image distortion.13 However, to our knowledge,no study has examined the association between a widerange of dietary behaviours and multiple physical and men-tal health outcomes in Korean adolescents. The presentstudy investigated the associations between dietary patternsand self-reported physical and mental health status in a rep-resentative sample of Korean adolescents.

Methods

We used data from the 2016 Korea Youth Risk BehaviourWeb-based Survey (KYRBS), an anonymous online self-report survey conducted by the Korea Centres for DiseaseControl and Prevention. The KYRBS sought to identifyhealth behaviours among middle- and high-school students.It utilised stratified cluster sampling to extract middle- andhigh-school student samples representative of the adoles-cent population in South Korea. After a trained teacher fullyexplained the survey and gave instructions, students whoagreed to participate in the study pressed a button on thescreen and completed the anonymous, self-administeredquestionnaire in their school’s computer laboratory. In the2016 KYRBS, a total of 65 528 respondents from800 schools (400 middle schools and 400 high schools)participated. Details regarding the sampling and surveymethods can be obtained elsewhere.14 The KYRBS wasapproved by the institutional review board of the KoreaCentres for Disease Control and Prevention (2014-06EXP-02-P-A). This manuscript was written in accordance withthe STROBE (STrengthening the Reporting of OBservationalstudies in Epidemiology) statement.

The predictor variables comprised three encouraged die-tary behaviours (i.e. consumption of fruits, vegetables, andmilk) and three discouraged dietary behaviours (i.e. skip-ping breakfast, consuming fast food and soft drink) basedon the adolescent portion of the food-based dietary guide-lines for Koreans.15 The KYRBS asked students how oftenthey engaged in each dietary behaviour within the pastweek; based on their answers, participants were dividedinto two groups for all six dietary behaviours. For skippingbreakfast (not including only milk or juice), participantswere categorised into the high-skipping group (skipping5 or more times a week) and low-skipping group (skippingless than 5 times a week). For fast food (e.g. pizza, ham-burger, fried chicken) and soda consumption, they weredivided into high (three or more times a week) and low-frequency consumption groups (less than 3 times a week).For vegetable consumption, they were divided into high(twice or more each day) and low-frequency consumptiongroups (less than twice every day). For fruit (not includingfruit juice) and milk (including white milk and flavouredmilk) consumption, they were divided into high (once ormore every day) and low-frequency consumption groups

(less than once a day). The specific questions and answerchoices can be found elsewhere.14

Dietary behaviours must be considered as a whole,because they are interrelated.16 Therefore, the dietarybehaviour score was calculated by adding 1 point for eachencouraged dietary behaviour and the absence of each dis-couraged behaviour. Therefore, this dietary behaviour scoreranged from 0 to 6, with higher scores indicating healthierbehaviour.

Sociodemographic variables included gender, schoolgrades, residential area (i.e. rural area, small city, or largecity), and perceived socioeconomic status (SES).17 SESwas assessed on a 5-point Likert scale (e.g. (i) low, (ii)low-middle, (iii) middle, (iv) high-middle, and (v) high).17

The outcome variables included seven physical andmental health variables. Self-perceived general and oralhealth were evaluated with the following questions: ‘Howhealthy do you usually feel?’ and ‘Usually, how do youthink your oral health is, such as tooth and gum health?’,respectively. The response options were: (i) very healthy,(ii) healthy, (iii) average, (iv) unhealthy, or (v) veryunhealthy. Self-perceived happiness was evaluated with thefollowing question: ‘How happy do you usually feel?’ Par-ticipants chose one of the following response options:(i) very happy, (ii) happy, (iii) average, (iv) unhappy, or(v) very unhappy. The degree of sleep satisfaction was eval-uated with the question, ‘In the past 7 days, did you getadequate sleep to overcome fatigue?’ Participants chosefrom one of the following response options: (i) very ade-quate, (ii) adequate, (iii) average, (iv) not adequate, or(v) not very adequate. Perceived stress was measured withthe question ‘To what degree are you usually stressed?’ Par-ticipants chose from response options of (i) very much,(ii) somewhat, (iii) average, (iv) not so much, or (v) not atall. For all these above questions, participants were classi-fied into groups as follows: those who selected options1 or 2 were classified into an ‘above average group’ andthose who selected options 3–5 were classified as the ‘aver-age or below average’ group.17,18

For the statistical analysis, we used multiple logisticregression analysis to calculate the association of each phys-ical and mental health variable (main outcome) with eachdietary behaviour (principal predictor) after adjusting forsex, school grade, residential area, SES, and dietary behav-iours other than principal predictor. Odds ratios and 95%confidence intervals were derived. To investigate the associ-ation of general healthy dietary behaviour and perceivedhealth status, a series of logistic regression analyses was per-formed using perceived health as the main outcome vari-able and dietary behaviour score as the predictor whileadjusting for sex, school grade, residential area, and SES.We used SPSS Statistics 21.0 (IBM Corp., Armonk, NY),with statistical significance defined as an alpha level of0.05. To obtain the Bonferroni-corrected p-values, wedivided the original α-value by the number of analyses per-formed with the dependent variables. Thus, after the Bon-ferroni correction, the statistical threshold was adjusted to0.008 (P = 0.05/6 dietary behaviours).

Dietary behaviours and perceived health

© 2018 Dietitians Association of Australia 489

Results

Table 1 displays the demographic characteristics of partici-pants. Approximately 28% of participants skipped breakfastfive or more days a week, 27% drank soft drinks three ormore times a week, and 16% consumed fast food three ormore times a week. The mean dietary behaviour score was3.08 � 1.26.

As shown in Table 2, all discouraged dietary behaviourswere associated with lower odds of good physical and men-tal health. All encouraged dietary behaviours were associ-ated with higher odds of perceived general health, oralhealth, happiness, and sleep satisfaction but no associationwith perceived stress or depressive mood.

As shown in Table 3, healthier dietary behaviour wasassociated with higher odds of perceived general health,oral health, happiness, and sleep satisfaction, and lowerodds of perceived stress and depressive mood.

Discussion

According to our results, discouraged dietary behaviours(i.e. frequently skipping breakfast, consuming fast food,drinking soft drink) were consistently associated with lower

odds of perceived general health, oral health, happiness,and sleep satisfaction, and higher odds of perceived stress,depressive mood, and suicidal ideation. Encouraged dietarybehaviours (i.e. fruits, vegetables, milk consumption) wereonly associated with higher odds of perceived general andoral health, happiness, and sleep satisfaction. These resultsare compatible with a previous meta-analytic study report-ing consistent associations between unhealthy foods andworse mental health, but inconsistent associations ofhealthy foods with better mental health.6

Korean adolescents that frequently skip breakfast tendedto have lower perceived physical and mental health. Thisresult is consistent with previous studies showing that indi-viduals who frequently skip breakfast have lower perceivedgeneral health and higher stress level, depressed feeling,and emotional distress.19,20 Adolescents who skip breakfastoften have an unhealthy lifestyle21 such as consumingsnacks between main meals, foods high in sucrose, andalcohol more often than do those who regularly eat break-fast.22 This relationship could be explained by the role ofcarbohydrates. When individual fast during sleep, theirblood glucose concentration decreases, while adrenalin andcortisol are released, generating irritability and anxiety.23

When individuals eat breakfast, cortisol production isreduced, leading to a lower level of perceived stress andpositive mood.24 Therefore, skipping breakfast is associatedwith worse physical as well as worse mental health amongadolescents.

Frequent fast food consumption was related to worsephysical and mental health. Frequently consuming fast foodhas a detrimental effect on individual’s physical health, suchas causing endothelial dysfunction, inflammation, and car-diovascular disease (CVD).25,26 This decreased physicalhealth, in turn, can impact mental health(e.g. depression).27,28 In addition, fast food consumption isassociated with low happiness,29 higher depression,30 andpsychiatric distress (i.e. worry, depression, confusion,insomnia, anxiety, aggression, and worthless).31 However,stress and depression might also increase consumption offast food, leading to physical health problems.32

Drinking soft drinks has consistently been associated withworse physical health, such as type 2 diabetes and metabolicsyndrome, and dental health.33,34 This could be explained bythe association between soft drink consumption and higherenergy intake. Soda consumption have high-glycaemicindexes,33 which makes individuals consume more fastfood;35 this, in turn, can be detrimental to their health.36

Soda consumption has been associated with depression,stress-related problems, suicidal ideation,37 psychologicaldistress,31 and lower happiness.29 Sugar-sweetened softdrinks were also associated with hyperactivity,38 behaviouralproblems,35 and lethargy.39

Frequent consumption of fruits and vegetables was asso-ciated with higher odds of perceived general health, oralhealth, happiness, and sleep satisfaction. According to pre-vious studies,40,41 fruit and vegetable consumption reducesthe risk of CVD, cancer, and even all-cause mortality. Thefibre contained in fruits and vegetables is related to a lower

Table 1 Demographic and health characteristics of Koreanadolescents (n = 65 528)

n (%)

SexMale 33 803 (51.6)Female 31 725 (48.4)

School gradeMiddle school, 1st 10 483 (16.0)Middle school, 2nd 10 517 (16.0)Middle school, 3rd 11 219 (17.1)High school, 1st 11 355 (17.3)High school, 2nd 11 070 (16.9)High school, 3rd 10 884 (16.6)

Age, mean � SD (years) 14.99 � 1.74Weight, mean � SD (kg) 57.81 � 12.01Residential area

Large city 29 046 (44.3)Small city 31 626 (48.3)Rural area 4856 (7.4)

Socioeconomic statusHigh 6247 (9.5)High-middle 17 997 (27.5)Middle 31 056 (47.4)Low-middle 8324 (12.7)Low 1904 (2.9)

Dietary behavioursSkipping breakfast ≥5 days/week 18 532 (28.3)Fruits ≥1 times/day 14 995 (22.9)Vegetables ≥2 times/day 19 369 (29.6)Milk ≥1 times/day 18 144 (27.7)Fast food ≥3 times/week 10 734 (16.4)Soft drinks ≥3 times/week 17 740 (27.1)

S. Park et al.

490 © 2018 Dietitians Association of Australia

risk of CVD,42 obesity,43 and cancer.44 In addition, fruitsand vegetables contain vitamins and minerals that preventvarious physical diseases45 and non-restorative sleep.46

Regarding oral health, consuming fruits and vegetables wasfound to reduce the risk of oral cancer by 49 and 50%,respectively.47 However, as noted in a previous meta-analysis,6 fruit and vegetable consumption has an inconsis-tent effect on mental health, so further studies are neededto clarify this relationship.

Adolescents who drank milk once or more per dayshowed higher odds of perceived general health, oralhealth, happiness, and sleep satisfaction. This is compatiblewith previous studies showing that dairy products mightreduce the risk of chronic diseases such as high-bloodpressure,48 type 2 diabetes,17 and metabolic syndrome,49 aswell as improve sleep satisfaction46 and oral health.34 Fur-thermore, low-milk consumption during adolescence leadsto low-bone mass and risk of fracture during adulthood 50.

Although none of the encouraged dietary behaviours hadany specific effects on perceived stress or depression,healthy dietary behaviour in general (i.e. more encourageddietary behaviour and less discouraged dietary behaviour)was associated with better self-perceived physical and men-tal health status among Korean adolescents. The lack of anyobserved benefits of encouraged dietary behaviours onmental health is perhaps because less than 30% of partici-pants met the recommended daily intake of these encour-aged foods. However, further studies are necessary to clarifythe relationship between consuming healthy foods andmental health.

The present study had several limitations. First, across-sectional design was used. Therefore, we can suggestonly associations between dietary behaviours and healthstatus in Korean adolescents, but not causal relationships.Future studies should investigate the longitudinal effectsof dietary behaviours on the health of Korean adolescents.Second, all data were collected through self-reports, sothe student-provided information might be inaccurate.However, self-reports of health status are often consistentwith actual medical records51 and save time for largestudy samples.T

able

2Associatio

nsof

each

dietarybehaviou

rwith

perceivedhealth

status

inKoreanadolescents

Skipping

breakfast

≥5days/week

Fruits≥1

times/day

Vegetables≥3

times/day

Milk

≥2tim

es/day

Fastfood

≥3

times/week

Softdrinks

≥3

times/week

AOR(95%

CI)

AOR(95%

CI)

AOR(95%

CI)

AOR(95%

CI)

AOR(95%

CI)

AOR(95%

CI)

Perceivedgeneralhealth

0.86

(0.82–

0.89

)*1.10

(1.06–

1.15

)*1.29

(1.23–

1.34

)*1.21

(1.16–

1.27

)*0.82

(0.78–

0.86

)*0.90

(0.86–

0.94

)*Perceivedoralhealth

0.84

(0.81–

0.87

)*1.23

(1.19–

1.28

)*1.37

(1.33–

1.42

)*1.14

(1.10–

1.18

)*0.86

(0.82–

0.90

)*0.81

(0.78–

0.85

)*Perceivedhapp

iness

0.80

(0.77–

0.83

)*1.17

(1.12–

1.23

)*1.23

(1.19–

1.28

)*1.11

(1.07–

1.15

)*0.83

(0.80–

0.87

)*0.88

(0.85–

0.92

)*Perceivedsleepsatisfaction

0.84

(0.80–

0.87

)*1.13

(1.08 –

1.18

)*1.16

(1.11–

1.20

)*1.09

(1.04–

1.13

)*0.81

(0.77–

0.85

)*0.85

(0.81–

0.89

)*Perceivedstress

1.23

(1.18–

1.27

)*0.96

(0.92–

1.00

)1.01

(0.98–

1.05

)*0.97

(0.94–

1.01

)1.25

(1.20–

1.31

)*1.17

(1.13–

1.22

)*Depressivemood

1.22

(1.18–

1.27

)*1.03

(0.99–

1.08

)1.01

(0.97–

1.06

)1.03

(0.98–

1.07

)1.42

(1.36–

1.49

)*1.26

(1.20–

1.31

)*

Mod

elswereadjusted

forsex,

scho

olgrade,residentialarea,socioecono

micstatus,andotherdietarybehaviou

rs.

AOR,adjustedod

dsratio

;CI,confi

denceinterval.

*P<0.00

1.

Table 3 Associations of dietary behaviour scores with per-ceived health and health behaviours in Korean adolescents

Dietary behaviour scores(per 1 score increase)

AOR (95% CI)

Perceived general health 1.18 (1.17–1.20)*Perceived oral health 1.23 (1.21–1.24)*Perceived happiness 1.18 (1.17–1.20)*Perceived sleep satisfaction 1.16 (1.14–1.17)*Perceived stress 0.90 (0.89–0.92)*Depressive mood 0.89 (0.88–0.91)*

Models were adjusted for sex, school grade, residential area, andsocioeconomic status.AOR, adjusted odds ratio; CI, confidence interval.*P < 0.001.

Dietary behaviours and perceived health

© 2018 Dietitians Association of Australia 491

In conclusion, despite these limitations, the presentstudy provides an analysis of the association betweendietary behaviours and health status among 65 528 ado-lescents from a nationally representative sample (thusensuring good external validity of the study results). Theresults indicate that discouraged dietary habits have rela-tively detrimental effects on perceived physical and men-tal health in Korean adolescents. Education on healthydietary behaviours is needed for adolescents who skipbreakfast frequently and who regularly consume fastfood and soft drinks.

Funding source

This work was supported by a clinical research grant(No. 2016-03) from the National Centre for Mental Health,Republic of Korea.

Conflict of interest

The authors have no conflicts of interest to declare.

Authorship

SP analysed the data and wrote the manuscript. SR and JHLparticipated in writing the manuscript.

References

1 Fergusson DM, Woodward LJ. Mental health, educational, andsocial role outcomes of adolescents with depression. Arch GenPsychiatry 2002; 59: 225–31.

2 Burt MR. Why should we Invest in Adolescents? Washington,D.C.: Urban Institute, 1996.

3 Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR,Walters EE. Lifetime prevalence and age-of-onset distributionsof DSM-IV disorders in the national comorbidity survey repli-cation. Arch Gen Psychiatry 2005; 62: 593–602.

4 Burdette AM, Needham BL, Taylor MG, Hill TD. Health life-styles in adolescence and self-rated health into adulthood. JHealth Soc Behav 2017; 58: 520–36.

5 WHO. In: Nutrition in Adolescence—Issues and Challenges for theHealth Sector. Geneva: WHO Press, 2005.

6 O’Neil A, Quirk SE, Housden S et al. Relationship between dietand mental health in children and adolescents: a systematicreview. Am J Public Health 2014; 104: 31–42.

7 Jacka FN, Kremer PJ, Berk M et al. A prospective study of dietquality and mental health in adolescents. PLOS ONE 2011; 6:e24805.

8 Oddy WH, Robinson M, Ambrosini GL et al. The associationbetween dietary patterns and mental health in early adoles-cence. Prev Med 2009; 49: 39–44.

9 Hong CH. Current health issues in Korean adolescents. KoreanJ Pediatr 2011; 54: 395–400.

10 Neumark-Sztainer D, Story M, Resnick MD, Blum RW. Corre-lates of inadequate fruit and vegetable consumption amongadolescents. Prev Med 1996; 25: 497–505.

11 Song Y, Park MJ, Paik H-Y, Joung H. Secular trends in dietarypatterns and obesity-related risk factors in Korean adolescentsaged 10–19 years. Int J Obes (Lond) 2010; 34: 48–56.

12 Kim T-H, Choi J-Y, Lee H-H, Park Y. Associations between die-tary pattern and depression in Korean adolescent girls. J PediatrAdolesc Gynecol 2015; 28: 533–7.

13 Lee J, Lee Y. The association of body image distortion withweight control behaviours, diet behaviours, physical activity,sadness, and suicidal ideation among Korean high school stu-dents: a cross-sectional study. BMC Public Health 2016; 16: 39.

14 Korea Centers for Disease Control and Prevention. Korea YouthRisk Behaviour Web-based Survey (KYRBS). Osong: Korea Cen-ters for Disease Control and Prevention, 2017.

15 The Korea Health Industry Development Institute. DietaryGuidelines for Koreans. Osong: Ministry of Health and Wel-fare, 2010.

16 Jacques PF, Tucker KL. Are dietary patterns useful for under-standing the role of diet in chronic disease? Am Soc Nutrition2001; 73: 1–2.

17 Park S, Lee Y, Jang H, Jo M. Violence victimization in Koreanadolescents: risk factors and psychological problems. Int J Envi-ron Res Public Health 2017; 14: 541–52.

18 Park S, Kim Y. Prevalence, correlates, and associated psycho-logical problems of substance use in Korean adolescents. BMCPublic Health 2016; 16: 79.

19 Huang CJ, Hu HT, Fan YC, Liao YM, Tsai PS. Associations ofbreakfast skipping with obesity and health-related quality oflife: evidence from a national survey in Taiwan. Int J Obesity2010; 34: 720–5.

20 Kaneita Y, Ohida T, Osaki Y et al. Insomnia among Japaneseadolescents: a nationwide representative survey. Sleep 2006;29: 1543–50.

21 Keski-Rahkonen A, Kaprio J, Rissanen A, Virkkunen M,Rose RJ. Breakfast skipping and health-compromising behav-iours in adolescents and adults. Eur J Clin Nutr 2003; 57:842–53.

22 Sjoberg A, Halberg L, Hoglund D, Hulthein L. Meal pattern,food choice, nutrient intake and lifestyle factors in The Gote-borg Adolescence Study. Eur J Clin Nutr 2003; 57: 1569–78.

23 Fishbein DH, Pease SE. Diet, nutrition, and aggression. JOffender Rehabil 1994; 21: 117–44.

24 Duclos M, Gouarne C, Bonnemaison D. Acute and chroniceffects of exercise on tissue sensitivity to glucocorticoids. J ApplPhysiol 2003; 94: 869–75.

25 Lopez-Garcia E, Schulze MB, Fung TT et al. Major dietary pat-terns are related to plasma concentrations of markers of inflam-mation and endothelial dysfunction. Am J Clin Nutr 2004; 80:1029–35.

26 Fung TT, Willett WC, Stampfer MJ, Manson JE, Hu FB. Die-tary patterns and the risk of coronary heart disease in women.Arch Int Med 2001; 161: 1857–62.

27 Vaccarino V, Johnson BD, Sheps DS et al. Depression, inflam-mation, and incident cardiovascular disease in women withsuspected coronary ischemia: the National Heart, Lung, andBlood Institute–sponsored WISE study. J Am Coll Cardiol2007; 50: 2044–50.

28 Lavoie KL, Pelletier R, Arsenault A, Dupuis J, Bacon SL. Associ-ation between clinical depression and endothelial functionmeasured by forearm hyperemic reactivity. Psychosom Med2010; 72: 20–6.

29 Chang H-H, Nayga RM. Childhood obesity and unhappiness:the influence of soft drinks and fast food consumption. J Hap-piness Stud 2010; 11: 261–75.

30 Sánchez-Villegas A, Toledo E, de Irala J et al. Fast-food andcommercial baked goods consumption and the risk of depres-sion. Public Health Nutr 2011; 15: 424–32.

S. Park et al.

492 © 2018 Dietitians Association of Australia

31 Zahedi H, Kelishadi R, Heshmat R et al. Association betweenjunk food consumption and mental health in a national sampleof Iranian children and adolescents: the CASPIAN-IV study.Nutrition 2014; 30: 1391–7.

32 Torres SJ, Nowson CA. Relationship between stress, eatingbehaviour, and obesity. Nutrition 2007; 23: 887–94.

33 Schulze MB, Manson JE, Ludwig DS et al. Sugar-sweetenedbeverages, weight gain, and incidence of type 2 diabetes inyoung and middle-aged women. JAMA 2004; 292: 927–34.

34 Moynihan P, Petersen PE. Diet, nutrition and the prevention ofdental diseases. Public Health Nutr 2004; 7: 201–26.

35 �verby N, Høigaard R. Diet and behavioural problems at schoolin Norwegian adolescents. Food Nutr Res 2012; 56: 17231.

36 Bes-Rastrollo M, Sánchez-Villegas A, Gómez-Gracia E,Martínez JA, Pajares RM, Martínez-González MA. Predictors ofweight gain in a Mediterranean cohort: the Seguimiento Uni-versidad de Navarra study. Am J Clin Nutr 2006; 83: 362–70.

37 Shi Z, Taylor AW, Wittert G, Goldney R, Gill TK. Soft drinkconsumption and mental health problems among adults inAustralia. Public Health Nutr 2010; 13: 1073–9.

38 Lien L, Lien N, Heyerdahl S, Thoresen M, Bjertness E. Con-sumption of soft drinks and hyperactivity, mental distress, andconduct problems among adolescents in Oslo, Norway.Am J Public Health 2006; 96: 1815–20.

39 Wiles NJ, Northstone K, Emmett P, Lewis G. “Junk food” dietand childhood behavioural problems: results from the ALSPACcohort. Eur J Clin Nutr 2009; 63: 491–8.

40 Bellavia A, Larsson SC, Bottai M, Wolk A, Orsini N. Fruit andvegetable consumption and all-cause mortality: a dose-responseanalysis. Am J Clin Nutr 2013; 98: 454–9.

41 Wang X, Ouyang Y, Liu J et al. Fruit and vegetable consump-tion and mortality from all causes, cardiovascular disease, andcancer: systematic review and dose-response meta-analysis ofprospective cohort studies. BMJ 2014; 349: g4490.

42 Streppel MT, Ocké MC, Boshuizen HC, Kok FJ, Kromhout D.Dietary fiber intake in relation to coronary heart disease andall-cause mortality over 40 y: the Zutphen study. Am J ClinNutr 2008; 88: 1119–25.

43 Tucker LA, Thomas KS. Increasing total fiber intake reduces riskof weight and fat gains in women. J Nutr 2009; 139: 576–81.

44 Dong J-Y, He K, Wang P, Qin L-Q. Dietary fiber intake andrisk of breast cancer: a meta-analysis of prospective cohortstudies. Am J Clin Nutr 2011; 94: 900–5.

45 Traber MG, Stevens JF. Vitamins C and E: beneficial effectsfrom a mechanistic perspective. Free Radic Biol Med 2011; 51:1000–13.

46 Grandner MA, Jackson N, Gerstner JR, Knutson KL. Sleepsymptoms associated with intake of specific dietary nutrients. JSleep Res 2014; 23: 22–34.

47 Pavia M, Pileggi C, Nobile CG, Angelillo IF. Associationbetween fruit and vegetable consumption and oral cancer: ameta-analysis of observational studies. Am J Clin Nutr 2006;83: 1126–34.

48 Azadbakht L, Mirmiran P, Esmaillzadeh A, Azizi T, Azizi F.Beneficial effects of a dietary approaches to stop hypertensioneating plan on features of the metabolic syndrome. DiabetesCare 2005; 28: 2823–31.

49 Elwood PC, Pickering JE, Fehily AM. Milk and dairyconsumption, diabetes and the metabolic syndrome: the Caer-philly prospective study. J Epidemiol Commun Health 2007;61: 695–8.

50 Kalkwarf HJ, Khoury JC, Lanphear BP. Milk intake duringchildhood and adolescence, adult bone density, and osteopo-rotic fractures in US women. Am J Clin Nutr 2003; 77:257–65.

51 Thomas J III, Paulet M, Rajpura JR. Consistency between self-reported and recorded values for clinical measures. Cardiol ResPract 2016; 2016: 4364761.

Dietary behaviours and perceived health

© 2018 Dietitians Association of Australia 493

ORIGINAL RESEARCH

Characteristics of older adults with diabetes: Whatdoes the current aged care resident look like?

Olivia FARRER ,1 Alison YAXLEY ,1 Karen WALTON2 and Michelle MILLER1

1Department of Nutrition and Dietetics, Flinders University, Adelaide, South Australia and 2Nutrition and Dietetics,School of Medicine, University of Wollongong, Wollongong, New South Wales, Australia

AbstractAim: The prevalence of diabetes in older adults in residential aged care (RAC) is twice that of community dwellingolder adults. Older adults with diabetes have been highlighted as being at high risk of frailty and malnutrition, particu-larly when managed on a therapeutic diet. However, assumptions may be based on clinical presentation of our oldestold which may be at conflict with the clinical presentation of younger older adults. The aim of this retrospective auditwas to identify the characteristics of aged care residents with diabetes, their comorbidities and malnutrition risk.Methods: Residents’ with diabetes paper-based records were audited for demographic data, diet code, medical andmedication history, malnutrition risk screening scores and anthropometric data.Results: A total of 295 residents with diabetes records were audited across 13 sites in South Australia. Youngerolder adults (65–74 years) were generally in a healthy weight range (23–30 kg/m2) or obese (>30 kg/m2) and hadgained weight since admission to RAC; whereas the oldest old were more likely to have lost weight and had a lowerbody mass index range 22.6–28.9 kg/m2. An unexpected finding was the variability between RAC sites in dietarymanagement of diabetes and potential inappropriate dietary management of malnutrition, as indicated by their foodservice diet codes.Conclusions: Findings suggest that younger older adults in RAC are more likely to be overweight which is main-tained over a typical length of stay. Study findings indicate that diabetes management is inconsistent and highlightsthe need for mandated RAC menu guidelines.

Key words: diabetes, diet, older adult, residential aged care.

Introduction

The proportion of older adults over 65 years is growingand is projected to increase to 25% of Australia’s popula-tion in the next 30 years.1 Diabetes, particularly type 2 dia-betes, is more prevalent in older adults (>65 years) and isstrongly associated with obesity, as well as being an inde-pendent risk factor for residential aged care (RAC).2

There is limited literature addressing the clinical presen-tation of older adults with diabetes, however, dietary man-agement of this group has come under much scrutiny overthe last 20 years.3,4 This is particularly as a result of thehigh rates of malnutrition seen in RAC settings, with litera-ture suggesting approximately half of all institutionalisedolder adults are at risk of malnutrition or malnourished.5

Therapeutic diabetic diets have traditionally imposed some

restriction in food choice or macronutrient distribution foroptimal glycaemic and weight control and therefore thiscohort have been particularly highlighted as a vulnerablegroup for malnutrition risk.6

Subsequently, current best practice recommendations forRAC7 have been revised to encourage a standard menu forall residents without dietary restriction, which incorporatesnourishing meals and snacks.3,4,7–9 The recommendationspropose that the liberalisation of low fat and sugar restric-tions will not significantly affect glycaemic control for resi-dents with diabetes.3,4,7 A recent systematic review hasfound the evidence underpinning these recommendationsto be of low quality, with many of the conclusions derivedfrom author consensus.10 Studies frequently cited in sup-port of a liberalised diet are those by Coulston et al.11 andTariq et al.12 which propose older adults with diabetes areat higher risk of poor oral intake and therefore malnutri-tion. However, on examination of their studied cohorts allwere within a healthy weight range (body mass index, BMI,20–25 kg/m2), and a therapeutic diet had no impact onweight change (loss or gain).

Similarly in the wider literature, albeit limited, whichincludes anthropometric data for older adults with diabetes,cohorts are typically described as ‘generally well nourished’13

O. Farrer, BSc Hons Dietetics, APD, LecturerA. Yaxley, PhD, APD, LecturerK. Walton, PhD, AdvAPD, Associate ProfessorM. Miller, PhD, AdvAPD, DeanCorrespondence: M. Miller, Department of Nutrition and Dietetics,Flinders University, GPO Box 2100 Adelaide, SA 5001, Australia.Email: michelle.miller@flinders.edu.au

Accepted January 2018

© 2018 Dietitians Association of Australia494

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 494–499 DOI: 10.1111/1747-0080.12418

or as having BMI’s ranging between 24 and 28 kg/m2.2 Anenergy dense diet has long been recognised as appropriateindividualised nutrition support for malnourished individ-uals. However, provision of this style of diet to older adultswith diabetes who are not malnourished, may pose issuessuch as unwanted weight gain which could impact mobilityas well as chronic hyperglycaemia. Suboptimal glycaemicmanagement has the potential to cause short-term symp-toms such as polyuria, confusion and falls which could eas-ily be mistaken for typical consequences of ageing. In thelonger term, hyperglycaemia can also significantly impact onresidents’ risk of cognitive decline and subsequently theirquality of life.14,15 This is particularly relevant for the cur-rent ageing population or ‘baby boomers’ (characterised asadults born between 1946 and 1965) who are more fre-quently visiting health services for management of cardiovas-cular problems, obesity and type 2 diabetes than everbefore.16 While individualising dietary recommendations isrecommended, it has been recognised that individualisingan institutionalised diet is problematic.17

The aim of this retrospective audit is to contribute to theliterature a more current snapshot of typical characteristicsof RAC residents with diabetes. The study will also explorethe risk of malnutrition in this cohort, weight trends overtypical RAC length of stay (LOS), as well as the presence ofpre-existing comorbidities more aligned with obesity.

Methods

A retrospective audit of resident records was conducted ona convenience sample of n = 295 aged care residents withknown diabetes in 13 aged care facilities, 12 of which weremetropolitan and 1 rural site, all in South Australia. Thesample size was calculated with a 95% CI (�5%). This wasbased on a typical total RAC population18 and approxi-mately 25% of all residents as having diabetes. Thus,requiring a sample size of n = 289.5 individuals for statisti-cal significance and ability to generalise to the wider popu-lation. Aged care organisations with facilities in differentgeographical and socioeconomic areas of the state andrepresenting both metropolitan and regional areas wereapproached for auditing. Data collection included review ofgeneral admission and nursing documentation, current die-tary management codes and malnutrition screening scoresfrom ‘in house’ screening using either the Malnutrition Uni-versal Screening Tool (MUST [BAPEN, Worcestershire,UK]) tool19 or Mini Nutrition Assessment Tool (MNA[Nestle Nutrition Institute, Switzerland]) short form.20 Clin-ical presentation of residents was determined through theaudit of these documents as well as medication charts andgeneral observation and medical notes. A standardised formfor data collection was used, and all data were collectedwithin a 4-month period, on one occasion only and by oneresearcher for consistency. Following ethical approval fromthe Flinders University Social and Behavioural ethics com-mittee, residents with diabetes were identified by nursingstaff and confirmed via their paper-based records providedfor onsite audit.

All residents with diagnosed diabetes were included inthe audit. Written consent from the residents was waivedby the ethics committee to help to eliminate a skew inresults towards only having data for those residents able toconsent. Additionally, no exclusion criteria relating to LOS,diagnosis or permanency of placement was applied for thepurpose of data collection.

Data were collected as de-identified information andincluded demographic information: age (years), gender,date of admission to RACF, comorbidities (based on comor-bidities listed in General Practice Management of type 2diabetes21), all medical history, most recent diabetes bio-chemistry and albumin, and finally a list of medications.The clinical targets used to define optimal management ofdiabetes were based on the International Diabetes Federa-tion global guideline for management of type 2 diabetes inolder adults for HbA1c15 and general practice guidelines fortotal cholesterol:14

i HbA1c: generally 7.0%–8.0% (range: 53–64 mmol/mol)ii Total cholesterol: <4.0 mmol/L

Biochemistry targets were specific to type 2 diabetes dueto the known higher prevalence of type 2 diabetes in RACbut these specific targets could also apply to individualswith type 1 diabetes.15 Physical assessment data includedcurrent and admission weight (kg), height (cm), malnutri-tion screening score (generated either from a MUST orMNA screening tool) and RAC diet code with a BMI(kg/m2) for admission and time of audit being derivedwhen both weight and height were recorded. For the pur-poses of the analysis in this study, the revised BMI healthyrange of 23–30 kg/m2 was used for benchmarking a healthyweight range.

Nutrition scores were generated for all residents by eitherthe facility dietitian or aged care staff during regular screen-ing. The MUST19 or the MNA short form20 was used byfacilities in all audited sites to identify residents who werealready malnourished or at risk of malnutrition; with ascore of ‘0’ or ‘12–14’ representing no risk of under nutri-tion or normal nutritional status; ‘1’ or ‘8–11’ being a mod-erate risk or at risk of malnutrition and a score of ‘2’ or‘0–7’ meaning the resident was at high risk or malnour-ished, respectively.19,20 For statistical analysis, residentsscoring either moderate risk or at risk (1 or 8–11) andmoderate to high risk of malnutrition (2 or 0–7) were col-lapsed into one grouping for comparison against those resi-dents identified as not at risk. Residents without a nutritionscore (n = 37) were those who had either not required areview by the dietitian or who had not been screened in thelast 12 months by the facility care staff and were notincluded in the analysis.

Statistical analysis calculated frequencies (%) and mea-sures of distribution, describing the general characteristicsof older adults with diabetes such as distribution of gender,typical medical and dietary management of residents, andanthropometry. Data were analysed for the whole cohortinitially and also categorised into three older adult agegroups: young old (65–74 years), old-old (75–84 years)and oldest old (>85 years) and presented to determine

Characteristics of older adults with diabetes

© 2018 Dietitians Association of Australia 495

generational differences in clinical presentation. Where thedata are not normally distributed, a Kruskal-Wallis test wasused for comparison across the groups or where data werecategorical a chi-squared (χ2) test of association was used.

Those residents with a nutrition screening score wereincluded for additional analysis (n = 258) of differences incharacteristics between residents not at risk and those atrisk of malnutrition. Not normally distributed data wereanalysed using a Mann-Whitney U test for continuous dataand χ2 test for categorical data. All tests assumed a signifi-cant P-value of <0.05 and were analysed using IBM SPSSstatistics version 22.0 (IBM Corp, Armonk, NY).

Results

Of the 13 RAC included in this study, a total of n = 295residents were identified as having diabetes out of a totalpopulation of n = 1298 residents across all 13 sites provid-ing a prevalence estimate of 23%. Overall, 22% of all resi-dents had either type 1 or type 2 diabetes and their generalcharacteristics are presented in Table 1. Residents with dia-betes were predominantly female (67%, n = 198) whichalthough statistically significant is an expected finding foran ageing cohort (P = 0.026).22 The largest proportion ofall residents with diabetes regardless of gender, were in theoldest old category (≥85 years) representing 51% of allolder adults with diabetes. Median age for the whole cohortwas 85 years (interquartile range, IQR: 80–89 years). Theaverage LOS ranged from 1 to 204 months for all residents(n = 295) with a median LOS of 24 months which was alsosignificantly different across the age groups (P = 0.027). Ofthe whole cohort, 41% of residents (n = 121) were withinan adjusted healthy BMI range (23–30 kg/m2) and 39%(n = 39) were obese (BMI > 30 kg/m2).

Residents were generally within desirable levels forHbA1c (n = 150) with median HbA1c of 6.6 mmol/L, IQR:6.0–7.3, despite significant differences between residentsmanaged by diet or by medication for diabetes (P < 0.001).The remaining residents (n = 145) had no results forHbA1c from the past 12 months at the time of audit. Inaddition, residents with diabetes had close to ideal levelsfor cholesterol (median: 4.2 mmol/L, IQR: 3.5–5.2) with nosignificant differences for biochemistry across the agegroups (Table 1). Of the total cohort (n = 295), 83%(n = 246) residents with diabetes presented with comorbid-ities including raised blood pressure, cholesterol or cardio-vascular disease (CVD).

The young old represented 11% (n = 33) of the cohortand more than half (n = 17, 55%) were managed with oralhypoglycaemic agents (OHAs) or OHAs and insulin (n = 4,13%) as compared to the oldest old (>85 years) who werepredominantly managed by diet alone (n = 90, 60%) asshown in Table 1. BMI ranged from 17.4 to 53 kg/m2, withthe median BMI for the young old being the highest for thewhole cohort at 31 kg/m2 (IQR: 26.5–35). Median BMIshowed a downward trend for the two older generations,with the oldest old (>85 years) reporting a median BMI25.6 kg/m2 (IQR: 22.6–28.9). Conversely, residents’ weight

typically increased between admission and data collection(current weight) for residents aged between 65–74 yearsand 75–84 years, but for the oldest old ≥85 years, weightwas more likely to have decreased since admission, asshown in Table 1. This proved statistically significant whencomparing weight at admission and current weight acrossthe age groups; P < 0.001 for both.

Independent of age, when examining the relationshipbetween weight and malnutrition risk, change in weightwas significantly different for those at moderate to high riskof malnutrition (mean SD: −2.76 � 8.3 kg) and those resi-dents not at risk (2.37 � 7.5 kg), P < 0.001 (Table 2). Thiswas also reflected in a significant difference in BMI betweengroups (P < 0.001). Residents at risk of malnutrition weremore likely to have a median BMI of 24 kg/m2, IQR of20–28. This compared to those not at risk, who had ahigher median BMI of 29 kg/m2, IQR of 25–33.

Of the residents with diabetes who had been screenedfor malnutrition risk (n = 258), 35% (n = 90) were identi-fied as being at risk of malnutrition or malnourished(Table 2). There were significant differences between thosenot at risk and those at risk of malnutrition when compar-ing to individual diet codes or texture modified diets,P < 0.001. But on closer examination of the data, propor-tionally more residents not at risk of malnutrition were ondiets considered restrictive, such as diabetic (n = 94, 56%)and weight reducing (n = 16, 9%), than those residentsidentified as at risk or malnourished; n = 21 (23%) andn = 3 (3%), respectively.

Discussion

Generational differences in our ageing population are likelyto bring about differing clinical presentations in older adultsand new challenges to our health care services, particularlyRAC.23 Currently there are over a quarter of a million olderadults residing in Australian RAC facilities,18 approximatelya quarter of them will have diabetes.2 The largest age groupin RAC is still the oldest old (>85 years)31 who are cur-rently the ‘depression babies’, born in a time of food ration-ing and scarcity. This is likely to be significantly different tothe experience of our younger RAC resident and prospec-tive users of RAC, or ‘baby boomers’; who have had abun-dant access to food and who, it is postulated, will likelypresent to RAC with obesity and comorbidity.16,21,24,25

The findings of this study were consistent with priorstudies’ examining typical BMI for older adults withdiabetes,2,13 and found that based on a median BMI of27.6 kg/m2 it could be interpreted that the population isindeed ‘well nourished’.2,13 There has been some debate asto whether BMI is sensitive enough to determine malnutri-tion risk, particularly where sarcopenia may be present butnot clinically assessed.26 However, a recent study27 evalu-ated applicability of using BMI to measure nutritional sta-tus. The study found that a relationship does exist betweenpoor nutritional status as measured by a low BMI, and anincreased risk of mortality associated with BMI <23 kg/m2

but also >30 kg/m2. Similarly, more recent meta-analyses

O. Farrer et al.

496 © 2018 Dietitians Association of Australia

(2014 and 2015) of BMI and all-cause mortality in olderadults28,29 found that a BMI <23 kg/m2 is associated withincreased mortality risk in the general aged care population.The studies also noted a ‘U’-shaped relationship in the datameaning a BMI >30 kg/m2 was also shown to be associatedwith negative health outcomes.

Stratifying the results by age groups presented an interest-ing finding which was the generational differences in weightand nutritional status. Younger older adults more frequentlypresented with a higher BMI than the cohort median range(median BMI: 29–31 kg/m2), but more importantly main-tained or in some cases gained weight from time of admis-sion over a median LOS of 24 months. This is in contrast tothe oldest old (>85 years) who were more likely to havebeen admitted with a BMI under the median (median BMI:25.6 kg/m2) and who subsequently lost weight over the sameperiod (Table 1). This clinical presentation of the oldest oldis much more indicative of the ‘at risk’ population discussedin the early literature which supported a change in dieteticmanagement and subsequent liberalised diet for olderadults in RAC.11,12 Whereas the clinical presentation ofyounger older adults in this study presents new informationwhich supports opinion that baby boomers will more fre-quently present to services as overweight or obese and with

significant comorbidity.1,16 This study found that the major-ity of the residents (83%) had comorbidities associated withCVD. Therefore, a diet which may encourage unintentionalweight gain and which is not supportive of comorbiditymanagement or diabetes, could therefore be problematic.

Finally, an additional finding of this study concurredwith recent literature proposing that dietary managementof older adults with diabetes is highly variable.30 Butunexpectedly this was not restricted to delivery of a thera-peutic diabetic diet versus a liberalised menu. Residentsidentified as at risk of malnutrition in this audit wouldideally have all been on a high energy/high protein dietcode—or at the very least not on a diabetic or weightreducing diet as per aged care best practice guidelines.7

Results from this audit highlighted that 23% residents atrisk of malnutrition were still being managed on a diabeticdiet; 3% on a weight reducing diet, while 24% remainedon the standard diet code according to facility systems. Intotal, 50% (n = 45) of all residents with diabetes who wereidentified as at risk of malnutrition, or malnourished werenot being managed on a liberalised or fortified diet. Thisagain raises the need for widely endorsed nutrition guide-lines to promote consistent practice and best outcomes forall residents.

Table 1 Residents with diabetes and their characteristics

Total cohort(n = 295)

65–74 years,33 (11%)

75–84 years,111 (38%)

≥85 years,151 (51%) P = <0.05(a)

Gender, n (%)Male 97 (33) 16 (48) 41 (37) 40 (27) 0.026Female 198 (67) 17 (52) 70 (63) 111 (73)

LOS (months), range 1–204 2–125 1–204 1–154 0.027Type 2 diabetes, n (%) 280 (95) 29 (88) 104 (94) 147 (97) 0.061Diabetes management(b), n (%)

Diet alone 31 10 (32) 40 (37) 90 (60)Diet and OHAs 109 17 (55) 57 (52) 54 (36) <0.001Diet and OHAs and/or insulin 149 4 (13) 12 (11) 5 (3)

Weight (kg), median (IQR)Admission 71.1 (61–81) 80.4 (73–97) 73.4 (65–88) 68 (58–75) <0.001Current 72.3 (60–84) 87 (71–90) 76.4 (65–90) 66.6 (57–76) <0.001

Current BMI (kg/m2), median (IQR) 27.6 (23.3–31.4) 31.0 (26.5–35.0) 29.2 (24.1–33.4) 25.6 (22.6–28.9) <0.001Adjusted current BMI (kg/m2), n (%)

<23 60 (20) 3 (10) 17 (17) 39 (29)23–30 121 (41) 11 (35) 41 (41) 73 (54) <0.001

>30 114 (39) 17 (55) 43 (43) 23 (17)HbA1c (%), median (IQR)(c) 6.6 (6.0–7.3) 6.2 (5.6–6.7) 6.8 (6.1–7.7) 6.4 (6.0–7.3) 0.076Total cholesterol (mmol/L)(d),

median (IQR)4.2 (3.5–5.2) 4.0 (2.7–4.9) 4.4 (3.4–5.5) 4.2 (3.5–4.9) 0.427

Albumin (mmol/L), median (IQR) 35 (33–39) 35 (32–37.5) 36 (33–40) 35 (33–38) 0.653Comorbidities (%)(e)

Hypercholesterolemia/hypertension 144 (49) 12 (36) 48 (43) 84 (56) 0.038Cardiovascular disease 200 (68) 22 (67) 79 (71) 99 (66) 0.665

Significant values in bold text.BMI, body mass index; IQR, interquartile range; LOS, length of stay; OHAs, oral hypoglycaemic agents, n, number.(a) Significance between age groups determined via Kruskal-Walls or χ2 tests.(b) Diabetes management missing data: n = 150.(c) HbA1c data: n = 150.(d) Total cholesterol: n = 208.(e) Data for comorbidities was recorded as separate entries and so individuals may be represented twice for hypertension and cardiovascular disease.

Characteristics of older adults with diabetes

© 2018 Dietitians Association of Australia 497

A limitation of this study is the nature of the studydesign which offers only a snapshot of the characteristicsof this subgroup with no rationale for diet code alloca-tion or diabetes management. In addition, the data takenfrom aged care records were variable in detail and fre-quency of entries and biochemistry was not routinelyordered to include HbA1c or not ordered at all. The datahave also not been discussed within the context of whatdifference the geographical and socioeconomic regionsmay have, as this was outside the scope of this study.However, it may be useful to consider in future research.Despite limitations, sufficient data to demonstrate statisti-cal significance were collected and findings are consistentwith prior literature, while also providing the new datapertinent to generational differences in older adults. Alongitudinal design may be more appropriate for morethorough data collection and particularly in determiningcorrelation between malnutrition risk and diet composi-tion and controlling for confounders; while a qualitativestudy to explore resident experience of dietary manage-ment may also be warranted.

In summary, there are significant differences in the pre-sentation of older adults with diabetes across the genera-tions, particularly with regard to weight and BMI both atadmission and after an average LOS of 24 months. Youngerolder adults generally present as more overweight with thisweight increasing throughout their admission rather thandecreasing. Overall, resident characteristics identified in thisstudy are more frequently congruent with healthy weight orobesity than underweight, and residents are more often notat risk of malnutrition or malnourished. This has practicalimplications for using a ‘one diet fits all’ approach for a het-erogeneous population, and where an individualisedapproach for all residents with diabetes is problematic forcurrent food service systems. Future research looking at

menu design in RAC and which addresses the heterogeneousnature of residents and their changing clinical presentationmay be warranted.

References

1 Australian Bureau of Statistics (ABS). Population ProjectionsAustralia, 2012 to 2101. Canberra: ABS, 2013.

2 Fagot-Campagna A, Bourdel-Marchasson I, Simon D. Burdenof diabetes in an aging population: prevalence, incidence, mor-tality, characteristics and quality of care. Diabetes Metab 2005;31: 5S35–52.

3 Niedert K. Position of the American Dietetic Association: liberal-ization of the diet prescription improves quality of life for olderadults in long-term care. J Am Diet Assoc 2005; 105: 1955–65.

4 Dorner B. A liberalized approach to diets for diabetes in longterm care. Director 2002; 10: 132–3.

5 Agarwal E, Marshall S, Miller M, Isenring E. Optimising nutri-tion in residential aged care: a narrative review. Maturitas2016; 92: 70–8.

6 Nathan DM. The diabetes control and complications trial/epi-demiology of diabetes interventions and complications study at30 years: overview. Diabetes Care 2014; 37: 9–16.

7 Bartl R, Bunney C. Best Practice Food and Nutrition Manual forAged Care Facilities, 2nd edn. St Leonards, NSW: Nutrition Ser-vices, Central Coast Health produced document, Chapter 6Tips to Maximise the Nutrition Content of Foods Offered onthe Menu, Northern Sydney Central Coast Area Health, 2012.

8 American Diabetes Association. Translation of the diabetesnutrition recommendations for health care institutions: posi-tion statement. Diabetes Care 2003; 26: S70–2.

9 Richardson TSA. Healthy Eating & Diabetes: A Guide for AgedCare Facilities. The goals of dietary management of diabetes inaged care (pg.8), 1st edn. Queen Elizabeth Hospital. DiabetesCentre South Australia: 2011.

10 Farrer O, Yaxley A, Walton K, Healy E, Miller M. Systematicreview of the evidence for a liberalized diet in the management

Table 2 Typical characteristics for residents with diabetes compared to malnutrition risk score and differences betweengroups

Total (n = 258)Not at risk of malnutrition

(n = 168, 65%)

Moderate to high riskof malnutrition(n = 90, 35%) P-value

Age, years (median, IQR) 85 (80–89) 84 (78–89) 85 (80–90) 0.220Current BMI (median, IQR) 27.6 (23.3–31.4) 29 (25–33) 24 (20–28) <0.001Length of stay, months (median, IQR) 24 (12–46) 24.5 (12–46) 27.5 (13–47) 0.338Weight change (kg) (mean, SD) 0.44 � 8.1 2.37 � 7.5 −2.76 � 8.3 <0.001Resident diet code, n (%)(a)

Standard 62 (24) 40 (24) 22 (24) <0.001Diabetic 115 (45) 94 (56) 21 (23)HEHP 62 (24) 18 (11) 44 (49)Weight reducing 19 (7) 16 (9) 3 (3)

Texture modified diet code, n (%)Soft—Texture A 16 (34) 10 (6) 6 (7)Minced and moist—Texture B 23 (49) 2 (1) 21(23)Smooth pureed—Texture C 2 (4) 0 (0) 2 (2)

Significant values in bold text.HEHP, high energy and high protein; IQR, interquartile range; n, number; OHAs, oral hyperglycaemic agents; WR, weight reducing.(a) Diet code n = 258, of which n = 41 were texture modified.

O. Farrer et al.

498 © 2018 Dietitians Association of Australia

of diabetes mellitus in older adults residing in aged care facili-ties. Diabetes Res Clin Pract 2015; 108: 7–14.

11 Coulston AM, Mandelbaum D, Reaven RM. Dietary manage-ment of nursing home residents with non-insulin-dependentdiabetes mellitus. Am J Clin Nutr 1990; 51: 67–71.

12 Tariq S, Karcic E, Thomas D et al. The use of a no-concen-trated-sweets diet in the management of type 2 diabetes innursing homes. J Am Diet Assoc 2001; 101: 1463–6.

13 Turnbull P, Sinclair A. Evaluation of nutritional status and itsrelationship with functional status in older citizens with diabe-tes mellitus using the mini nutritional assessment (MNA) tool.A preliminary investigation. J Nutr Health Aging 2002;6: 185–9.

14 The Royal Australian College of General Practitioners andDiabetes Australia. General Practice Management of Type 2Diabetes—2014–2015. Melbourne: The Royal Australian Col-lege of General Practitioners and Diabetes Australia, 2014.

15 International Diabetes Federation. Global Guideline for Manag-ing Older People with Type 2 Diabetes. Brussels: IDF, 2013.

16 King DE, Matheson E, Chirina S, Shankar A, Broman-Fulks J.The status of baby boomers: health in the United States—thehealthiest generation? JAMA Intern Med 2013; 173: 385–6.

17 Tinker LF, Heins JM, Holler HJ. Commentary and translation:1994 nutrition recommendations for diabetes. J Am Diet Assoc1994; 94: 507–11.

18 Australian Institute of Health and Welfare. Residential AgedCare in Australia 2010–11: A Statistical Overview. Canberra:AIHW, 2012.

19 Todorovic V, Russell C, Elia M. The ’MUST’ explanatory book-let: BAPEN, 2011. (Available from: http://www.bapen.org.uk/pdfs/must/must_explan.pdf, accessed 1 September 2017).

20 McGee M, Jensen GL. Mini nutritional assessment (MNA):research and practice in the elderly. Am J Clin Nutr 2000;71: 158.

21 The Royal Australian College of General Practitioners. GeneralPractice Management of Type 2 Diabetes: 2016–18. Melbourne:RACGP, 2016.

22 Australian Bureau of Statistics. Mature Age Persons StatisticalProfile: Health. ABS: Canberra, 2004.

23 Forster S, Gariballa S. Age as a determinant of nutritional sta-tus: a cross sectional study. Nutr J 2005; 4: 28.

24 Adams K, Anderson JB, Archuleta M, Smith Kudin J. Definingskilled nursing facility residents’ dining style preferences. J NutrGerontol Geriatr 2013; 32: 213–32.

25 King DE, Xiang J, Brown A. Intake of key chronic disease-related nutrients among baby boomers. South Med J 2014;107: 342–7.

26 Cook Z, Kirk S, Lawrenson S, Sandford S. Use of BMI in theassessment of undernutrition in older subjects: reflecting onpractice. Proc Nutr Soc 2005; 64: 313–7.

27 Miller MD, Thomas JM, Cameron ID et al. BMI: a simple, rapidand clinically meaningful index of under-nutrition in the oldestold? Br J Nutr 2009; 101: 1300–5.

28 Winter JE, MacInnis RJ, Wattanapenpaiboon N, Nowson CA.BMI and all-cause mortality in older adults: a meta-analysis.Am J Clin Nutr 2014; 99: 875–90.

29 Veronese N, Cereda E, Solmi M et al. Inverse relationshipbetween body mass index and mortality in older nursing homeresidents: a meta-analysis of 19,538 elderly subjects. Obes Rev2015; 16: 1001–15.

30 Farrer O, Yaxley A, Walton K, Milte R, Miller M. The ‘diabeticdiet’: a web based survey for determining the incidence, ratio-nale, composition and implications in Australian residentialaged care facilities. J Nursing Home Res 2017; 3: 50–3.

31 Gaskill D, Black LJ, Isenring EA, Hassall S, Sanders F,Bauer JD. Malnutrition prevalence and nutrition issues in resi-dential aged care facilities. Australas J Ageing 2008; 27:189–94.

Characteristics of older adults with diabetes

© 2018 Dietitians Association of Australia 499

ORIGINAL RESEARCH

‘Breaking Barriers, Breaking Bread’: Pilot study toevaluate acceptability of a school breakfast programutilising donated food

Natika DEAVIN,1 Anne-Therese MCMAHON,1 Karen WALTON1 and Karen CHARLTON 1,2

1School of Medicine, University of Wollongong and 2Illawarra Health and Medical Research Institute, Wollongong,New South Wales, Australia

AbstractAim: One in seven Australian schoolchildren do not consume breakfast. School-based breakfast programs assist at-risk children to meet their nutritional requirements and provide cognitive and behavioural benefits, but may result insignificant food costs and waste. The present study aimed to explore acceptability and perceived benefits of a novelfree primary school-based breakfast program utilising donated food.Methods: Process evaluation included quantification of amount of food donated, number of meals provided andnutritional analysis of some of the meal items. Impact evaluation was based on thematic analysis of focus groupsheld with students, parents and teachers to explore their perspectives about the program. Breakfast diaries and hun-ger rating visual analogue scales were used to evaluate students’ breakfast habits at home over fiveconsecutive days.Results: The program saved 14.4 t of food from landfill through conversion into 44 000 meals. One-fifth of childreninterviewed arrived at school without having breakfast at least once per week while one-third of students reportedbeing hungry on arrival at school. Benefits of participation in the program included increased willingness to attendschool, improved alertness and behaviour, as well as creation of an equitable, supportive environment beneficial forlow income or food insecure families.Conclusions: This novel breakfast program based on donated food was widely accepted by students, teachers andparents, providing benefits beyond the mere provision of food. It provides a model for school-based interventions tocombine breakfast programs with sustainable food production approaches.

Key words: breakfast, food waste, nutrition, primary school, program evaluation.

Introduction

Primary schools provide an ideal setting to address healthand health-promoting behaviours in young children. Amajor focus of school-based nutrition intervention programsis to provide a nutritious breakfast before school. Fifteen percent of Australian schoolchildren skip breakfast each daywhich equates to one in seven children missing this meal.1

This figure is concerning as breakfast assists children to meettheir nutrient requirements.2,3 Children who skip breakfastdo not make up for the missing nutritional benefits through-out the rest of the day. This places the children at anincreased risk for nutritional deficiencies3 as well as

adversely impacts their academic achievement.2,4,5 Childrenmost at risk of arriving at school without having had break-fast are often from lower socioeconomic families whoexperience food insecurity,1,2,4 and are further disadvan-taged through a higher risk of poor behavioural andcognitive outcomes at school.6

School breakfast programs are known to be beneficial forchildren’s behavioural and cognitive learning4 because ofthe increased provision of essential nutrients and the expe-riential nature of the programs, which aids in food skill andknowledge development.7 However, while there are signifi-cant benefits associated with school breakfast programs,they may impact negatively on the environment. One studyhas reported that almost half of the foods served duringbreakfast programs are thrown away, contributing to bothprogram cost, and environmental impacts with methane gasgeneration from food waste in landfill.8 Thus, evaluation ofbreakfast programs should consider totality of food provi-sion, cost and waste.

Other reasons for evaluating novel school breakfast pro-grams include the effect on dietary intake of students, their

N. Deavin, BND, APD, DietitianA. McMahon, PhD, APD, Senior LecturerK. Walton, PhD, AdvAPD, Associate ProfessorK. Charlton, PhD, AdvAPD, RPHNutr, Associate ProfessorCorrespondence: K. Charlton, School of Medicine, University ofWollongong, Wollongong, NSW 2522, Australia. Tel: 02 4221 4754.Email: karenc@uow.edu.au

Accepted August 2018

© 2018 Dietitians Association of Australia500

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 500–508 DOI: 10.1111/1747-0080.12478

cognition and concentration in class and overall behaviour.Provision of breakfast has also been linked to improvedhealth with respect to body weight status,9 and better nutri-tional intake.10 In addition, social psychological outcomes,such as improved academic performance,11 reduction inabsenteeism and an overall improved nutritional status havebeen identified with breakfast program delivery.9,12,13 Prac-tical factors that should also be considered with respect tosustainability of breakfast programs include: total food cost,integration of the program into school policy, parental per-ception of the program and dependency on a key person toobtain food donations and champion the continuation ofthe service.14–16

Breaking Bread, Breaking Barriers is a breakfast programdesigned to address two priority issues: poor breakfast con-sumption before school for children attending a regionalpublic school and the large amounts of food waste gener-ated in today’s society. The program was delivered by anot-for-profit organisation founded in 2012, and it is novelbecause it collected and utilised donated food from super-markets and small local businesses that would otherwisehave been discarded, delivering it to children using amobile food van. ALL Sustainable Futures received a grantfrom the Environmental Trust and Environmental Protec-tion Authority to purchase a refrigerated van utilised forfood transportation, storage and preparation.

The aim of the present pilot study was to explore accept-ability and perceived benefits of a novel free primaryschool-based breakfast program ‘Breaking Bread, BreakingBarriers’ utilising donated food. This included examining itseffect on food waste, stakeholders’ perceptions of the pro-gram and its effect on students’ food security, school atten-dance and behaviour and their nutrition-related knowledge,attitudes and practices.

Methods

The Breaking Bread, Breaking Barriers breakfast program wasconducted at a public primary school that is situated in theIllawarra region of NSW. This school was selected becauseof its location in an area of socioeconomic disadvantage(ranked 481 out of 2568 regions in NSW in 2011).17,18

The breakfast program began at the school in term 4 of2016 and ran for two consecutive terms. The same three tofour volunteers ran the breakfast program at the schooleach week, to ensure stability and familiarity for thestudents.

Food was prepared and provided to all participating stu-dents, families and teachers, to avoid potential stigmatisa-tion around free breakfast provision for economicallydisadvantaged families, on Friday morning each weekbefore school. Breakfast was consumed close to the proxim-ity of the food van in a suitable open space. Remaining foodscraps were fed to the school’s chickens or composted andused in the school’s food garden.

Evaluation of the pilot breakfast program interventionincluded process and impact evaluative methods, informedby the social cognitive model (SCM).19,20 The SCM

acknowledges an individual’s behaviour is multifaceted andhas interrelated influences inclusive of biology, cognitiveand affective constructs, as well as socio-environmental bar-riers and enablers. Hence, the study design explored per-sonal perspectives (cognitive and affective constructs)through qualitative enquiry utilising focus groups with indi-vidual stakeholder groups (students, parents and teachers)as well as accounted for socioeconomic factors (barriers andenablers), such as identifying SES levels, enabling exposureto breakfast meal preparation, to gain an understanding ofkey influences on behaviour. To demonstrate the integrativeframework for analysis of indicators included in the evalua-tion, a LOGIC model21 is shown in Figure 1. The presentstudy did not extend to outcome evaluation.

The research team performing the evaluation of the pro-gram consisted of three academic dietitians, an Honoursresearch student and a research assistant. For the processevaluation, objective measures were collected throughoutthe program delivery regarding the amount and type offood received from the not-for-profit partner. This includedrecording of the type and quantity of donated food(i.e. saved from landfill), weighed using used a commercialWedderburn scale. The number of meals made using thedonated produce was based on the weight of cooked itemsdivided by 270 g, used as a proxy serve size for a child’s‘meal’. Many students had multiple ‘meals’ at a single break-fast. In the case of a composite meal using multiple foods,as in the case of a bacon and egg roll, the number of mealswere calculated according to the limiting factor, that is,number of bacon slices available. The number of bacon andegg rolls can be calculated by the number of bacon slicesleftover at the end of the breakfast session. The final num-ber factored into ‘waste’ estimation how many wanted chil-dren requested either only bacon or egg. Then, numberconsumed was multiplied by the weight of a bacon roll. Forfruit, the amount donated was weighed before and afterbreakfast. The amount of food provided for compost wascalculated as leftover produce after the breakfast(in weight). No food went to landfill as the compost wasgiven to school chickens. Process evaluation also includednutritional analysis of the breakfast program using recipesprovided from volunteer staff and nutritionally analysing anestimated serve of food using FoodWorks, Version 21(Xyris Software, Brisbane, Australia).22

Impact evaluation assessment drew upon qualitativeapproaches through focus groups with three groups of keystakeholders to explore their perspectives about the pro-gram.23 During term 3 of 2017, following the initial deliv-ery of the breakfast program, students, teachers and parent/guardians (program recipients) were invited to participatein focus groups. Teachers who attended the breakfast pro-gram were invited as they were well placed to observeimpacts of the program on children’s behaviour during andpost the breakfast. In order to recruit participants for focusgroups and interviews, a participant information sheet, flyerand consent form was sent home with each child from year3 to 6. Children who returned parental consent forms wereincluded in the study sample. Parents were also invited to

Evaluation of a school breakfast program

© 2018 Dietitians Association of Australia 501

attend focus groups using the same communication strategyand a $25 gift voucher was offered to parents as remunera-tion for their time.

Two trained facilitators conducted focus groups asrecommended by Swift and Tischler23 using a semistruc-tured facilitators guide tailored for each target group. Theyexplored experiences of the program, perspectives on theappropriateness as well as opinions on options for futureprograms. Generic questions were as follows: How often didyou go to the breakfast program? What are your thoughts aboutthe program? What are your thoughts on a nutrition educationprogram for (students/teachers/your children)? Do you haveany suggestions or comments about the breakfast program?

Additional questions for each of the stakeholder groupswere as follows:

Students: What do you like at the breakfast? Is there any-thing you do not like about the breakfast? Did you try a foodyou had not before at the breakfast? Who do you go to thebreakfast with? What do you know about the program? Has thebreakfast program changed what you eat for breakfast and, ifyes, how?

Teachers: What are your thoughts on choosing Friday forthe breakfast program? Did you see a difference in the behav-iour of students after the program on a Friday?

Parents: Is there anything you would change about thebreakfast? Did you notice any changes in your children whenthe breakfast program was running?

Focus groups were held exclusively with stakeholdergroups, with one each comprising teachers and parents(both held late afternoon, after school hours). Because thechildren were primary school-aged and the researchers didnot want to introduce stigma nor distract from classroomtime, information was obtained from children using a mod-ified focus group approach. In collaboration with the schoolprincipal, class activities were organised for one lesson

period across years 3–6. A fruit-tasting session was con-ducted with students in these classes. Thereafter, childrenwhose parents had consented to their inclusion in the studywere invited to have a discussion in one area of the class-room with the facilitator, while their peers completed food-related class activities. These included memory games, wordsearches and picture identification, all related to fruits andvegetables.

Focus groups were voice recorded, transcribed verbatim,coded using content analysis24 and further categorised intothemes that emerged, with the support of the NVIVO the-matic analysis software version 10 (NVivo qualitative dataanalysis software; QSR International Pty Ltd., 2012).25 Theprimary researcher (KC) immersed herself in the data andconducted a number of iterations as recommended by Sal-daña.26 To ensure trustworthiness of final categories andthemes, these were discussed and agreed through consensuswith the other researchers (AM, KW and ND) as recom-mended by Swift and Tischler.23 Exemplar quotes wereidentified to authenticate findings.

Breakfast record templates were provided as part of theimpact evaluation to children in six classes at the school.Students filled out these records each school morning forfive consecutive days during the first lesson of the day,under the supervision of class teachers. They were asked torate their hunger/satiety using a validated ‘Teddy Bear’visual analogue scale, with scores ranging from 1 to 5 indi-cating, if they were hungry or full, respectively.27 Responseswere collated by the research team and summarised overthe 5-day period for numbers of responses within each ofthe five categories.

All data were de-identified and stored in lockabledrawers in the researchers’ offices. Ethical approval wasgranted by the University of Wollongong Human ResearchEthics committee (HE17/186) and the Department of

INPUT IMPACT OUTCOMES

•

•

•

•

•

•

•

•

•

•

•

•

•

•

•

ACTIVITIES OUTPUTS

•••

•

•

Figure 1 LOGIC model for evaluation of breakfast program.

N. Deavin et al.

502 © 2018 Dietitians Association of Australia

Education Ethics (SERAP No. 2017169). The study wasperformed in accordance with these ethical standards andall participants gave informed consent prior to inclusion inthe study.

Results

Process evaluationDonated food and waste: Table 1 provides a summary of

the amount of food that was obtained, donors, direct costof the program, volunteer hours and food waste prevented.It was estimated that the program utilised 14.4 t of donatedfood that would otherwise have gone to landfill.

Foods provided: Foods provided at the breakfast programvaried significantly each week because of the variability inthe foods donated/collected and included a variety of fruitsand vegetables, cereal-based foods, such as pancakes and hotcross buns, protein-based meals, such as frittatas and Frenchtoast, as well as full-cream milk and banana smoothies. Someof the foods that children reported as being most frequentlyconsumed were assessed for energy and protein content. Theanalysis was conducted based on an ingredients list providedby the volunteers and an estimated average serving size.While actual intake was unable to be obtained and calculatedfor the students, a typical intake at the breakfast could beexpected to have provided 250–2000 kJ and 1–20 g of pro-tein. The children commonly consumed a variation of theitems and could therefore have had higher energy and pro-tein intakes than those suggested.

Impact evaluationBreakfast records: Fifty-four students from 3 years 3/4

classes completed the breakfast records for five school days.Five of these records were incomplete and excluded.Despite many efforts, only three classes returned the com-pleted breakfast records, which are summarised in Table 2.On one-third (32%) of occasions over a recording period ofa week, children came to school hungry (determined bycombining hunger levels 1 and 2), with more than halfof these children reported being ‘very hungry’ (hungerlevel 1). The types of food that children reported consum-ing for breakfast were as follows: bread (n = 66 occasions);

unknown cereal type (n = 22); highly refined cereal such asfruit Loops (n = 44); high fibre, low sugar cereal (n = 46);fruit (n = 34); yoghurt and milk (n = 11); protein-basedmeals such as eggs (n = 11); snack foods such as biscuits(n = 10); other foods such as chicken (n = 10) and noth-ing (n = 16).

Of the breakfast reports from 49 children over 5 days,there were 16 occasions (7%) where no breakfast had beenconsumed. The types of food that the children reported con-suming on the days that they reported being hungry or veryhungry is shown in Table 3. On 31% of the occasions thatchildren reported being ‘very hungry’, they had come toschool without having had breakfast while on 40% of theseoccasions children had consumed highly refined cereals.This hunger may be attributed to inadequate amounts ofbreakfast or the poor nutritional quality of the cereal. Highlyrefined cereals are often low in protein and fibre; thus, limit-ing satiety-inducing properties of these foods. Over a periodof a week, one-fifth (n = 10; 20%) of children who providedbreakfast records were not having breakfast on at least oneweekday, 4% of children (n = 2) were not having breakfaston 2 days, and a further 4% (n = 2) on three or more days.This number is higher than the national average of studentswho skip breakfast (14%) highlighting this to be a group athigh risk of food insecurity.1

Twenty-one schoolchildren (11 male; 10 female) agedbetween 8 and 12 years (school years 3–6) participated inthree focus groups. Of the six teachers who participated in asingle focus group, three were male and three female whilethe two parent attendees were both female. The programwas well attended, with all teachers and almost all childrenreporting that they had attended every breakfast. Of the par-ents interviewed, one had never attended, while the otherattended every session. All participants believed there to beadequate amounts and varieties of foods provided and weremore varied compared to breakfast options served at home.

Six main themes were apparent from the qualitativefocus groups, including Breakfast benefits, Skill acquisition,Environmental impacts, Food security impacts, Program signifi-cance and Student engagement. Breakdown of each themeand sub-themes are shown in Figure 2, and identifies theme

Table 1 Details of the school breakfast program

Activity Results

Program duration Two school termsFood recovered/

recycled14.4 t of food saved Coles

Helensburgh $4320 in wastecharges

Food waste 140 kg of food scraps fed to chickensOther waste Cardboard boxes used for mulchFood donation

sourcesColes Helensburgh; Crinis Fruit

Market; Albion Park Poultry; TerryStreet Meat and Seafood

Volunteers Volunteers contributed 4708 hoursProject cost $230 340 including consumables,

lost salary and van

Table 2 Data collected from student breakfast recordsusing Teddy Bear27 visual analogue scale to rate hunger andsatiety

Hunger level

Number ofoccasions (n = 49participants)(a)

(n = 245)Percentageof occasions

1. Really hungry 47 192. Quite hungry 31 133. Not hungry,

not full76 31

4. Quite full 53 215. Very full 38 15

a Each participant reported on 5 days (n = 49 ×5 responses = 245).

Evaluation of a school breakfast program

© 2018 Dietitians Association of Australia 503

relevance according to stakeholder group. Comments madeby participant groups often overlapped, so findings weretherefore collectively combined to demonstrate themes.

Theme 1: Breakfast benefitsBreakfast benefits encompass additional effects of the

program incorporating known benefits related to in-class behaviour—focus and alertness to broader

psychosocial outcomes including confidence, energy andschool attendance. In particular, focus and alertness wereidentified as an additional benefit. It was noted that stu-dent’s focus around school classwork increased on thedays of the breakfast program. Students themselvescommented that breakfast helped them to be calmerand less ‘silly’. The teachers affirmed these commentsand most importantly, were able to identify that their

Table 3 Foods eaten on occasions when children reported being hungry (scores 1 and 2) using Teddy Bear27 visual analoguescale

Hunger score 1 (very hungry) Hunger score 2 (quite hungry)

Number of times Percentage of occasions Number of times Percentage of occasions

Breads 2 4 5 16Highly refined cereal 19 40 9 29High fibre, low sugar cereal 9 19 7 22Fruit 0 0 1 3Protein 0 0 3 10Yoghurt 0 0 0 0Milk 1 2 1 13Snack foods 1 2 4 13Other 0 0 0 0Nothing 15 31 1 3

Breakfast program

Student Engagement

Breakfast benefits

Environmental impacts

Food security impacts

Program significance

Eq

uit

y

Fo

od

was

te

man

agem

ent

Skill acquisition

Expan

ded

foo

d

experien

ce

Focus and Alertness

Figure 2 Thematic analysis of stakeholder perceptions related to the breakfast program. ( ) Students; ( ) teachers;( ) parents.

N. Deavin et al.

504 © 2018 Dietitians Association of Australia

students were more focused in class on the day of thebreakfast program.

‘They were definitely more settled in the class and it’s theones that don’t get the breakfast you noticed obviouslythe biggest difference in [are] absolutely more settled,[and] more focused’.—Teacher

School attendance/arrival and excitement: Studentsreported that on the days of the breakfast program, theywere more likely to get to school early and it was some-thing to look forward to all week.

‘I’ve been excited every morning at the breakfast becausewhenever I woke up, I asked my mum to get to schoolreally quick so I can just get there and I’d rate it fivestars’.—Student (year 3/4)

Teachers also mentioned that throughout the durationthat the program was running, their student’s attendancewas almost 100%.

‘Very rarely on Fridays we had anyone away... we actu-ally knew they were really sick’.—Teacher

Both parents reflected that their children were moreenthusiastic about school on Friday (breakfast programday) and frequently demonstrated positive behaviours suchas getting ready more quickly than on other school days.

Theme 2: Skill acquisitionSkill acquisition incorporated learning experiences includ-

ing knowledge development, expanded food experience and foodskill development through the various learning opportunitiesoffered by the breakfast program.

Knowledge development: Students spontaneously couldreport that breakfast can increase their energy levels and providethem with fuel for the day ahead, facts which they attributed tolearnings provided by the program. Student’s confidence alsoseemed to increase, in terms of children being able to imple-ment the behaviour themselves. One student remarked:

‘They’re showing us how easy it is just to make yourbreakfast in the morning... they made it easily for us sowe should be able to make it easily for ourselves’.—Stu-dent (year 5/6)

The teachers also noted that the program generated anawareness of extended choices in terms of what the chil-dren can have for breakfast before school. This was recog-nised as a particularly useful outcome given the school’sclose proximity to a fast food outlet, where schoolchildrencommonly obtained takeaways. Children in the groupsstrongly expressed positive responses with respect to theimportance of breakfast for the overall health.

Foods tried: The program enabled students to taste andexperience new foods, with some reporting that foods firsttried at the breakfast were now consumed on a regular basisoutside of school.

‘I’ve actually changed my habits a lot. I used to hatebacon and eggs... and then I tried it and now I’ve chan-ged my tastebuds’.—Student 2 (year 5/6)

‘I did try some... passionfruit, I eat it like every daynow’.—Student (year 3/4)

‘I started baking quiches at dinner time and my kids loveit, even with spinach which they’ve never eaten so it’sdefinitely allowed us to experiment more with differenthealthy foods at home’.—Parent

The teachers observed that children were trying newfoods, particularly those that they did not recognise. Parentsalso reported observing the benefits of the tasting opportu-nities for their children, with one parent reporting beingable to experiment with meals at home, as the childrenwere more willing to try different foods.

Food skill development: While the program was notdesigned for the students to assist in food preparation, somestudents elected to arrive early to offer assistance with foodpreparation. The students reported feeling being highly moti-vated by being allowed to assist in the preparation and oftenencouraged their friends to join them. The teachers alsoobserved the willingness of the students to help out.

‘There’s a lot of kids who [didn’t have] the breakfast pro-gram but helped with it, assisted, and they were the firstto go “When’s it come back” because they loved prepar-ing’.—Teacher

This highlights the likelihood of the students acceptingand enjoying a cooking program within the school.

Theme 3: Environmental impactsThis theme incorporates social effects related to aspects,

such as school community engagement and food waste manage-ment. All participants identified the social aspect of the break-fast program. Each group commented on the atmosphereand the interactions that the breakfast program provided.

‘It was really exciting. You got to meet your friends reallyearly before school and you could play around. Now onschool days and Fridays, we have to wait up at the officeand you can’t do anything’.—Student (year 5/6)

Teachers mentioned how the social environment pro-vided a platform for enhanced positive social interactionsbetween the students, teachers and parents, noticing par-ents interacting and engaging within the school environ-ment. One parent mentioned that it was an enjoyableatmosphere, observing students, parents and teachersengage positively together.

Theme 4: Food security impactsFood access: All groups were openly able to identify that

some of the school families had insufficient funds to pur-chase adequate amounts of food, and that the breakfast pro-gram was useful to address this. While low income was nota commonly discussed topic by the children, it was evident

Evaluation of a school breakfast program

© 2018 Dietitians Association of Australia 505

which children had experience within their own or otherfamilies with limited income to purchase foods.

‘[It] helped the families because everything was free andmost families here are really poor’.—Student 1 (year 3/4)

‘Lots of people need it because my family didn’t haveenough money for a while so I got to have free break-fast’.—Student 2 (year 3/4)

Teachers highlighted that the program was beneficial forlow-income families and commented that they felt wholefamilies would attend to ensure they would be able to get afull meal for the day. Parents also mentioned the benefit ofchildren being able to access a variety of foods, which theyotherwise would not be able to do because of restrictedhousehold incomes.

Equity: Teachers discussed how the program createdequality and reduced the stigma associated with provisionof free breakfast for only the disadvantaged within theschool.

‘There was no [food] limit... they could have what theyneeded... Those who obviously have issues with havingfood at home, enough food, they knew they could comehere and they were on the same wavelength as everyother student that came here’.—Teacher

Theme 5: Program significanceThis theme incorporated four aspects: affinity, health-

related impacts, consistency of program delivery and enhance-ment for future.

Affinity: Students overall response was wholeheartedlypositive in terms of loving the program, the food provided,the social interaction of the experience and the program isprovided free of charge. The parents affirmed that the chil-dren’s enthusiasm towards the program, stating that thewhole program had been a success.

‘I know my kids pretty much loved everything... all thekids down there were happy, they were socialising, theywere eating, they were walking around trying new foodswhich they probably might not do at home’.—Parent

Health-related impacts: A number of references weremade to the potential health impacts of the program. Stu-dents believed the program to be healthy, particularly incomparison to what they are familiar with.

‘You can go to school and have a better, healthier break-fast instead of what you have at home’.—Student(year 5/6)

The teachers believed the program to be beneficial inreducing the amount of processed and packaged foods thechildren ate, and that the children were able to try differentfoods that they might not otherwise have had the opportu-nity to try at home. The teachers had some concernsregarding the nutritional quality of some of the foods pro-vided; however, parents believed the program to be healthy

and provide a variety of foods that are not often servedat home.

‘[It] definitely showed that your average healthy breakfastdidn’t have to be like what the norm is in a household...[the children] were amazed what kind of [foods could beeaten]’.—Parent

Consistency of program delivery: Every child involved inthe focus groups reported that they were very upset thatthe program had been discontinued.

‘I don’t want to come to school anymore on Fridays’.—Student 1 (year 3/4)

Parents and teachers confirmed that the children weresaddened over the discontinuation of the program, fre-quently questioned why it was ceased, and wanted to haveagreement about consistency for its delivery into the future.

Enhancement for future: Notably, the significance of theprogram was underpinned by additional suggestions forfuture delivery. These included options for modification tothe quality of the foods provided and inclusion of seasonalfoods. Teachers reported that children’s skill levels in cook-ing were not strong, most likely reflecting their limitedengagement with cooking at home. They suggested that theaddition of a cooking program would be highly beneficial toimprove the children’s knowledge and skills in this regards.This concept was further developed through the suggestionof a school recipe book to assist the students with healthymeal planning and increase independent cooking skills.

Theme 6: Student engagementStudent engagement included elements that enveloped a

number of characteristics that underpinned their interest inbeing overall engaged with the program. These included ele-ments such as program frequency, attendance, breakfast at home,food preference, food quality and variety, and foods offered.

Discussion

The present study reports on the evaluation of a novel freebreakfast program offered to primary school-aged childrenattending a school in an area of socioeconomic disadvantagethat used donated food to deliver meals on-site using amobile food van. Key findings were that the breakfast pro-gram was enthusiastically received by children, teachers andparents and resulted in improved school attendance, betterbehaviour of children, and feelings of empowerment by chil-dren to try new foods. It is well recognised that childrenwho consume breakfast have improved focus, alertness andconcentration during the morning.2 These factors also con-tribute to a less disruptive classroom environment which canenhance learning opportunities.2 The improved school atten-dance rates recognised by the teachers on the days that thefree breakfast program was offered is consistent with findingsfrom other studies, as reviewed by Rampersaud et al.3

Another additional benefit may be an experiential learn-ing opportunity in line with the vicarious learning identified

N. Deavin et al.

506 © 2018 Dietitians Association of Australia

in Bandura’s SCM.20 Outcomes from the focus groups indi-cated that children learnt new information about nutrition,while consuming and trying new foods at the breakfast pro-gram. Tasting experiences have been reported to be themost effective way to improve food preferences and aremore effective than simply visualising a food.28 The morediverse and unique experiences are, the higher the likeli-hood of new learnings.28 In the current evaluation, all chil-dren expressed strong positive perceptions regarding theimportance of breakfast following participation after twoterms, a finding that has been reported by others.29 In linewith the social cognitive theory,20 interaction between peersand their food environment play a significant role in theacceptance and uptake of foods30,31

Equity was an important aspect related to acceptability ofthe current breakfast program. Bailey-Davis et al.32 reportedthat peers would bully students participating in free breakfastprograms and consequently these children preferred to go toschool hungry. To overcome such stigma, all members of theschool community were invited to attend the breakfast pro-gram. Students from low-income families are much morelikely to skip breakfast, resulting in unhealthy intakes withlower amounts of fruit and vegetables33 and thus are mostlikely to benefit from breakfast programs. It is noteworthythat one-third of students in the present study were going toschool hungry, with one-fifth of children skipping breakfastbefore school at least once per week.

In terms of improvements to the program, some teachersexpressed concern over the type of foods being providednot always being considered healthy, such as bacon and eggrolls on weeks that these foods had been donated. Moststudies suggest that any form of breakfast is important forchildren’s health, behaviour and overall well-being.3 Bentonet al.34 found, however, that a lower glycaemic index(GI) breakfast was related to increased attention and mem-ory scores, compared with those consuming higher GIfoods. It is therefore suggested that, first and foremost, abreakfast-consuming behaviour is established, and then thenutritional quality of the breakfast foods can be evaluated.3

The pilot breakfast program that was evaluated wasinformed by the SCM in its approach, and was aligned withthe Ottawa Charter actions of strengthening communityaction and creating a supportive environment.35 It strength-ened community action by empowering the school commu-nity through the continued provision of recovered food bythe volunteer program to target food security, hunger andfood waste. A supportive environment was provided within aschool setting where a socially acceptable meal could beaccessed without stigmatisation associated with breakfastprograms being for low-income individuals.32 The programalso aligned with two goals of the United Nations’ sustainabledevelopmental goals, namely ‘Zero Hunger’ and ‘ResponsibleConsumption and Production’.36 As well as being wellreceived by the target audience, the program model has beenshown to be effective in its ability to redistribute food other-wise destined for landfill to needy families.

The primary limitation of the present study is that it wasquasi-experimental and did not include a control school.

Another limitation is that a not-for-profit organisation ranthe program and therefore determined its starting date andunexpected completion. This meant that pre-interventiondata could not be collected prior to commencement andearly discontinuation prevented collection of all processdata. Food provision data were provided by the partnerorganisation and could therefore not be verified. The pro-gram was also limited to one session per week, potentiallylimiting the impact of the program.

Because of the composition of the focus groups and onlyone setting, findings need to be explored in other settingsfor confirmation. The low numbers of parents who partici-pated in the evaluation also limit the depth of the findingsfrom this stakeholder group. The school principal identifiedwhich classes could be accessed for the focus groups withchildren who may have resulted in some additional per-spectives not being captured, and was a participant in theteachers’ focus group, which may have resulted in limitedopen commentary from the teachers.

In conclusion, a free school breakfast program based ona model that encourages distribution of recovered foodfrom local providers has been shown to be highly acceptedby schoolchildren, teachers and parents at a primary schoollocated in a geographical area of high socioeconomic disad-vantage. The novel breakfast program encompassed inclu-sivity of all students and overcame perceived stigmaassociated with poverty and the need for a breakfast pro-gram. Importantly, the program provides a model whichcould be potentially scaled up and delivered to similarlydisadvantaged schools, for maximum benefit of school-children in the Illawarra region.

Funding source

A $9000 University of Wollongong Community Engage-ment Grant (CEGS) was utilised for this research to assistwith data collection and transcription as well as develop-ment of nutritional resources post-intervention. All Sustain-able Futures designed and delivered the breakfast program,and partnered with the research team for an evaluation ofthe acceptability of the program.

Conflict of interest

The authors have no conflict of interest to declare.

Authorship

KC, the lead researcher designed the project. ND (student),KW and KC designed the data collection and analysis ofthe research and ND drafted the manuscript. AM providedexpert guidance with qualitative data collection, analysisand interpretation. All authors edited the manuscript. Allauthors are in agreement with the manuscript and declarethat the content in this manuscript has not been publishedelsewhere.

Evaluation of a school breakfast program

© 2018 Dietitians Association of Australia 507

References

1 Australian Bureau of Statistics. Census at School Australia. Can-berra: Australian Bureau of Statistics, 2011. (Available from:http://www.abs.gov.au/websitedbs/CaSHome.nsf/Home/2012+CensusAtSchool+Summary+Data, accessed 7 March 2017).

2 Adolphus K, Lawton CL, Dye L. The effects of breakfast onbehavior and academic performance in children and adoles-cents. Front Hum Neurosci 2013; 7: 425–31.

3 Rampersaud GC, Pereira MA, Girard BL, Adams J, Metzl JD.Breakfast habits, nutritional status, body weight, and academicperformance in children and adolescents. J Am Diet Assoc2005; 105: 743–60.

4 Hochfeld T, Graham L, Patel L, Moodley J, Ross E. Does schoolbreakfast make a difference? An evaluation of an in-school break-fast programme in South Africa. Int J Educ Dev 2016; 51: 1–9.

5 Petralias A, Papadimitriou E, Riza E et al. The impact of aschool food aid program on household food insecurity. Eur JPublic Health 2016; 26: 290–6.

6 Rodgers YV, Milewska M. Food assistance through the schoolsystem. J Child Poverty 2007; 13: 75–95.

7 Kolb DA. Experiential Learning: Experience as the Source ofLearning and Development, Vol. 377. Englewood Cliffs, NJ:Prentice Hall, 1984. (Available from: http://academic.regis.edu/ed205/Kolb.pdf, accessed 7 June 2017).

8 Byker CJ, Farris AR, Marcenelle M, Davis GC, Serrano EL.Food waste in a school nutrition program after implementationof new lunch program guidelines. J Nutr Educ Behav 2014; 46:406–11.

9 Deshmukh-Taskar PR, Nicklas TA, O’Neil CE, Keast DR,Radcliffe JD, Cho S. The relationship of breakfast skipping andtype of breakfast consumption with nutrient intake and weightstatus in children and adolescents: the National Health andNutrition Examination Survey 1999–2006. J Am Diet Assoc2010; 110: 869–78.

10 Briefel RR, Wilson A, Gleason PM. Consumption of low-nutri-ent, energy-dense foods and beverages at school, home, andother locations among school lunch participants and nonpar-ticipants. J Am Diet Assoc 2009; 109: S79–90.

11 Kleinman RE, Hall S, Green H et al. Diet, breakfast and aca-demic performance in children. Ann Nutr Metab 2012;46: 24–30.

12 Saini V, Johari A, Potaliya M. Analytical review on effect ofbreakfast consumption on cognitive processes of children. AnnAgri Bio Res 2014; 19: 628–9.

13 Murphy JM, Pagano ME, Nachmani J, Sperling P, Kane S,Kleinman RE. The relationship of school breakfast to psychoso-cial and academic functioning: cross-sectional and longitudinalobservations in an inner-city school sample. Arch Pediatr Ado-lesc Med 1988; 152: 899–907.

14 Spruance LA, Harrison C, Brady P, Woolford M, LeBlanc H.Who eats school breakfast? Parent perceptions of school break-fast in a state with very low participation. J Sch Health 2018;88: 139–49.

15 Hearst MO, Shanafelt A, Wang Q, Leduc R, Nanney MS. Alter-ing the school breakfast environment reduces barriers to schoolbreakfast participation among diverse rural youth. J Sch Health2018; 88: 3–8.

16 Wilson DK, Griffin S, Saunders RP et al. Formative evaluationof a motivational intervention for increasing physical activity inunderserved youth. Eval Program Plann 2006; 29: 260–8.

17 Australian Bureau of Statistics. 2011 Census Data—QuickStats.Canberra: Australian Government, 2012. (Available from:

http://www.abs.gov.au/websitedbs/censushome.nsf/home/historicaldata2011?opendocument&navpos=280, accessed21 March 2017).

18 Community Childcare Cooperative NSW. Socio Economic Indexesfor Areas. Merrickville, Sydney: Community Childcare Coopera-tive, 2013; 27. (Available from: http://ccccnsw.org.au/pdf/socio-economic-index-for-areas.pdf, accessed 3 March 2017).

19 UNICEF, The Centre for Disease Control and Prevention. TheSocial Cognitive Model A Framework for Prevention. Sydney:UNICEF, 2014; 4. (Available from: http://www.cdc.gov/violenceprevention/overview/social-ecologicalmodel.html,accessed 3 March 2017).

20 Bandura A. Social Foundations of Thought and Action: A SocialCognitive Theory. Englewood Cliffs, NJ: Prentice Hall, 1986.

21 Cooksy LJ, Gill P, Kelly PA. The program logic model as anintegrative framework for a multimethod evaluation. Eval Pro-gram Plann 2001; 24: 119–28.

22 Foodworks Version 8 [Computer Software]. Qld, Australia:Xyris Software, 2015.

23 Swift JA, Tischler V. Qualitative research in nutrition and die-tetics: getting started. J Hum Nutr Diet 2010; 23: 559–66.

24 Harris JE, Gleason PM, Sheean PM, Boushey C, Beto JA,Bruemmer B. An introduction to qualitative research for foodand nutrition professionals. J Am Diet Assoc 2009; 109: 80–90.

25 NVIVO Qualitative data analysis software. Version 10. QSRInternational Pty Ltd., 2014.

26 Saldaña J. The Coding Manual for Qualitative Researchers, Vol. 1.London: Sage, 2015; 333.

27 Bennett C, Blissett J. Measuring hunger and satiety in primaryschool children. Validation of a new picture rating scale. Appe-tite 2014; 78: 40–8.

28 Pollitt E. Does breakfast make a difference in school? J Am DietAssoc 1995; 95: 1134–9.

29 Birch LL, McPhee L, Shoba BC, Pirok E, Steinberg L. Whatkind of exposure reduces children’s food neophobia? Lookingvs. tasting. Appetite 1987; 9: 171–8.

30 Smith SL, Cunningham-Sabo L. Food choice, plate waste andnutrient intake of elementary-and middle-school students par-ticipating in the US National School Lunch Program. PublicHealth Nutr 2014; 17: 1255–63.

31 King SC, Weber AJ, Meiselman HL, Lv N. The effect of mealsituation, social interaction, physical environment and choiceon food acceptability. Food Qual Prefer 2004; 15: 645–53.

32 Bailey-Davis L, Virus A, McCoy TA, Wojtanowski A, VanderVeur SS, Foster GD. Middle school student and parent percep-tions of government-sponsored free school breakfast and con-sumption: a qualitative inquiry in an urban setting. J Acad NutrDiet 2013; 113: 251–7.

33 Moore GF, Murphy S, Chaplin K, Lyons RA, Atkinson M,Moore L. Impacts of the primary school free breakfast initiativeon socio-economic inequalities in breakfast consumptionamong 9–11-year-old schoolchildren in Wales. Public HealthNutr 2014; 17: 1280–9.

34 Benton D, Maconie A, Williams C. The influence of the glycae-mic load of breakfast on the behaviour of children in school.Physiol Behav 2007; 92: 717–24.

35 World Health Organization. The Ottawa Charter for Health Promo-tion. Malaysia: WHO, 2017. (Available from: http://www.who.int/healthpromotion/conferences/previous/ottawa/en/index1.html,accessed 3 March 2017).

36 United Nations. Sustainable Development Goals. United Nations,2017. (Available from: http://www.un.org/sustainabledevelopment/hunger/, accessed 31 January 2017).

N. Deavin et al.

508 © 2018 Dietitians Association of Australia

ORIGINAL RESEARCH

Communicating health—Optimising young adults’engagement with health messages using social media:Study protocol

Catherine LOMBARD,1 Linda BRENNAN,2 Michael REID,3 Karen M. KLASSEN,1 Claire PALERMO ,1

Troy WALKER,1 Megan S.C. LIM,4 Moira DEAN,5,6,7 Tracy A. MCCAFFREY1 and Helen TRUBY 1

1Department of Nutrition, Dietetics and Food and 6School of Public Health and Preventive Medicine, MonashUniversity, School of 2Media and Communication and 3Economics, Finance and Marketing, RMIT University, 4BurnetInstitute and 7School of Population and Global Health, University of Melbourne, Melbourne, Victoria, Australia; and5School of Biological Sciences, Institute for Global Food Security, Queen’s University Belfast, Belfast, UK

AbstractBackground: Obesity is a global health problem. Understanding how to utilise social media (SM) as a platform forintervention and engagement with young adults (YAs) will help the practitioners to harness this media more effec-tively for obesity prevention.Aim: Communicating health (CH) aims to understand the use of SM by YAs, including Aboriginal YAs, and in doingso will improve the effectiveness of SM strategies to motivate, engage and retain YAs in interventions to reduce therisk of obesity, and identify and disseminate effective ways for health professionals to deliver obesity preventioninterventions via SM.Methods: The present study describes the theoretical framework and methodologies for the CH study, which isorganised into four interrelated phases, each building on the outcomes of preceding phases. Phase 1 is a mixedmethods approach to understand how YAs use SM to navigate their health issues, including healthy eating. Phase2 utilises co-creation workshops where YAs and public health practitioners collaboratively generate healthy eatingmessages and communication strategies. Phase 3 evaluates these messages in a real-world setting. Phase 4 is thetranslation phase where public health practitioners use outcomes from CH to inform future strategies and to developtools for SM for use by stakeholders and the research community.Discussion: The outcomes will include a rich understanding of psychosocial drivers and behaviours associated withhealthy eating and will provide insight into the use of SM to reach and influence the health and eating behaviours of YAs.

Key words: Aboriginal young adults, healthy eating, obesity prevention, social marketing, social media, youngadults.

Introduction

Globally, 39% of adults aged 18 years and over were over-weight or obese in 2014.1 In young adults (YAs), defined as

those aged between 18 and 24 years,2 overweight or obesityhas risen from 9% in 1995 to 39% in 2014.3 Young Aborig-inal and Torres Strait Islander Australians (hereafter referredto as Aboriginal people) have a higher prevalence, withalmost 60% (59%) of women aged 18–24 years being over-weight or obese.4 Obesity is a well-established risk factorfor metabolic and psychological ill health.5,6

Potentially modifiable health behaviours, in particulareating and physical activity, underpin weight gain.7–9 Eatingrelated behaviours of greatest concern in this target agegroup include: (i) low fruit and vegetable intake10,11 and(ii) high intake of sugar-sweetened beverages.3,12,13 This ishappening with a backdrop of increased concern aboutbody image and utilisation of more extreme dieting behav-iours.14,15 Behavioural barriers to healthy eating for YAsinclude: (i) a general apathy and lack of motivation towardstheir dietary choices; (ii) widespread availability ofunhealthy foods; (iii) lack of access and relative expense of

C. Lombard, PhD, Associate ProfessorL. Brennan, PhD, Professor of AdvertisingM. Reid, PhD, Professor of Economics, Finance and MarketingK.M. Klassen, PhD, Research FellowC. Palermo, PhD, Associate ProfessorT. Walker, MClinChiro, Research AssistantM.S.C. Lim, PhD, Deputy Program Director, Research FellowM. Dean, PhD, ProfessorT.A. McCaffrey, PhD, Senior LecturerH. Truby, PhD, Head of DepartmentCorrespondence: H. Truby, Department of Nutrition, Dietetics andFood, Monash University, Level 1 264 Ferntree Gully Road,Melbourne, VIC 3168, Australia.Email: helen.truby@monash.edu

Accepted May 2018

© 2018 Dietitians Association of Australia 509

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 509–519 DOI: 10.1111/1747-0080.12448

healthy foods; (iv) social norms where the unhealthy diet ofpeers plus peer influences support the expected consump-tion of unhealthy foods; (v) lack of time and facilities toplan, shop, and prepare foods and (vi) lack of knowledgeand skills to prepare healthy foods.16–19 For some youngAboriginal Australians, an additional challenge is findingthe balance between an easily available Westernised diethigh in processed foods and obtaining access to a more tra-ditional dietary pattern which incorporates the food knowl-edge from older generations.20–22

The rate of weight gain during the young adulthoodperiod is the fastest across adulthood.23 A period in whichcreating an identity and learning cooking skills take place17

this group are particularly vulnerable to food industry mar-keting.24 Interventions designed by health professionals andsocial marketers can be problematic if they do not ‘connect’with the target audience.25 Social marketing is the use ofcommercial marketing tools and techniques, such assegmentation and targeted marketing activities, to addresssocial issues.26 The use of SM for marketing purposes is

widespread in commercial applications.27,28 However, theuse of SM in social marketing is still in its infancy, withevolving intervention designs and little research that exam-ines its efficacy.29–31 While there is a high level of interestin using SM for nutrition/dietary interventions and commu-nications there remains limited understanding of the con-nection between online engagement and behaviour change.This research study seeks to examine these issues and deter-mine the value and effectiveness of SM for reaching YA. Aglossary of terms (see Table 1) is provided to assist with theunderstanding of marketing language used throughout thispaper.

Social media (SM) is defined as any web-based commu-nications channel dedicated to community-based input,interaction, content-sharing and collaboration. Types of SMinclude websites and applications dedicated to forums,microblogging, social networking, social bookmarking,social curation, and wikis.32 Globally, SM use has reached2.8 billion users and continues to grow.33 Ninety-eight per-cent of Australian YAs use the Internet and of these, 95%

Table 1 Glossary of terms used in this protocol

Term Definition

Co-creation When two or more people create something together, collaboratively and in agreement with eachother about desired outcomes. Note: this is not co-production whereby the ideation may occuroutside the group producing the artefact

Co-design Co-design is a form of design that actively involves a variety of stakeholders in the design process.Its roots are embedded in theories of participatory design

Consumersegmentation

Market segmentation is the process of dividing a market of potential customers into groups, basedon different characteristics. The segments created are composed of consumers who will respondsimilarly to marketing strategies and who share traits such as similar interests, needs, or locations.The idea behind segmentation is to create and resource different marketing strategies for differentgroups of consumers

Design thinking Design thinking is a method adopted by designers to collaboratively solve complex problems. Designthinking is a process of collaborative systemic reasoning about the issue, the end user andpossible outcomes; considering what is viable and feasible given the circumstances

Digital ethnography Digital ethnography describes the process and methodology of doing ethnographic research in adigital space. The digital field site is sometimes comprised of text, video or images and mayinclude social interactions

Living and eatingfor healthsegment (LEHS)

These segments will be defined based on the outcomes of Phase 1 and evaluated throughout theproject. Short descriptive segmentation personas will be developed to aid program development

Online conversations An OC is a multi-way dialogue between participants in an internet environment. It is informal,unstructured and dialogic (not mono-logic) in nature. It involves both listening and answeringand develops over a period of time. It is not an online chat or interview

Social marketing Social marketing seeks to develop and integrate marketing concepts with other approaches toinfluence behaviours that benefit individuals and communities for the greater social good. It seeksto integrate research, best practice, theory, audience and partnership insight, to inform thedelivery of competition sensitive and segmented social change programs that are effective,efficient, equitable and sustainable: Consensus definition International Social MarketingAssociation34

Social media Websites and applications that enable users to create and share content or to participate in socialnetworking

Wicked Problems The term was first coined by Rittel & Webber35 and is widely used to describe an issue (e.g. obesityprevention) that is socially complex, with many interdependencies and no universal solution. Forsocial marketers, the inherently conflicting consumer and stakeholder behaviours that are difficultto define represent a ‘Wicked Problem’

26

C. Lombard et al.

510 © 2018 Dietitians Association of Australia

use the Internet to access social networking sites (SNS).36

SNSs provide a platform for engaging with young Aborigi-nal Australians whose SM use is high.37 Influencing andreaching YAs using traditional channels of health promotionand education campaigns are increasingly challenging.38,39

Thus SM could be acceptable tool to reach and engageeffectively with this target audience.40–43

A significant advantage of using SM is the opportunity touse customised messages for specific sub-groups or con-sumer segments in a way that has not previously availableto health practitioners.44 Social marketers know that toimpact and engage effectively with consumers prior seg-mentation of the target market is required in order to tailorspecific messages and delivery platforms for sub-groupswithin the population.45 Consumer segmentation is knownto be a strategy to improve effectiveness of social marketingcampaigns,46 including for addressing obesity15 and alcoholconsumption.47 However, segmentation by demographiccharacteristics alone may not be as effective as applyingpsycho-behavioural segmentation schemes typically definedby behaviours, attitudes, knowledge, opinions, goals or life-styles.45,48 Psycho-behavioural consumer segments aremore indicative of how and why people actually behave inthe way they do and their likely reactions to interventionsand communication efforts.49 In the context of this study,segmenting YAs into groups that share similar lifestyles,including attitudes, opinions and SM use, means that socialmarketers are better able to develop interventions. Movingbeyond the “one size fits all” approach, consumer segmenta-tion enables a focus on targeted message content and mediachannels which the empirical literature highlights as beingcritical to behaviour change.50,51

Engagement with YAs in SM interventions, adopting aprocess of co-design and co-creation is essential.52 CHenables YAs to participate directly in the design of a SMcampaign, by contributing to the development of campaignmessages (a bottom-up process). This includes determina-tion of the best mix of SM, and sharing messages in realtime with friends, family and networks.41 Using a co-creation approach draws on social marketing principles andallows the research team and stakeholders to connect morepersuasively, effectively integrate the ‘voice of the customer’within change-related communications and intervention,and have a larger societal impact.42 As using SM for healthresearch is relatively new, we still are learning which chan-nels and types of messages to use to reach our target audi-ence, and how to evaluate audience engagement and theeffectiveness of campaign messages. This research protocolfor CH describes the process of a social-marketing-SMintervention design aiming to affect healthy eating in allYAs, especially, Aboriginal Australians.

CH is an interdisciplinary, multi-stage–multi-methodstudy that draws on academics and professionals from com-munications, business and social marketing, nutrition anddietetics, Aboriginal community health organisation mem-bers and public health practitioners.53 Additionally, stake-holders supporting this project are not-for-profitorganisations and an Australian government health

department. Throughout the CH protocol the term ‘indus-try partners’ describes not-for-profit organisations and pub-lic health practitioners who are actively involved in healthpromotion campaigns.

CH is a four-year study comprised of four phases(Figure 1). Phase 1 is a formative phase that aims to gain adeep understanding of YAs’, including AboriginalAustralians’, psychosocial characteristics and behavioursrelated to accessing healthy eating content on SM, as wellas defining Living and Eating for Health Segments (LEHSs).The results of Phase 1 inform Phase 2, where stakeholdersand segments of YAs will co-create healthy eating campaignstrategies and content. In Phase 3, co-created content willbe evaluated in a real-world setting using novel engagementand retention tools. Phase 4 is the translation phase inwhich the outcomes from previous phases are used todevelop tools that can be applied and useful for both stake-holders and the research community.

The overarching aims of ‘CH’ are to:1 understand the how YAs, especially Aboriginal YAs,

engage and utilise SM in relation to health and healthyeating,

2 improve the effectiveness of SM strategies to motivate,engage and retain YAs, especially Aboriginal YAs, ininterventions to reduce the risk of obesity, and

3 identify and disseminate effective ways to deliver theseSM interventions for the target group.

Methods

CH will use a multi-stage mixed methods design54

grounded in the integrative model of behaviour change(IMBC) (Figure 2).55 There are several paradigms whichhave been used to describe and understand behaviourchange. These models range in scope from rational cogni-tive models of individuals making active choices through tobehavioural ecological systems models of social changewhere the individual may not even be aware that they aremaking a choice.26 A model that incorporates the middleground between autonomous consumer decision-makingand systems thinking, introduced by Fishbein and Yzer,55

which updates earlier rational choice models56–59 but alsoincludes interactions within the decision-making environ-ment (i.e. the behavioural ecosystem).

The IMBC theory suggests a higher probability that abehaviour will be performed if one has the intention to per-form the behaviour, has the necessary skills and abilitiesrequired, and there are few environmental or other con-straints. A strong intention to perform the behaviour is inturn influenced by attitudes, perceived norms and self-efficacy relating to behaviour. In CH, the behaviour ofinterest is eating behaviour.5 The protocol recognises themultiple environmental compartments in the obesogenicenvironment.60 In addition to learning about the impact ofthese compartments on YAs, CH seeks to understand indi-vidualised aspects of behaviour change in relation toSM. While IMBC theory is in use within the health domain,there is a paucity of knowledge regarding the specific

Communicating health study protocol

© 2018 Dietitians Association of Australia 511

Phase 1 Phase 4 Phase 2 Phase 3

Think Tank

(design)

Think Tank

(development)

Think Tank

Implementation

Online conversations

(insight into social

media use and

application) *4 weeks

n=200

Survey

instrument

development

Online survey

n=2000

HERLS

Co-creation of

campaigns

(Wicked Problems

Studio)

Online conversations

Evaluation & evolution

of co-created

campaign

n=100

Engagement &

effectiveness

evaluation

Refinement and

testing

esp. engagement &

retention

Real social

media

campaign

Observe and

measure outcomes

* 6 months

Strategies to

optimise campaigns

Social media and

health metrics

Refined concepts, strategies

Measure effectiveness

Immediate scale up

Key

Industry

& govt

partners

Data

collection

& analysis

Investigator

activity

Outcome

Think Tank

Evaluation

Translation

Tools to measure

effectiveness

(esp. engagement &

retention)

Figure 1 Communicating health overview (LEHS, living and eating for health segments).

Figure 2 Integrated model of behaviour change (IMBC, adapted for this study).

C. Lombard et al.

512 © 2018 Dietitians Association of Australia

influences on behaviour—especially precursors of decision-making, such as exposure to prior multiple interventions,SM usage and social motivations for healthy eating. Hence,deeper examination of context and lived experiences isrequired before designing interventions to promote health-ier lifestyles.61 To this end, a multi-method iterativeapproach was designed including a variety of disciplinesand approaches ranging from digital ethnography (Phases1, 2 and 3) health promotion (Phase 1) through designthinking (Phase 2) towards participatory action researchusing stakeholders and community members to design andtest interventions (Phases 3 and 4). The CH study is there-fore described as a sequence of steps and which iterativelybuild on the outcomes of each other. Data triangulationenhances the transferability and dependability of findings.CH incorporates six data collection procedures incorporat-ing both qualitative and quantitative data.

Phase 1—Think Tank 1 Design: The Think Tanks (TTs)are a series of workshop style round table engagementswith partners and experts. TT members are drawn fromgovernment health organisations, non-government organi-sations, not-for-profit organisations, as well as academiccontent and process matter experts. The first TT will be aworkshop for project ideation, research design and scopingpriorities of industry partners. Subsequent TTs occur ineach phase (see Figure 1). TT-2 will focus on developmentof SM campaigns that will be relevant and applicable to thepartner organisations. The focus of TT-3 is to ensure thatcampaigns based on the research and co-creation work-shops will be implementable and pragmatic in execution.TT-4 will evaluate the project, campaigns and participate indefining translation.

Online conversations: Using the principles of digitalethnography,62 online conversations (OCs) involve the useof a powerful online community methodology (taking placein specifically created virtual lounge rooms) used globallyby major companies to obtain insights into consumerbehaviour.63 OCs generate rich insights that go beyond thecapabilities of traditional research approaches such as focusgroups and face-to-face interviews as participants in theOCs converse over a longer period of time responding toquestioning from group moderators on lifestyle, media useand health-related behaviours. CH Phase 1 will be an OCswith YAs discussing a series of 12 health- and eating-relatedtopics, hosted and moderated by an independent marketingresearch agency.

Participants and recruitment: YAs (18–24 years) using SMat least twice daily and living in Australia are eligible to par-ticipate in this phase and 200 will be recruited by anAustralian Market & Social Research Society-certified fieldhouse. This sample size will provide a rich data pool forunderstanding eating-related behaviours and is based onprevious work using similar methodology.47 Participantswill be assigned to one of four communities based on theirage and predisposition to health messages as determined bytheir responses to an online questionnaire about their atti-tudes towards their health.

Once assigned to a community, participants will enterinto an extended conversation about health and SM, reflecton their experiences and observations and have the oppor-tunity to respond to the insights of other participants andto prompts guided by the online moderators. It is likelythat Aboriginal young people from urban areas may berecruited to the OC, however to ensure adequate data areavailable, up to 30 young Aboriginal adults will be recruitedseparately through Aboriginal Community ControlledHealth Organisations (ACCHOs) across Victoria. Theseadditional regional and rural participants will be recruitedto supplement the online (mainly urban) participant pool.A modified version of the OC described above will be con-ducted with this group and facilitated by an Aboriginalresearcher (TW). The researchers acknowledge the nuancesin Aboriginal people’s construction of the concept ofhealth,64,65 engagement in health promotion and in Englishcommunication66 and the likely need for different commu-nication messages. The researcher (TW) will determinewhether the conversations should be held online or as face-to-face focus groups or interviews.

Analysis: Researchers will use thematic analysis to exam-ine the qualitative data collected from participants’ discus-sions and identify themes using a constant comparisonapproach.67 Data will be in the form of text responses toquestions and uploaded images, including memes, videoclips and selfies. Common themes associated with eatingand/or health behaviour, attitudes and motivation will bedrawn out and used as a basis for considering likely LEHSs.Additionally, comparative analysis will be used on a contin-ual basis to compare people and each group to check thatLEHS types are appropriately represented by the data.

Survey development: The survey instrument will be devel-oped based on the findings of the OCs and the antecedentsof behaviour change.55 The interdisciplinary research teamwill evaluate the survey items, including those from existingvalidated surveys to ensure they are representative of eachscale’s domain (i.e. face validity).68 Where necessary, origi-nal scales will be developed using procedures suggested byDeVellis.69 The survey will be piloted (n = 30) to ensureconvergent and discriminant validity between variables.70

Exploratory factor analysis will be undertaken within eachof the model components in order to finalise stable fac-tors.71 Confirmatory factor analysis will be undertaken toexamine sample level psychometric properties, that is,individual-item reliabilities, convergent and discriminantvalidity.72 The outcome will be a robust instrument for usein the online survey. The process of development andtesting of the instrument follows a structured path73,74

and involves instrument development and refinement.Depending on the results from the OCs with young Aborig-inal adults, the same questionnaire may be used or mayneed to be modified. The team will work with Aboriginalhealth partners to assure validity for Aboriginal Australians.

Online survey questionnaire: This stage aims to evaluatethe LEHS developed in the OCs and further define thetarget group segments in relation to the IMBC.55

Communicating health study protocol

© 2018 Dietitians Association of Australia 513

Participants and recruitment: A field house will recruit2000 participants, including Aboriginal YP for an onlinesurvey, based on predetermined criteria such as age, gen-der, location to ensure rural and remote participants arerecruited. This methodology is an adaption of that usedsuccessfully in our recent alcohol and lifestyles research.47

Data analysis: Relationships will be hypothesised prior tothe analysis based on the outcomes of the OCs. Regressionand structural equation modelling using smartPLS375 willbe used to examine the strength and direction relationshipsbetween intentions, attitudes and values, lifestyle, SM useand behaviour. Cluster analysis will be used to develop anddescribe population segment profiles. In using structuralequation modelling to investigate the relationships, it islikely that some elements from the behaviour change modelwill be modelled using latent variables while others will beobserved directly.71 As a result of this procedure, the LEHSwill be defined and associations between LEHS and specificvariables of the IMBC can be examined.

Phase 1 outcomes—Healthy eating-related lifestyle segments:The outcomes from Phase 1 will be a rich understanding ofYAs, including Aboriginal Australians’ attitudes, beliefs andbehaviours related to health communication viaSM. Variation in uptake and acceptability of social market-ing messages according to LEHS will be assessed. At theend of Phase 1, the research will have identified the mosteffective channels, the message tone (e.g. positive vs nega-tive, use of intense emotional appeals) and content (images,videos, websites, applications) with greatest potential forfurther development for each target segment.

Phase 2—TT development: The TT of experts drawn fromour partner organisations and their networks will reviewour learnings from Phase 1, which includes the Phase1 communities’ feedback on existing health promotioncampaigns. Based on Phase 1 results, the TT experts willsuggest evidence-based, obesity-prevention (focusing onhealthy eating), target messages for the different LEHS.These suggestions will be used in the following co-creationactivity.

‘Obesity: the Wicked Problem’—co-creation studio: A designstudio workshop where YAs studying a variety of disci-plines ranging from health to digital media drawn from dif-ferent universities will participate. The interdisciplinaryteams will be briefed together on ‘Obesity: the WickedProblem’.49 Facilitators will use design-thinking tech-niques76 (see Figure 3) over a 3-week period to generate

content strategies and potential messages which resonateand are pertinent for obesity prevention. The teams will besupported by industry expert partners, designers, contentand other experts as required for ideation and creationactivities.

Following the design studio workshops, the researchteam will select co-created campaigns to be evaluated andevolved using an online content co-creation activity,77

including an effectiveness survey. The selected campaignswill be uploaded to a purpose built website and SM seedingwill occur through the partners’ websites and SM platformsidentified from Phase 1. Partners in this sense will be allthose contributing to the project, including the universities,project partners and industry. During the evaluation, thewebsite and SM channels will be publicly available, whichwill allow participants in Phases 2 and 3 to organically cre-ate and share content.

Participants recruitment and procedure—evaluation andevolution stage: A new online community will be recruited(n = 100), similar to the procedure described in Phase1. Community members will complete a screening ques-tionnaire developed from Phase 1 which will segment theminto separate groups (LEHS). Using an iterative process,participants from each segment will be engaged in furtherco-creating content (such as memes, life hacks and videos).As the materials are discussed, debated and generated bythe users, the evolved (co-created) outcomes will be evalu-ated according to their value and worth for each segment.The user generated outcomes will be evaluated in an onlinesurvey to the same participants (n = 100) which will alsoproduce analytics for engagement and effectiveness of theco-created SM campaigns.78 Young Aboriginal adults willbe included in the same co-creation activity and will berecruited by the Aboriginal researcher (TW).

Data analysis for online co-creation evolution and evaluationsurvey: Using mixed-methods data analysis, the online con-versations will capture qualitative information such as recallof messages and the strength of negative or positive feelingstowards the materials. The research team will assess engage-ment by having participants view, respond to and shareimages, videos, text, containing newly developed healthmessages and evaluate SM channel metrics. Responses willbe recorded on motivation, interest and personal relevance.The approach to analysis of all data forms will be herme-neutic, assessing the phenomena as they arise within thecommunity.79 The aim of the conversations and online

Figure 3 Generic design thinking process.

C. Lombard et al.

514 © 2018 Dietitians Association of Australia

co-creation phase is to understand engagement and to iden-tify potential elements of successful SM activities. CH willalso assess how the target group manages multiple riskbehaviour messages (e.g. alcohol, drugs and sexual health)as well as healthy eating and how they assimilate or priori-tise multiple messages.

The evaluation survey will be quantitative in nature andthe same principles of analysis will be applied as thatdescribed in Phase 1. The aim is to measure engagementand to establish metrics for evaluating effectiveness of co-created outcomes, thereby building on the outcomes of theearlier qualitative stages of Phase 2.

Phase 2 outcomes—Healthy eating messages and strategiesfor working with YAs: The outcomes of Phase 2 will includea variety of health message content for each LEHS segment,including Aboriginal YAs for use in Phase 3. Outcomes willbe a deep understanding of effective SM content creationdesigned to connect with YAs. The design-thinking ‘WickedProblems’ Studio is unique to CH and a description of pro-cedures involved will be an important adjunct for partnersworking with YAs and SM in health-related areas. Anotherimportant outcome of Phase 2 will be an understanding ofwhat works and what does not work when it comes to SMinterventions aiming to influence healthy eating behavioursin YAs.

Phase 3—Implementation: Using the outcomes from Phase2, the research team and partners will evaluate the materialsand select content ideas and strategies for implementation.The TT participants will be invited to contribute theirexpertise to evaluate the content and research outcomes tocome up with a series of potential ‘best’ solutions. Industrypartners will be invited to adapt these ideas into their owncampaigns using the skills of the ‘Wicked Problems’ Studio

teams. It is expected that at least two concepts will becomereal SM campaigns associated with industry or not for profitpartner organisations.

Real-world SM campaign: In order to assess the campaignelements, real-world industry partners will adapt the con-cepts and strategies identified from Phases 1 and 2 and con-duct a SM campaign launched from their own sites, linkingwith the CH website. The campaign will incorporate con-tent and message design (e.g. memes, life hacks), seeding ofrelevant SM platforms (e.g. Facebook, Instagram), andtracking of outcomes (e.g. shares, likes). Mobile socialapplications may be developed depending on the outcomesof Phases 1 and 2. Fundamental to the design of SM cam-paigns is the setting of baseline objectives and tracking theoutcomes over time. Figure 4 summarises the steps in SMcampaign development. The industry campaign partnerswill be assisted with design and technical advice by theexperts and students who participated in the ‘Wicked Prob-lems’ Studio, as well as members of the research team.

Participants and recruitment: Subsequent to campaignlaunch, relevant participants, including a group of youngAboriginal Australians, will be invited using a modifiedrespondent-driven sampling approach. This will apply asystematic approach to identification and recruitment ofhard to reach populations and previous participants will beinvited to participate in this new phase. Participants will beasked to recruit others in their social network using a snow-balling technique.80 On consenting to participate in theintervention, participants will complete the online LEHSinstrument (developed in Phase 1) to identify their segmentprofile and complete the questionnaire used in Phase 1. Amarket research agency (see Phase 1) will provide the initialparticipant pool with incentives for participation provided

Figure 4 Social media campaign development.

Communicating health study protocol

© 2018 Dietitians Association of Australia 515

to others. It is important that a group of related social net-works (affinity groups) is established to ensure authenticsocial interactions between participants.

Procedure: Project partners will send messages via onlinesocial networks who will be invited to ‘share’ the campaign.No identifiable information is collected; however, the mar-ket research agency will passively follow URLs to determinewhat happens to the messages even if participants do noth-ing. De-identified data will also be collected by the partnersfrom their SM sites and provided to the research team foranalysis. CH will then systematically test single and combi-nations of channels with content matched to the target seg-ment. All SM channels will link to the CH test websitewhich will contain pages for each LEHS segment andinclude: (i) evidence-based health information; (ii) anopportunity to participate in a simple engaging activity, forexample, watch a video/game and (iii) the opportunity toopt-in to the SM campaign aimed at healthy eating behav-iours. The opt-in option is important to establish levels ofengagement with the campaign, it being the final categoryof commitment to the campaign message. In addition tostandard web metrics, engagement will be measured on ascale of low (e.g. failure to click on the link to the website)to high engagement (e.g. signing up for the CH communityoutside the incentivised market research process). Partici-pants will continue to receive the full SM intervention for6 months with materials tailored closely to the target popu-lation segments. Retention will be measured by signing upfor the 6-month intervention and measuring quality andquantity of interactions with the program over that time.

Data analysis: The research team will passively trackURLs and SM channel usage for participants who receivethe initial SM message. Their response to communications,for example, site visits, actions, will be tracked even if theydo not proceed. SM usage (e.g. Facebook) will be correlatedwith channel use (i.e. did this channel only work for highlevel Facebook users?). The retention measures will be cor-related with target segment and demographic characteris-tics, SES (income, education) and minority populationgroups. The size of friend networks, frequency of activity,and other measures of SM use, along with participant char-acteristics will be evaluated as potential moderators of inter-vention effects, identifying those groups who benefit mostor the least from these types of interventions. Finally, wewill use this information to test an engagement and reten-tion measurement scale. The aim is for this scale to beapplied to future social marketing campaigns to measureand benchmark success.

Phase 3 outcomes: The outcomes of interest are the levelof engagement, interaction, retention and impact of the SMcampaign on health-related behaviours in YAs.

Phase 4—Translation: Phase 4 is designed to ensure thatthe learnings from Phases 1 to 3 are available to the socialmarketing and health communities in a form that is readilyaccessible and easy to use. Research reports based on theoutcomes will be provided to participating partners, inaddition to academic peer-reviewed publications reportingon the outcomes. Furthermore, the TT will be reconvened

in order to evaluate the outcomes of previous phases interms of its applicability to the landscape of obesityresearch. The TT will establish the parameters for the devel-opment of the toolkit, which is envisaged as a guide forsocial-media-social-marketing aimed at young people andhealthy eating. This toolkit will potentially be downloadablefrom the CH website. In addition to tools to measure effec-tiveness and engagement, the toolkit will suggest strategiesto optimise campaigns and outline metrics for effective SMand healthy eating campaigns for young Australians, espe-cially Aboriginal YAs. This will permit an immediate scal-ability of the research and possible extension to otherhealth domains.

Discussion

An evidence-based approach drawing expertise and knowl-edge of research design from several disciplines was used todesign this series of studies which comprise CH. The itera-tive approach will first understand the target population,lead to segmentation and then co-create intervention con-tent with the target group and expert stakeholders, andfinally will test engagement with the content in a real lifecampaign. The overall approach recognises the importanceof participatory research and embraces the unknown in acontinuous cycle of designing, testing and improving,rather than using the traditional approach of creating anintervention, deploying it and only discovering issues andproblems during the final evaluation, rather than through-out the design phase.

CH uses and adapts methodologies that have previouslybeen developed in marketing research. A major benefit ofthe OC approach is that it is ethnographic in nature, and assuch allows researchers to understand the cultural nature ofcommunity dialogue; specific inclusion of AboriginalAustralians is crucial and integral to the methodologicaldesign. These approaches enable us to understand the why,how, when and what, for each behaviour under scrutiny. Inturn, this allows rich insights into participants’ values,assumptions, beliefs, rituals, language and identity, in rela-tion to the role of healthy eating in their lives. In a virtuallounge room, the moderator has to stimulate conversationsand demonstrate authenticity to engage members effec-tively. In doing so, the researcher becomes part of the tribein order to understand it. Furthermore, relative to dataextracted from some of the more conventional researchmethods, online communities are a source of naturallyoccurring data and are more suited to uncover culturalinfluences and insights that are not biased either in theirgeneration or in reporting. A powerful application of thisapproach is the ability of participants to co-create contentsuch a health messages.

In addition to input from the target group, this projectinvolves multiple stakeholders from government and non-government organisations at all stages. This partnershipbetween researchers and those who will eventually imple-ment social marketing campaigns has been recommended

C. Lombard et al.

516 © 2018 Dietitians Association of Australia

as an aid to successful research and program delivery andin translation and adoption of findings into practice.81,82

The strength of Phase 3 is the ability to test content inthe real world, as opposed to a ‘controlled’ research setting,a strategy which means that the results will be readily appli-cable and translatable into practice.78 Despite the real-worldsetting, CH will employ a robust research methodology,including utilisation of a suite of metrics and indicators, toevaluate the intervention.

The iterative nature of the project, incorporating a rangeof methodologies is the key strength. This allows the gath-ering of a range of data and provides deeper insights to bet-ter understand the problem and how to address it. Theparticipatory nature of the content development willincrease uptake and engagement in the target group. Stud-ies addressing the use of SM to reach this group with healthmessages are limited,83 and few have incorporated an in-depth ethnographic multi-level mixed-methods approach inintervention development.

CH is a novel protocol that aims to synthesise the effec-tiveness of SM strategies to motivate, engage and retain YAs,especially Aboriginal YAs, in interventions to reduce the riskof obesity, and to identify and disseminate effective ways todeliver these interventions via SM. This will be achievedusing mixed methods, an iterative co-creation of contentprocess, and real-world implementation and evaluation,drawing on behavioural change theories and social market-ing strategies from an experienced interdisciplinary team.

Funding source

The present study was supported by the National Healthand Medical Research Council (grant number:GNT1115496). Partner organisations involved include VicHealth, the Victorian Department of Health and HumanServices, Cancer Council Victoria, Nutrition Australia, HeartFoundation, Victorian Aboriginal Community ControlledHealth Organisation, Janette Kendall Consulting and FitnessAustralia.

Conflict of interest

The authors declare that they have no competing interests.

Authorship

CL, CP, MSCL, TAM and HT contributed to study concep-tion and design and manuscript revision; LB contributed todrafting the manuscript; study conception and design, spe-cifically in developing the methodology and measures; pro-viding critical revision for intellectual content; MRcontributed to study conception and design, specifically indeveloping the methodology and measures; KMK contrib-uted to drafting the manuscript and implementing themethodology and measures; TW contributed to studymethodologies in particular the engagement strategies forAboriginal and Torres Strait Islander people; MD contrib-uted to study design and manuscript revision. All authors

provided final approval of the version to be published andhave agreed to be accountable for all aspects of the work inensuring that questions related to the accuracy or integrityof any part of the work are appropriately investigated andresolved.

References

1 World Health Organization. Obesity and overweight: Factsheet.World Health Organization, 2016. (Available from: http://www.who.int/mediacentre/factsheets/fs311/en/:, accessed 8 September2017).

2 National Health and Medical Research Council. Targeted call forresearch into engaging and retaining young adults in interven-tions to improve eating behaviours and health outcomes(Preventing Obesity in 18–24 year olds TCR), 2015. (Availablefrom: https://www.nhmrc.gov.au/grants-funding/apply-funding/targeted-and-urgent-calls-research/preventing-obesity-18-24-year-olds-t, accessed 8 September 2017).

3 Australian Bureau of Statistics. 4364.0.55.001—AustralianHealth Survey: First Results, 2011–12. Canberra, Australia:Australian Bureau of Statistics, 2012.

4 Australian Bureau of Statistics. 4727.0.55.001—AustralianAboriginal and Torres Strait Islander Health Survey: First Results,Australia, 2012–13. Canberra, Australia: Australian Bureau ofStatistics, 2013.

5 Lim SS, Vos T, Flaxman AD et al. A comparative risk assess-ment of burden of disease and injury attributable to 67 riskfactors and risk factor clusters in 21 regions, 1990–2010: a sys-tematic analysis for the Global Burden of Disease Study 2010.Lancet 2012; 380: 2224–60.

6 Lobstein T, Baur L, Uauy R, TaskForce IIO. Obesity in childrenand young people: a crisis in public health. Obes Rev 2004; 5(S1): 4–104.

7 Ashton LM, Hutchesson MJ, Rollo ME, Morgan PJ, Collins CE.Motivators and barriers to engaging in healthy eating and phys-ical activity: a cross-sectional survey in young adult men.Am J Mens Health 2017; 11: 330–43.

8 Ashton LM, Hutchesson MJ, Rollo ME, Morgan PJ,Thompson DI, Collins CE. Young adult males’ motivators andperceived barriers towards eating healthily and being active: aqualitative study. Int J Behav Nutr Phys Act 2015; 12: 93.

9 Ashton LM, Morgan PJ, Hutchesson MJ, Rollo ME, Young MD,Collins CE. A systematic review of SNAPO (smoking, nutrition,alcohol, physical activity and obesity) randomized controlledtrials in young adult men. Prev Med 2015; 81: 221–31.

10 Chapman K, Havill M, Watson WL et al. Time to address con-tinued poor vegetable intake in Australia for prevention ofchronic disease. Appetite 2016; 107: 295–302.

11 Menozzi D, Sogari G, Mora C. Explaining vegetable consump-tion among young adults: an application of the theory ofplanned behaviour. Nutrients 2015; 7: 7633–50.

12 Bleich SN, Wang YC, Wang Y, Gortmaker SL. Increasing con-sumption of sugar-sweetened beverages among US adults:1988–1994 to 1999–2004. Am J Clin Nutr 2009; 89: 372–81.

13 Sui ZX, Wong WK, Louie JCY, Rangan A. Discretionary foodand beverage consumption and its association with demo-graphic characteristics, weight status, and fruit and vegetableintakes in Australian adults. Public Health Nutr 2017; 20:274–81.

14 Warin M, Zivkovic T, Moore V, Ward PR, Jones M. Short hori-zons and obesity futures: Disjunctures between public health

Communicating health study protocol

© 2018 Dietitians Association of Australia 517

interventions and everyday temporalities. Soc Sci Med 2015;128: 309–15.

15 Wills J, Crichton N, Lorenc A, Kelly M. Using population seg-mentation to inform local obesity strategy in England. HealthPromot Int 2014; 30: 658–66.

16 Munt AE, Partridge SR, Allman-Farinelli M. The barriers andenablers of healthy eating among young adults: a missing piece ofthe obesity puzzle: a scoping review. Obes Rev 2017; 18: 1–17.

17 Hollywood LE, Cuskelly GJ, O’Brien M et al. Healthful groceryshopping. Perceptions and barriers. Appetite 2013; 70: 119–26.

18 Palermo C, Perera-Schulz D, Kannan A et al. Development andevaluation of an iPad app for measuring the cost of a nutritiousdiet. JMIR mHealth uHealth 2014; 2: e50.

19 Paxton SJ, McLean SA. Peer interactions and relationships—risk factors for eating disorders. In: Wade TD, ed. Encyclopediaof Feeding and Eating Disorders. Singapore: Springer, 2015.

20 Brimblecombe J, Maypilama E, Colles S et al. Factors influenc-ing food choice in an Australian Aboriginal community. QualHealth Res 2014; 24: 387–400.

21 Colles SL, Maypilama E, Brimblecombe J. Food, food choiceand nutrition promotion in a remote Australian Aboriginalcommunity. Aust J Prim Health 2014; 20: 365–72.

22 Kerpan ST, Humbert ML, Henry CJ. Determinants of diet forurban aboriginal youth: implications for health promotion.Health Promot Pract 2015; 16: 392–400.

23 Allman-Farinelli MA, Chey T, Bauman AE, Gill T, James WPT.Age, period and birth cohort effects on prevalence of over-weight and obesity in Australian adults from 1990 to 2000.Eur J Clin Nutr 2008; 62: 898–907.

24 Freeman B, Kelly B, Vandevijvere S, Baur L. Young adults:beloved by food and drink marketers and forgotten by publichealth? Health Promot Int 2016; 31: 954–61.

25 Nigg CR, Geller KS, Motl RW, Horwath CC, Wertin KK,Dishman RK. A research agenda to examine the efficacy andrelevance of the transtheoretical model for physical activitybehavior. Psychol Sport Exerc 2011; 12: 7–12.

26 Brennan L, Binney W, Parker L, Aleti T, Nguyen D. Social Mar-keting and Behaviour Change: Models, Theory and Applications.Cheltenham, UK: Edward Elgar, 2014.

27 Ashley C, Tuten T. Creative strategies in social media market-ing: an exploratory study of branded social content and con-sumer engagement. Psychol Market 2015; 32: 15–27.

28 Voorveld HA, van Noort G. Social media in advertising cam-paigns: examining the effects on perceived persuasive intent,campaign and brand responses. J Creative Commun 2014; 9:253–68.

29 Willis EA, Szabo-Reed AN, Ptomey LT et al. Do weight man-agement interventions delivered by online social networkseffectively improve body weight, body composition, andchronic disease risk factors? A systematic review. J Telemed Tel-ecare 2017; 23: 263–72.

30 Wiederhold BK. Beyond direct benefits: indirect health benefitsof social media use. Cyberpsychol Behav Soc Netw 2017;20: 1–2.

31 Pappa GL, Cunha TO, Bicalho PV et al. Factors associated withweight change in online weight management communities: acase study in the LoseIt Reddit community. J Med Internet Res2017; 19: e17.

32 Rouse M. Definition: social media, 2016. (Available from:https://protect-au.mimecast.com/s/M41pB3Sp0letnA?domain=whatis.techtarget.com, accessed 12 September 2017).

33 Hootsuite Media Inc. We are social Ltd. digital in 2017 globaloverview, 2017. (Available from: https://protect-au.mimecast.

com/s/wxnVBQFMnzmsqY?domain=slideshare.net, accessed 12September 2017).

34 Australian Association of Social Marketing, International SocialMarketing Association, European Social Marketing Association.Consensus definition of social marketing. http://www.i-socialmarketing.org/assets/social_marketing_definition.pdf. 2013.Accessed September 12 2017.

35 Rittel WJ, Webber MM. Dilemmas in a general theory of plan-ning. Policy Sciences. 1973; 4(2): 155–169.

36 Australian Bureau of Statistics. 8146.0—Household Use of Infor-mation Technology: Australia, 2014–15. Canberra, Australia:Australian Bureau of Statistics, 2016.

37 Rice ES, Haynes E, Royce P, Thompson SC. Social media anddigital technology use among Indigenous young people inAustralia: a literature review. Int J Equity Health 2016; 15: 81.

38 Hebden L, Chey T, Allman-Farinelli M. Lifestyle interventionfor preventing weight gain in young adults: a systematicreview and meta-analysis of RCTs. Obes Rev 2012; 13:692–710.

39 Truong VD, Dang NV. Reviewing research evidence for socialmarketing: systematic literature reviews. In: Kubacki K,Rundle-Thiele S, eds. Formative Research in Social Marketing.Singapore: Springer, 2017; 183–250.

40 Vaterlaus JM, Patten EV, Roche C, Young JA. #Gettinghealthy:the perceived influence of social media on young adult healthbehaviors. Comp Human Behav 2015; 45: 151–7.

41 Vandelanotte C, Kirwan M, Rebar A et al. Examining the use ofevidence-based and social media supported tools in freelyaccessible physical activity intervention websites. Int J BehavNutr Phys Act 2014; 11: 105.

42 Greenhalgh T, Jackson C, Shaw S, Janamian T. Achievingresearch impact through co-creation in community-basedhealth services. Literature review and case study. Milbank Q2016; 94: 392–429.

43 Lim MS, Hocking JS, Aitken CK et al. Impact of text andemail messaging on the sexual health of young people: a rando-mised controlled trial. J Epidemiol Community Health 2012;66: 69–74.

44 Dooley JA, Jones SC, Iverson D. Social marketing and healthpromotion communication: from traditional marketing to Web2.0. Cases Public Health Commun Market 2013; 7: 120–39.

45 Slater MD. Theory and method in health audience segmenta-tion. J Health Commun 1996; 1: 267–83.

46 Moss HB, Kirby SD, Donodeo F. Characterizing and reachinghigh-risk drinkers using audience segmentation. Alcohol ClinExp Res 2009; 33: 1336–45.

47 VicHealth. Drinking-related Lifestyles: Exploring the Role of Alco-hol in Victorians’ Lives. Melbourne, Australia: Victorian HealthPromotion Foundation, 2013.

48 Maibach EW, Maxfield A, Ladin K, Slater M. Translating healthpsychology into effective health communication: the Americanhealthstyles audience segmentation project. J Health Psychol1996; 1: 261–77.

49 Brennan L, Previte J, Fry ML. Social marketing’s consumermyopia applying a behavioural ecological model to addresswicked problems. J Soc Market 2016; 6: 219–39.

50 Kreuter MW, Oswald DL, Bull FC, Clark EM. Are tailoredhealth education materials always more effective than non-tailored materials? Health Educ Res 2000; 15: 305–15.

51 Kubacki K, Dietrich T, Rundle-Thiele S, eds.. Segmentation insocial marketing: why we should do it more often that we cur-rently do. In: Segmentation in Social Marketing: Process, Methodsand Application. Singapore: Springer, 2017; 1–6.

C. Lombard et al.

518 © 2018 Dietitians Association of Australia

52 Buyucek N, Kubacki K, Rundle-Thiele S, Pang B. A systematicreview of stakeholder involvement in social marketing inter-ventions. Austral Market J 2016; 24: 8–19.

53 Choi BC, Pak AW. Multidisciplinarity, interdisciplinarity andtransdisciplinarity in health research, services, education andpolicy: 1. Definitions, objectives, and evidence of effectiveness.Clin Invest Med 2006; 29: 351–64.

54 Creswell JW. Research Design: Qualitative, Quantitative andMixed Methods Approaches, 2nd edn. Thousand Oaks, CA: SagePublications, 2003.

55 Fishbein M, Yzer MC. Using theory to design effective healthbehavior interventions. Commun Theor 2003; 13: 164–83.

56 Fishbein M. The role of theory in HIV prevention. AIDS Care2000; 12: 273–8.

57 Fishbein M, Cappella JN. The role of theory in developingeffective health communications. J Commun 2006; 56: S1–S17.

58 Fishbein M, Ajzen I. Belief, Attitude, Intention, and Behavior: AnIntroduction to Theory and Research. Reading, MA: Addison-Wesley, 1975.

59 Tversky A, Kahneman D. Judgment under uncertainty: heuris-tics and biases. In: Utility, Probability, and Human Decision Mak-ing. Singapore: Springer, 1975; 141–62.

60 Vandenbroeck P, Goossens J, Clemens M. Tackling obesities:future choices—building the obesity system map. UK Govern-ment Office for Science (Ed.). In: Foresight Programme.London: UK Government, 2007.

61 Lombard C, Deeks A, Jolley D, Ball K, Teede H. A low inten-sity, community based lifestyle programme to prevent weightgain in women with young children: cluster randomised con-trolled trial. BMJ 2010; 341: c3215.

62 Pink S. Digital Ethnography. Singapore: Springer, 2016.63 Dessart L, Veloutsou C, Morgan-Thomas A. Consumer engage-

ment in online brand communities: a social media perspective.J Product Brand Manag 2015; 24: 28–42.

64 Morgan DL, Slade MD. Aboriginal philosophy and its impacton health care outcomes. Aust N Z J Public Health 1997; 21:597–601.

65 Barnett L, Kendall E. Culturally appropriate methods forenhancing the participation of Aboriginal Australians inhealth-promoting programs. Health Promot J Aust 2011;22: 27–32.

66 Sharifian F. Cultural conceptualisations in intercultural com-munication: a study of Aboriginal and non-AboriginalAustralians. J Pragmat 2010; 42: 3367–76.

67 Boeije H. A purposeful approach to the constant comparativemethod in the analysis of qualitative interviews. Qual Quant2002; 36: 391–409.

68 Borsboom D. Measuring the Mind: Conceptual Issues in Contem-porary Psychometrics. Cambridge, UK: University Press, 2005.

69 DeVellis RF. Scale Development: Theory and Applications, 4thedn, Vol. 26. Los Angeles, CA: Sage Publications, 2016.

70 Farrell AM. Insufficient discriminant validity: a comment onBove, Pervan, Beatty, and Shiu (2009). J Bus Res 2010;63: 324–7.

71 Hair J, Black W, Babin B, Anderson R, Tatham R. MultivariateData Analysis. New Jersey, NJ: Pearson Education Limited, 2013.

72 Sarkar MB, Echambadi R, Cavusgil ST, Aulakh P. The influenceof complementarity, compatibility, and relationship capital onalliance performance. J Acad Market Sci 2001; 29: 358–73.

73 Sweeney JC, Soutar GN. Consumer perceived value: the devel-opment of a multiple item scale. J Retailing 2001; 77: 203–20.

74 Grunert KG. Towards a concept of food-related life style. Appe-tite 1993; 21: 151–5.

75 Hair JF Jr, Hult GTM, Ringle C, Sarstedt M. A Primer on PartialLeast Squares Structural Equation Modeling (PLS-SEM). California:Sage Publications, 2016.

76 Mummah SA, Robinson TN, King AC, Gardner CD, Sutton S.IDEAS (integrate, design, assess, and share): a framework andtoolkit of strategies for the development of more effective digi-tal interventions to change health behavior. J Med Internet Res2016; 18: e317.

77 Verloigne M, Altenburg TM, Chinapaw MJM, Chastin S,Cardon G, De Bourdeaudhuij I. Using a co-creational approachto develop, implement and evaluate an intervention to promotephysical activity in adolescent girls from vocational and techni-cal schools: a case control study. Int J Environ Res Public Health2017; 14: 2–17.

78 Lim MS, Wright CJ, Carrotte ER, Pedrana AE. Reach, engage-ment, and effectiveness: a systematic review of evaluationmethodologies used in health promotion via social networkingsites. Health Promot J Aust 2016; 27: 187–97.

79 Laverty SM. Hermeneutic phenomenology and phenomenol-ogy: a comparison of historical and methodological consider-ations. Int J Qual Methods 2003; 2: 21–35.

80 Durso LE, Latner JD. Understanding self-directed stigma:development of the weight bias internalization scale. Obesity2008; 16 (S2): S80–6.

81 Wolfenden L, Williams CM, Wiggers J, Nathan N, Yoong SL.Improving the translation of health promotion interventionsusing effectiveness—implementation hybrid designs in pro-gram evaluations. Health Promot J Aust 2017; 27: 204–7.

82 Milat AJ, Laws R, King L et al. Policy and practice impacts ofapplied research: a case study analysis of the New South Waleshealth promotion demonstration research grants scheme2000–2006. Health Res Policy Syst 2013; 11: 5.

83 Yonker LM, Zan S, Scirica CV, Jethwani K, Kinane TB. “Friend-ing” teens: systematic review of social media in adolescent andyoung adult health care. J Med Internet Res 2015; 17: e4.

Communicating health study protocol

© 2018 Dietitians Association of Australia 519

ORIGINAL RESEARCH

Use of hand grip strength in nutrition risk screeningof older patients admitted to general surgical wards

Angela BYRNES ,1,2 Alison MUDGE,3,4,5 Adrienne YOUNG ,2,6 Merrilyn BANKS2,6 andJudy BAUER 1

1Centre for Dietetics Research, School of Human Movement and Nutrition Sciences, The University of Queensland,2Nutrition and Dietetics Department and 3Internal Medicine and Aged Care Department, Royal Brisbane andWomen’s Hospital, 4Institute for Health and Biomedical Innovation, Queensland University of Technology, 5School ofMedicine, The University of Queensland and 6School of Exercise and Nutrition Sciences, Queensland University ofTechnology, Brisbane, Queensland, Australia

AbstractAim: Hand grip strength (HGS) has been proposed as an indicator of nutritional status that is objective, requires min-imal assessor training and is quick to administer, making it attractive for use in the acute setting. This study aimedto determine the discriminatory ability of impaired HGS to screen for malnutrition in an older hospital population andassess the added value of combining this with existing screening tools.Methods: Measures were undertaken during acute admission in patients ≥65 years admitted to general surgicalwards. Impaired HGS was defined as a mean value below the lower limit of the 95% CI of population norms andobserved HGS standardised as a percentage of this value. Nutritional risk was assessed using the MalnutritionScreening Tool (MST) and malnutrition defined as Patient-Generated Subjective Global Assessment (PG-SGA) rating Bor C. Discriminatory ability of impaired HGS to identify malnourished patients was tested using the area under thereceiver operating characteristic curve (AUC).Results: Seventy-five patients (mean age: 74.0 (SD 6.7) years, 60% male) were recruited. Impaired HGS did not accu-rately identify malnutrition (AUC (95% CI): 0.41 (0.25–0.58), P < 0.001), nor did it improve discriminatory ability ofthe MST (AUC (95% CI), MST: 0.83 (0.71–0.95), P = 0.32; MST/HGS combined: 0.68 (0.51–0.86), P = 0.035).Conclusions: HGS was not found to be suitable in screening older inpatients for malnutrition during admission tosurgical wards. As such, screening for nutrition risk using an existing validated tool to identify patients for furtherin-depth nutritional assessment by an appropriately trained clinician remains the preferred method.

Key words: aged, hand strength, inpatients, malnutrition, sensitivity, specificity.

Introduction

Protein–energy malnutrition—well-recognised in the acutesetting as an issue with clinical, financial and patient-levelconsequences1—has an estimated point prevalence acrossAustralian and New Zealand acute care facilities of 32%.2

Malnutrition and unintentional weight loss in the periodpreceding surgery have been linked to adverse clinicaloutcomes.3–5 This risk is pronounced in the elderly, who aremore likely to be malnourished prior to hospital admission.6

A number of nutrition screening tools have been validatedfor use in the clinical setting, many of which generally per-form well in discriminating between malnourished and well-nourished patients.7 Likewise, considerable variation existsin the methods used to assess nutritional status across stud-ies of hospitalised patients3,8,9 and includes anthropometricmeasures, biochemical markers, combinations of the two, orone of a number of validated nutrition assessment tools thatform an overall assessment based on the individual’s medicalhistory (e.g. weight change, functional capacity, nutritionimpact symptoms), food intake history and nutrition-focused physical examination.10 An example of one of thesetools is the Patient-Generated Subjective Global Assessment(PG-SGA).11

For a diagnostic or screening test to display clinical util-ity, there are four conditions that must be satisfied: accessi-bility, practicality, acceptability and appropriateness.12 Or,in other words, as a bedside measure it should be inexpen-sive, simple, non-invasive and valid. The limitation of toolssuch as the PG-SGA is they often require clinician training

A. Byrnes, APD, BHlthSc(Nutr&Diet), PhD CandidateA. Mudge, FRACP, PhD, Adjunct ProfessorA. Young, APD, PhD, Senior DietitianM. Banks, AdvAPD, PhD, Director Nutrition & DieteticsJ. Bauer, FDAA, PhD, Associate ProfessorCorrespondence: A. Byrnes, Centre for Dietetics Research, School ofHuman Movement and Nutrition Sciences, The University ofQueensland, Brisbane, QLD 4072, Australia. Tel: +61 413 461 412.Email: a.byrnes@uq.edu.au

Accepted February 2018

© 2018 Dietitians Association of Australia520

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 520–526 DOI: 10.1111/1747-0080.12422

(especially in undertaking nutrition-focused physical assess-ment), collection of detailed intake history is time consum-ing and they are subjective.13 It is therefore preferrable inthe busy clinical setting to employ quick and easy screeningthat requires minimal training to identify patients for fur-ther nutritional assessment by an appropriately trained cli-nician. One such tool commonly used in Australian acutecare facilities is the Malnutrition Screening Tool (MST).14

Hand grip strength (HGS) has been proposed as an indi-cator of protein–energy malnutrition or undernutrition thatmay be more sensitive to changes in muscle function sec-ondary to declining nutritional status than muscle mass orother body composition measures.15,16 Further advantagesinclude minimal assessor training, quick administration andobjectivity. When present in combination with other cri-teria, reduced HGS is recommended by the Academy ofNutrition and Dietetics and the American Society of Paren-teral and Enteral Nutrition (ASPEN) in identifying adultmalnutrition.10 A 2015 study by Demeny et al.,17 however,found that in the Australian acute care setting, few dietitiansare using HGS in the assessment of nutritional status or inmonitoring nutrition support strategies in older patientsassessed as malnourished or at risk of malnutrition.

HGS is not without limitations, with an array of factorsrelated to acute or chronic illness and associated treatmentthat may contribute to muscle weakness, such as diseaseseverity, comorbidities, inflammation, oxidative stress, cer-tain medications and electrolyte imbalances.16,18 Further-more, studies assessing its usefulness as an indicator ofnutritional risk19 and nutritional status in adult medicaland surgical inpatients20 and outpatients21–23 areconflicting.

The aim of this study was to determine the discrimina-tory ability of HGS in screening for malnutrition as assessedby the PG-SGA in an older (≥65 years) population admittedto general surgical wards and assess the added value ofcombining this with the MST. A secondary aim of thisstudy was to determine if HGS is correlated with PG-SGAtriage score. Throughout this report, ‘malnutrition’ will beused to refer to a state of undernutrition due to proteinand/or energy intake that is insufficient to meet require-ments, which may be elevated secondary to reducedabsorption, altered gastrointestinal transit, or impairednutrient utilisation.10 ‘Well-nourished’ has been used toindicate an absence of malnutrition.

Methods

The present study was a sub-study of the Collaborative forHospitalised Elders, Reducing the Impact of Stays in Hospi-tal (CHERISH) study.24 Consecutive eligible patients≥65 years were prospectively recruited (September 2015 toMarch 2016) from two general surgical wards at a large ter-tiary teaching hospital in Brisbane, Australia. Eligibility cri-teria included a length of stay ≥72 hours, not critically ill orpalliated, and informed consent could be obtained. Patientsunable to complete the test due to inability to comprehendor follow instructions (e.g. severe cognitive impairment,

limited English), or condition resulting in reduced handfunction were excluded (e.g. rheumatoid arthritis or carpaltunnel syndrome).

Participant descriptive data were retrieved from routinelycollected clinical and administrative information in the patientchart, with the exception of weight that was collected using asingle set of digital scales (Tanita BC 582 FitPlus InnerscanBody Composition Monitor; Tanita, Kewdale, Australia).Nutrition risk screening, nutrition assessment and HGS mea-sures were undertaken in that order during the same sessionby a single researcher, between day 4 and 6 of admission.

Nutrition risk was assessed from information obtaineddirectly from the patient using the MST,14 with at riskpatients scoring ≥2. Nutritional status was assessed usingthe scored PG-SGA,11 with information collected directlyfrom the patient and verified from the medical chart whereappropriate (e.g. weight history, presence of fever, cortico-steroid prescription). This resulted in a global rating ofnutritional status (i.e. PG-SGA A: well-nourished, B: sus-pected or moderate malnutrition or C: severely malnour-ished) and score. Malnutrition was defined as a combinedglobal rating of PG-SGA B or C. The PG-SGA is validatedfor use in acute patients25–27 and was specifically selecteddue to the addition of the triage score indicating nutritionimpact over the previous 2 weeks.

HGS was measured using a single Jamar hydraulic handdynamometer (Lafayette Instrument, Lafayette, IN) as perstandardised positioning and instruction prescribed by theAmerican Society of Hand Therapists (ASHT),28 and recordedas the mean of three trials (2–4-second isometric contractionwith minimum of 30-second break in between) in the domi-nant arm. In the absence of standardised recommendationsrelating to the use of dominant versus non-dominant side, thedominant arm was selected. In a systematic review of 72 stud-ies examining protocols for measuring HGS, the majority ofincluded studies used dominant hand; however, the authorsacknowledged evidence relating to dominance-specific differ-ences in grip strength is conflicting.29

All trials were undertaken with the dynamometer handlein setting II, as recommended by the ASHT.28 Existingresearch suggests handle position does not considerablyreduce accuracy of grip strength readings when adminis-tered at handle setting II.30,31 The procedure was describedand demonstrated to patients and correct position con-firmed prior to the beginning of the test, with verbalencouragement and feedback provided throughout. Timetaken to administer HGS was recorded using a digital timer.

Crude HGS scores were coded as impaired or notimpaired, with impaired HGS defined as a value below thelower limit of the 95% CI of the mean from age-, gender- andside-specific normative data,32,33 as suggested by Bohannonet al.33

MST and HGS were combined to form a compositescreening measure (MST/HGS combined), whereby thosepatients who had both an MST score ≥2 and impaired HGSwere considered to be at increased risk.

Dichotomised nutrition risk, HGS and MST/HGS com-bined were compared with global malnutrition rating to

HGS in nutrition risk screening

© 2018 Dietitians Association of Australia 521

determine discriminatory ability: sensitivity, specificity, posi-tive predictive value, negative predictive value and area underthe receiver operating characteristic curve (AUC). A highlysensitive and specific test is one with good ability to detectpatients with the disease (true positives), while also correctlyruling out patients without the disease (true negatives). Sensi-tivity and specificity values are impacted by disease preva-lence, therefore calculating the AUC (of a plot of sensitivityand one minus specificity for all thresholds) is a valuablemethod of determining discriminatory ability. AUC valuescloser to 1 indicate better discriminatory ability, that is, higherprobability that a randomly selected patient with the condi-tion of interest would have test results indicative of this condi-tion, than a randomly selected patient without the conditionof interest.34 Interpretation of AUC values were based uponestablished cut-off points: 0.9–1.0 = excellent; 0.8–0.9 = verygood; 0.7–0.8 = good; 0.6–0.7 = sufficient; 0.5–0.6 = poor;<0.5 = test not useful.35

Spearman’s rho (rs) was used to test for covariancebetween standardised HGS (i.e. mean HGS expressed aspercent of the lower limit of the 95% CI of the mean fromnormative data) and PG-SGA score.

Statistical significance was established at P < 0.05. Allstatistical analyses were carried out using the Software Pack-age for Social Sciences (SPSS) for Mac (version 23.0, IBMCorp., Chicago, IL). Post-hoc power analysis revealed thestudy to be adequately powered to detect an AUC of 0.705(Power: 80%, Type I error: 5% and allocation ratio: 4:1).

The study was approved by the HREC of the local healthdistrict (Ref No.: HREC/15/QRBW/95) and reported inaccordance with The Strengthening the Reporting of Obser-vational Studies in Epidemiology (STROBE) statement.36

Results

Of 273 eligible patients, 173 were enrolled into the CHER-ISH study (63.4% consent rate) and 75 patients (mean age:74.0 (SD 6.7) years, 60% male) were subsequently includedin this sub-study. The most common reason for exclusionfollowing enrolment was patient discharged before mea-sures were undertaken (n = 59). Thirty-two patients wereunable to perform HGS measures due to a pre-existinghand condition (n = 26) or inability to comprehend or fol-low instructions (n = 6), four declined to undertake HGSmeasures and a further three patients withdrew from thestudy or were palliated (Figure 1). Participant demographicand clinical characteristics are outlined in Table 1.

Time taken to measure and record HGS was mean 2.58(SD 0.52) minutes. Thirty-two percent (n = 24) of patientswere screened at nutritional risk, malnutrition was observedin 19% (n = 14) of patients (only one patient was assessedas being severely malnourished, i.e. PG-SGA C) and 36%(n = 27) had impaired HGS. Thirteen percent (n = 10) ofpatients were identified as at nutritional risk and hadimpaired HGS (MST/HGS combined). No difference wasobserved in mean crude HGS between malnourished andwell-nourished patients (26.3 (SD 9.4) kgf vs 27.8 (SD 8.4)kgf; P = 0.55).

Discriminatory ability of the MST, HGS and MST/HGScombined in screening patients for malnutrition are dis-played in Table 2. The MST was found to have good dis-criminatory ability in identifying patients with malnutrition.HGS was not found to be useful in identifying patients withmalnutrition, nor did combining HGS with the MSTincrease the discriminatory ability of the MST. The corre-sponding receiver operating characteristic curves are dis-played in Figure 2.

No covariance was observed between standardised HGSand PG-SGA score (rs: −0.129, P = 0.27).

Discussion

This study aimed to determine the discriminatory ability ofHGS in screening for malnutrition during admission to gen-eral surgical wards in an older (≥65 years) population andfound HGS does not appear to be useful as a screeningmeasure for malnutrition, or in identifying those most likelyto benefit from dietetic intervention when used as a standa-lone measure or when combined with the MST. AlthoughHGS was inexpensive, brief and simple to perform, thisstudy reveals issues with its practicality and validity withrespect to using it to screen for malnutrition.

The inappropriateness of HGS as a measure in patientswith upper limb injury, malformation or musculoskeletalcondition presents a challenge in older patients where thereis a high incidence of musculoskeletal problems such asosteoarthritis. In our sample, approximately 15% of patientswere excluded from HGS measurement for this reason.

The results from the present study are in disagreementwith the literature reporting discriminatory ability of HGSto detect malnutrition in adult inpatient populations. Floodet al.20 found HGS had good discriminatory ability indetecting malnutrition as determined by the PG-SGA in aheterogeneous sample of adult (≥18 years) hospital patients(n = 217) referred to dietetic services (AUC: 0.776 (95%CI: 0.698–0.853)). However, logistic regression modellingof proportion of variance in nutritional status shared bypercent of predicted HGS calculated using age-, gender-and BMI-adjusted predictive equations, found HGS sharedonly 9.3% of variability with nutritional status in the samegroup. Jeejeebhoy et al.37 in their study of 733 hospitalisedadults (≥18 years) across a variety of diagnostic categoriessimilarly found HGS was able to detect malnutrition asdetermined by SGA, however, this value indicated a testthat was only sufficient (AUC: 0.61 (95% CI: 0.57–0.65)),even when adjusted for age and sex (AUC: 0.63 (95% CI:0.59–0.67)). These studies differ from the present study inthe values selected for entry into the model, in thatobserved HGS was entered as a continuous variable. Cut-offs or HGS values for use in screening for malnutrition,however, were not suggested by the study authors.

A range of factors, other than nutritional status, mayhave impacted HGS in our sample. For example, a largeproportion of the sample had a primary diagnosis of cancer(45%) or other condition with related inflammation(e.g. inflammatory bowel disease, cholecystitis, trauma).

A. Byrnes et al.

522 © 2018 Dietitians Association of Australia

Inflammation may contribute to muscle weakness,38 poten-tially confounding the relationship between decreased HGSand nutritional status. Similarly, multimorbidity was preva-lent (evidenced by mean Charlson Comorbidity Index of2.6 (SD 2.1)), which has been shown to decrease HGS.18

Therefore, it is not unexpected that this group would pre-sent with lower HGS than their healthy counterparts. Addi-tionally, hand anthropometry, stature and body weight arereported to be predictors of HGS.18,39,40 These factors werenot accounted for in our study, as the study intended toassess the pragmatic use of HGS in the clinical setting.Without controlling for disease-related factors and anthro-pometric measures, HGS may have little use in screeningfor malnutrition.

It has also previously been suggested that HGS maymore closely reflect sarcopenia and frailty in older adults16

and, in combination with low muscle mass, is recom-mended by the European Working Group on Sarcopenia inOlder People in identifying sarcopenia.41 While undernutri-tion is recognised as a potential contributing factor to theoverarching geriatric syndrome of frailty,41,42 sarcopeniaand frailty can present in otherwise well-nourished individ-uals.43 For this reason, HGS is unlikely to be suitable in dif-ferentiating between nutrition- and age-related sarcopenia,without first establishing whether the patient is

malnourished. This further complicates the associationbetween reduced HGS and malnutrition in an older popula-tion. Whether indicative of nutritional status or sarcopenia,low HGS has been shown to be a predictor of increasedlength of stay, postoperative complications and requirementfor assistance at admission and at 6 months post-discharge.37,44–46 It may therefore be useful in screeningpatients who would most benefit from multidisciplinaryintervention and early discharge planning.

HGS may be useful as a sensitive measure of short-termchange in nutritional status, such as monitoring the effectof nutritional intervention.16,20 This is supported by theearly response of muscle strength to refeeding observed inearly starvation studies,47 however, whether this extends touse in an elderly population and the timeframe required toobserve change is unclear.48 If shown to be a valid measureof change, this would make it attractive for use in the acutecare setting where duration of observation is often shortand measurable improvements in patient weight or nutri-tional status as measured by a multicomponent tool may bedifficult.

The present study has some limitations. The large num-ber of patients enrolled in the parent study that were subse-quently excluded from the sub-study due to early dischargemay introduce selection bias, potentially reducing the

Figure 1 Flow diagram of participant inclusion in sub-study analysis. CHERISH, Collaborative for Hospitalised Elders,Reducing the Impact of Stays in Hospital.

HGS in nutrition risk screening

© 2018 Dietitians Association of Australia 523

generalisability of our results. The modest sample size con-tributed to reduced precision (indicated by the wide confi-dence intervals observed) and statistical power. As thepresent study was a sub-study of the CHERISH study,24 itwas not designed with malnutrition and HGS as primaryoutcome measures. However, as previously described, post-hoc power analysis revealed the study to be adequatelypowered to detect an AUC of 0.705, which indicates a testwith lower discriminatory ability than that found by Floodet al. in their study of a heterogeneous group of adultpatients.20 It is also to be expected that observed HGS in anacute population will be lower than that of norms derived

from a healthy population. Therefore, it is possible that theuse of a value below the lower limit of the 95% CI of themean from normative data in this study has resulted in highfalse positives, thus reducing discriminatory ability. Simi-larly, due to the low prevalence of severe malnutrition inour sample, we were not able to stratify by degree of mal-nutrition. Therefore, the study may benefit from beingrepeated with a larger sample size. The pragmatic approachemployed resulted in a heterogeneous sample, which couldbe viewed as both a limitation and a strength. A departurefrom highly controlled test conditions is necessary to pro-vide results that are generalisable to a wider population,

Table 1 Demographic and clinical characteristics of participants

Characteristic Total sample (n = 75) Malnourished (n = 14) Well-nourished (n = 61)

Age (years), mean (SD) 74.0 (6.7) 75.0 (7.1) 73.8 (6.6)Male, n (%) 45 (60) 10 (71) 35 (57)Admission category, n (%)

Elective 31 (41) 5 (36) 26 (43)Emergency 44 (59) 9 (64) 35 (57)

Usual place of residence, n (%)Independent living community 75 (100) 14 (100) 61 (100)RACF 0 (0) 0 (0) 0 (0)

CCI1, mean (SD) 2.6 (2.1) 2.7 (1.5) 2.6 (2.2)Weight (kg), mean (SD) 81.1 (19.1) 70.5 (13.4) 83.6 (19.5)BMI2 (kg/m2), n (%)

Underweight 13 (17) 6 (43) 7 (12)Normal weight 27 (36) 5 (36) 22 (36)Overweight 35 (47) 3 (21) 32 (52)

Primary diagnosis, n (%)Metastatic cancer 5 (7) 2 (14) 3 (5)Upper GI cancer 5 (7) 2 (14) 3 (5)Hepatopancreatobiliary cancer 9 (12) 3 (22) 6 (10)Colorectal cancer 15 (20) 2 (14) 13 (21)Other GI disease (non-malignant) 23 (31) 3 (22) 20 (33)Trauma 6 (8) 0 (0) 6 (10)Other 12 (16) 2 (14) 10 (16)

Crude HGS (kgf), mean (SD) 27.5 (8.6) 26.3 (9.4) 27.8 (8.4)Standardised HGS3, mean (SD) 110.1 (26.7) 101.4 (26.4) 112.1 (26.5)

BMI, body mass index; CCI, Charlson Comorbidity Index; GI, gastrointestinal; HGS, hand grip strength; RACF, residential aged carefacility.1 CCI score not age-adjusted.2 BMI ranges for older adults: underweight = <23.0; normal weight = 23.0–26.9; overweight = ≥27.0.3 Standardised HGS: percent of the lower limit of the 95% CI of the mean from age-, gender- and side-specific normative data.

Table 2 Discriminatory ability of screening measures to detect malnutrition (global rating B/C) as measured by the Patient-Generated Subjective Global Assessment

Screeningmeasure

Sensitivity(95% CI)

Specificity(95% CI)

PPV(95% CI)

NPV(95% CI)

AUC(95% CI)

AUCinterpretation

MST 86%(57%– 98%)

80%(68%–89%)

50%(37%–63%)

96%(87%–99%)

0.83(0.71–0.95)

Very good

HGS 50%(23%–77%)

67%(54%–79%)

26%(16%–40%)

85%(77%–91%)

0.41(0.25–0.58)

Test not useful

MST/HGScombined

43%(18%–71%)

93%(84%–98%)

60%(33%–82%)

88%(82%–92%)

0.68(0.51–0.86)

Sufficient

AUC, area under the receiver operating characteristic curve; HGS, hand grip strength; MST, Malnutrition Screening Tool; NPV, negative pre-dictive value; PPV, positive predictive value.

A. Byrnes et al.

524 © 2018 Dietitians Association of Australia

which is the reality of delivering care within the clinical set-ting. This study was unique in that it assessed HGS in anexclusively older population during admission.

HGS was not found to be suitable in screening older inpa-tients for malnutrition during admission to surgical wards, nordid it increase the discriminatory ability of the MST. In light ofthe limitations of our study, however, it is recommended thatour results are interpreted with caution. Screening for nutri-tion risk using an existing validated tool to identify patients forfurther in-depth nutritional assessment by an appropriatelytrained clinician remains the preferred method.

Funding source

This research was supported by an Australian GovernmentResearch Training Program Scholarship for author Byrnes.

Conflict of interest

The authors declare they have no conflicts of interest todisclose.

Authorship

All authors contributed to study design. AB collected, ana-lysed and reported data. AB prepared the manuscript. Allauthors critically reviewed the manuscript. All authors arein agreement with the manuscript and declare that the con-tent has not been published elsewhere.

References

1 Lim SL, Ong KCB, Chan YH, Loke WC, Ferguson M,Daniels L. Malnutrition and its impact on cost of

hospitalization, length of stay, readmission and 3-year mortal-ity. Clin Nutr 2012; 31: 345–50.

2 Agarwal E, Ferguson M, Banks M, Bauer J, Capra S, Isenring E.Nutritional status and dietary intake of acute care patients:results from the Nutrition Care Day Survey 2010. Clin Nutr2012; 31: 41–7.

3 van Stijn MFM, Korkic-Halilovic I, Bakker MSM, van derPloeg T, van Leeuwen PAM, Houdijk APJ. Preoperative nutri-tion status and postoperative outcome in elderly general sur-gery patients: a systematic review. J Parenter Enteral Nutr 2013;37: 37–43.

4 Garth AK, Newsome CM, Simmance N, Crowe TC. Nutritionalstatus, nutrition practices and post-operative complications inpatients with gastrointestinal cancer. J Hum Nutr Diet 2010;23: 393–401.

5 Sun Z, Kong X-J, Jing X, Deng R-J, Tian Z-B. Nutritional riskscreening 2002 as a predictor of postoperative outcomes inpatients undergoing abdominal surgery: a systematic reviewand meta-analysis of prospective cohort studies. PLoS One2015; 10: e0132857.

6 Dudrick SJ. Nutrition management of geriatric surgicalpatients. Surg Clin North Am 2011; 91: 877–96.

7 Young AM, Kidston S, Banks MD, Mudge AM, Isenring EA.Malnutrition screening tools: comparison against two validatednutrition assessment methods in older medical inpatients.Nutrition 2013; 29: 101–6.

8 Kubrak C, Jensen L. Malnutrition in acute care patients: a nar-rative review. Int J Nurs Stud 2007; 44: 1036–54.

9 Meijers JMM, Van Bokhorst-de van der Schueren MAE,Schols JMGA, Soeters PB, Halfens RJG. Defining malnutrition:mission or mission impossible? Nutrition 2010; 26: 432–40.

10 White JV, Guenter P, Jensen G et al. Consensus statement:academy of nutrition and dietetics and American Society forParenteral and Enteral Nutrition: characteristics recommendedfor the identification and documentation of adult malnutrition(undernutrition). J Parenter Enteral Nutr 2012; 36: 275–83.

11 Ottery FD. Definition of standardized nutritional assessment andinterventional pathways in oncology. Nutrition 1996; 12: S15–9.

(a) (b) (c)

Figure 2 Area under the receiver operating characteristic curve (AUC) for different screening measures in identifying mal-nourished patients (Patient-Generated Subjective Global Assessment global rating B/C). (a) Malnutrition screening tool (MST)score ≥2. AUC: 0.83 (95% CI: 0.71–0.95) (P < 0.001). (b) Impaired hand grip strength (value below the lower limit of the95% CI of the mean from age-, gender- and side-specific normative data). AUC: 0.41 (95% CI 0.25–0.58) (P = 0.32).(c) MST/HGS combined (patients with both an MST score ≥2 and impaired HGS). AUC: 0.68 (95% CI: 0.51–0.86)(P = 0.035).

HGS in nutrition risk screening

© 2018 Dietitians Association of Australia 525

12 Smart A. A multi-dimensional model of clinical utility. Interna-tional J Qual Health Care 2006; 18: 377–82.

13 da Silva Fink J, Daniel de Mello P, Daniel de Mello E. Subjec-tive global assessment of nutritional status—a systematicreview of the literature. Clin Nutr 2015; 34: 785–92.

14 Ferguson M, Banks M, Bauer J, Isenring E, Vivanti A, Capra S.Nutrition screening practices in Australian healthcare facilities:a decade later. Nutr Diet 2010; 67: 213–8.

15 Windsor JA, Hill GL. Grip strength: a measure of the proportionof protein loss in surgical patients. Br J Surg 1988; 75: 880–2.

16 Norman K, Stobaus N, Gonzalez MC, Schulzke JD, Pirlich M.Hand grip strength: outcome predictor and marker of nutri-tional status. Clin Nutr 2011; 30: 135–42.

17 Demeny D, Jukic K, Dawson B, O’Leary F. Current practices ofdietitians in the assessment and management of malnutritionin elderly patients. Nutr Diet 2015; 72: 254–60.

18 Guerra RS, Fonseca I, Pichel F, Restivo MT, Amaral TF. Hand-grip strength and associated factors in hospitalized patients.J Parenter Enteral Nutr 2015; 39: 322–30.

19 Guerra RS, Fonseca I, Pichel F, Restivo MT, Amaral TF. Hand-grip strength cutoff values for undernutrition screening at hos-pital admission. Eur J Clin Nutr 2014; 68: 1315–21.

20 Flood A, Chung A, Parker H, Kearns V, O’Sullivan TA. Theuse of hand grip strength as a predictor of nutrition status inhospital patients. Clin Nutr 2014; 33: 106–14.

21 Marr KJ, Shaheen A-A, Lam L, Stapleton M, Burak K,Raman M. Nutritional status and the performance of multiplebedside tools for nutrition assessment among patients waitingfor liver transplantation: a Canadian experience. Clin Nutr2017; 17: 68–74.

22 Garcia MF, Wazlawik E, Moreno YMF, Führ LM, González-Chica DA. Diagnostic accuracy of handgrip strength in theassessment of malnutrition in hemodialyzed patients. e-SPEN J2013; 8: e181–6.

23 Alvares-da-Silva MR, Reverbel da Silveira T. Comparisonbetween handgrip strength, subjective global assessment, andprognostic nutritional index in assessing malnutrition and pre-dicting clinical outcome in cirrhotic outpatients. Nutrition2005; 21: 113–7.

24 Mudge AM, Banks MD, Barnett AG et al. CHERISH (collaborationfor hospitalised elders reducing the impact of stays in hospital):protocol for a multi-site improvement program to reduce geriatricsyndromes in older inpatients. BMC Geriatr 2017; 17: 11.

25 Bauer J, Capra S, Ferguson M. Use of the scored Patient-Generated Subjective Global Assessment (PG-SGA) as a nutri-tion assessment tool in patients with cancer. Eur J Clin Nutr2002; 56: 779–85.

26 Laky B, Janda M, Cleghorn G, Obermair A. Comparison of dif-ferent nutritional assessments and body-composition measure-ments in detecting malnutrition among gynecologic cancerpatients. Am J Clin Nutr 2008; 87: 1678–85.

27 Du H, Liu B, Xie Y et al. Comparison of different methods fornutrition assessment in patients with tumors. Oncol Lett 2017;14: 165–70.

28 Fess EE. Grip strength. In: Casanova JS, ed. Clinical AssessmentRecommendations, 2nd edn. Chicago: American Society of HandTherapists, 1992; 41–5.

29 Sousa-Santos AR, Amaral TF. Differences in handgrip strengthprotocols to identify sarcopenia and frailty—a systematicreview. BMC Geriatr 2017; 17: 238.

30 Firrell JC, Crain GM. Which setting of the dynamometer pro-vides maximal grip strength? J Hand Surg [Am] 1996; 21:397–401.

31 Trampisch US, Franke J, Jedamzik N, Hinrichs T, Platen P.Optimal Jamar dynamometer handle position to assess maxi-mal isometric hand grip strength in epidemiological studies.J Hand Surg [Am] 2012; 37: 2368–73.

32 Bohannon RW, Bear-Lehman J, Desrosiers J, Massy-Westropp N, Mathiowetz V. Average grip strength: a meta-analysis of data obtained with a Jamar dynamometer fromindividuals 75 years or more of age. J Geriatr Phys Ther2007; 30: 28–30.

33 Bohannon RW, Peolsson A, Massy-Westropp N, Desrosiers J,Bear-Lehman J. Reference values for adult grip strength mea-sured with a Jamar dynamometer: a descriptive meta-analysis.Physiotherapy 2006; 92: 11–5.

34 Zhou XH, Obuchowski NA, McClish DK. Measures of diagnos-tic accuracy. In: Statistical Methods in Diagnostic Medicine, 2ndedn. Hoboken, NJ: John Wiley & Sons, Inc., 2011; 13–55.

35 Šimundi�c A-M. Measures of diagnostic accuracy: basic defini-tions. EJIFCC 2009; 19: 203–11.

36 Von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC,Vandenbroucke JP. The strengthening the reporting of observa-tional studies in epidemiology (STROBE) statement: guidelinesfor reporting observational studies. Lancet 2007; 370: 1453–7.

37 Jeejeebhoy KN, Keller H, Gramlich L et al. Nutritional assess-ment: comparison of clinical assessment and objective variablesfor the prediction of length of hospital stay and readmission.Am J Clin Nutr 2015; 101: 956–65.

38 Schaap LA, Pluijm SMF, Deeg DJH et al. Higher inflammatorymarker levels in older persons: associations with 5-year changein muscle mass and muscle strength. J Gerontol A Biol Sci MedSci 2009; 64A: 1183–9.

39 Budziareck MB, Pureza Duarte RR, Barbosa-Silva MCG. Refer-ence values and determinants for handgrip strength in healthysubjects. Clin Nutr 2008; 27: 357–62.

40 Norman K, Schütz T, Kemps M, Josef Lübke H, Lochs H,Pirlich M. The subjective global assessment reliably identifiesmalnutrition-related muscle dysfunction. Clin Nutr 2005; 24:143–50.

41 Cruz-Jentoft AJ, Baeyens JP, Bauer JM et al. Sarcopenia:European consensus on definition and diagnosis: report of theEuropean working group on sarcopenia in older people. AgeAgeing 2010; 39: 412–23.

42 Inouye SK, Studenski S, Tinetti ME, Kuchel GA. Geriatric syn-dromes: clinical, research, and policy implications of a coregeriatric concept. J Am Geriatr Soc 2007; 55: 780–91.

43 Landi F, Liperoti R, Russo A et al. Association of anorexia withsarcopenia in a community-dwelling elderly population: resultsfrom the ilSIRENTE study. Eur J Nutr 2013; 52: 1261–8.

44 Gariballa S, Alessa A. Impact of poor muscle strength on clini-cal and service outcomes of older people during both acute ill-ness and after recovery. BMC Geriatr 2017; 17: 123.

45 Shen Y, Hao Q, Zhou J, Dong B. The impact of frailty and sar-copenia on postoperative outcomes in older patients undergo-ing gastrectomy surgery: a systematic review and meta-analysis.BMC Geriatr 2017; 17: 188.

46 Antunes AC, Araújo DA, Veríssimo MT, Amaral TF. Sarcopeniaand hospitalisation costs in older adults: a cross-sectionalstudy. Nutr Diet 2017; 74: 46–50.

47 Jeejeebhoy KN. Muscle function and nutrition. Gut 1986;27: 25–39.

48 Milne AC, Potter J, Vivanti A, Avenell A. Protein and energysupplementation in elderly people at risk from malnutrition.Cochrane Database Syst Rev 2009: CD003288.

A. Byrnes et al.

526 © 2018 Dietitians Association of Australia

ORIGINAL RESEARCH

Buchanania obovata: An Australian Indigenous foodfor diet diversification

Selina A. FYFE ,1 Michael E. NETZEL,2 Ujang TINGGI,3 Eva M. BIEHL4 and Yasmina SULTANBAWA 2

1School of Agriculture and Food Sciences, The University of Queensland, 2Queensland Alliance for Agriculture andFood Innovation (QAAFI), Health and Food Sciences Precinct, The University of Queensland and 3Queensland HealthForensic and Scientific Services, Brisbane, Australia; and 4Technische Universität München, Freising, Germany

AbstractAim: Buchanania obovata Engl., the Green Plum, is a small green fruit eaten by Australian Indigenous peoples of theNorthern Territory and Western Australia that has had limited study and has potential as a source of food for dietdiversification. The flesh and seed of the fruit are eaten and the plant is used as bush medicine.Methods: Physical characteristics of the fruit were measured. The flesh and seed freeze dried powders weremeasured separately for proximates, mineral/trace elements and heavy metals, and folate analysis. Vitamin C wasanalysed in the flesh.Results: The flesh is high in protein (12.8 g/100 g dry weight (DW)) and both flesh and seed are high in dietary fibre(55.1 and 87.7 g/100 g DW, respectively). The flesh is high in potassium (2274.7 mg/100 g DW), and is a goodsource of magnesium (570.5 mg/100 g DW), calcium (426.0 mg/100 g DW) and phosphorous (216.8 mg/100 g DW),whereas the seed is high in iron (8.15 mg/100 g DW). The flesh contains folate at 752.4 μg/100 g DW and the seedcontains 109.5 μg/100 g DW as pteroylmonoglutamic acid equivalents.Conclusions: The flesh and seed have good nutritional properties and the results support the use of the Green Plumfor diet diversification and nutrition in Indigenous and non-Indigenous populations in Australia.

Key words: Buchanania obovata, food, folate, Indigenous Australia, nutritional profile.

Introduction

The Green Plum is the fruit of the tree Buchanania obovataEngl. and is a native Australian food eaten by IndigenousAustralians. It grows and is wild harvested in the northernparts of Australia, in the Northern Territory and WesternAustralia,1 and is also known as the wild plum, bushmango or wild mango. The Green Plum is a small greenfruit, shown in Figure 1, the flesh is eaten raw from the treeand the flesh and seed are mashed into a paste and eaten.2

The tree B. obovata grows to about 15 m high and hasrough brown bark and branchlets. The leaves are alternateand their shape is obovate to oblong-obovate growing to

5–25 x 1.5–10 cm in size. The flowers are bisexual and arecreamy white in colour.3

Buchanania obovata is in the family Anacardiaceae whichcontains well-known commercialised fruit including mango(Mangifera indica), cashew apple (Anacardium occidentale)and pistachio nuts (Pistacia vera).4

Studies of Australian native foods show they can have verygood nutritional and functional properties. Indigenous foodsof Australia have previously been underutilised but now havepotential for wider availability and commercialisation. TheKakadu Plum (Terminalia ferdinandiana) has such high levelsof ascorbic acid (vitamin C) and high-antioxidant capacity5,6

that it is used in the dietary supplement industry6 and a nat-ural preservative is made from it for commercial dipping ofprawns in Queensland, Australia, giving an extended shelflife.7 Other Australian native foods have been shown tohave high-antioxidant compounds, minerals and vitamins8

indicating that further study of Australian native foods iswarranted.

The Green Plum was selected for this study as it is usedby Australian Aboriginal people as a food and the plant asbush medicine. The Green Plum is commonly eaten bymany Aboriginal communities and is a favourite with chil-dren.2,9 Aboriginal communities have names for it in theirown languages, in Kija it is taluuny,10 in Murinpatha it iskilen, in Jandjung it is djamuru,2 in Rirratjunu it is

S. A. Fyfe, MFoodSciTech, StudentM. E. Netzel, PhD, Senior Research FellowU. Tinggi, PhD, Advanced ChemistE. M. Biehl, BSc, StudentY. Sultanbawa, PhD, Associate ProfessorCorrespondence: Y. Sultanbawa, Queensland Alliance for Agricultureand Food Innovation (QAAFI), The University of Queensland, Healthand Food Sciences Precinct, 39 Kessels Rd, Coopers Plains, PO Box156, Archerfield BC, Brisbane, QLD 4108, Australia.Tel: 07 3443 2471.Email: y.sultanbawa@uq.edu.au

Accepted April 2018

© 2018 Dietitians Association of Australia 527

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 527–532 DOI: 10.1111/1747-0080.12437

dhurrpinda,11 in Djambarrpuyngu it is munydjutj, in Bur-arra it is mulaagi, in Yanyuwa it is bigigee, in Garawa it isbikabaji, in Anindilyakwa it is mangkarrba, in Djinang it isyulwandi, in Alawa it is yulmuru, in Rirratjunu it is dhurr-pinda3 and the people of Groote Eylandt call it man-gkarrkba.12 Young stems, leaf ribs, inner bark from youngbranches and older stems have been used as bush medicinefor their antiseptic and analgesic qualities to treat toothache,skin infections and conditions and as an eye lotion.3,12

Indigenous communities in Australia have had very poorquality diets for at least three decades.13 Australian Aborigi-nals are at increased risk of malnutrition and chronic dis-ease causing poorer levels of health than non-IndigenousAustralians.14 Diabetes, cardiovascular disease and tobacco-related illnesses account for half of the health gap in theAustralian Indigenous population.15 Diversity in nationalfood system has been shown to increase human health andnutrition.16 Indigenous traditional foods could add diversityin nutrition to the Australian food system and be a goodsource of nutrients and micronutrients. Wider availabilityand a better understanding of native Australian foods couldlead to a higher intake by Australian Indigenous peoplesproviding better nutrition and health outcomes.

This is the first study of this kind on the Green Pluminvestigating so much of its nutritional properties for die-tary purposes. The data presented in this study are novelexcept for a small number of proximate and inorganic con-stituents and vitamin C of the flesh which were publishedin 1993,17 and mineral analysis which were done by Medleyand Bollhöfer and provided as supplementary material intheir study on the radium uptake in the tree.9 Including themineral analysis data in this study is important as theirstudy focused on the tree whereas this study is investigatingthe flesh and seed of the fruit for dietary purposes and alsoanalysed the proximates, vitamin C and folate, making thenutrition profile available to dietitians, nutritionists and thefood industry. Vitamin C and folate were chosen as vitaminsfor testing in this initial screening of the Green Plum as

vitamin C is a key vitamin in fruit, and folate, a critical vita-min, has previously been found in considerable amounts ina number of Australian native bush foods.8

Methods

Chemicals used in the analysis were from Sigma-Aldrich(St Louis, MO, USA) and Chem-Supply (Gillman, SA,Australia).

Approximately 800 individual B. obovata fruit were ran-domly picked slightly under-ripe from several trees locatednear Darwin, Northern Territory, Australia, in 2015, withthe assistance of the experienced Thamarrurr Rangers andChris Brady from the Northern Territory. The fruit werepicked under-ripe due to availability and ease of transporta-tion. They were frozen and transported to Brisbane wherethey remained frozen at −20�C. Samples for physical prop-erties and moisture content were randomly selected andthawed to room temperature before testing. From the sam-ple, 554 g of composite whole fruit with the seed intact(approximately 700 Green Plums) were lyophilised in aChrist Gamma 1–16 LSC Freeze Drying Unit (John MorrisScientific, Osterode, Germany) and the flesh was ground toa powder in a kitchen food processor. The seeds were sepa-rated from the flesh powder with a sieve and finely groundin a hammer mill (Lab Mill, Christy and Norris Ltd.Chelmsford, England). The composite flesh powder andseed powder were stored at room temperature in separateair tight plastic containers, protected from light.

Size of whole fruit: Ten Green Plums were chosen at ran-dom and measured for length, width and depth (heightlying on a bench) using 150 mm Digital Calliper (CraftrightEngineering Works, Jiangsu, China).

Weight: Ten whole fruit were weighed on laboratoryscales (Sartorius CP224S, Gottingen, Germany). The fruitwere cut down one side with a knife and the flesh peeledoff the seed, then weighed separately and the pulp/seedratio was calculated.

Figure 1 Green Plums (a) whole and (b) flesh and seed separated.

S. A. Fyfe et al.

528 © 2018 Dietitians Association of Australia

Colour: Colour was measured with a Konica MinoltaCR-400 Chroma Meter and DP-400 Data Processor (Osaka,Japan) on 10 Green Plums chosen for optimal testing. Twomeasurements were averaged for each result. The colour ofthe inside flesh of 10 pieces of fruit were measured, withboth measurements from the same area of flesh. Lightness(L*), Chroma (C*) and hue angle (h) were recorded.

Moisture Content: Moisture content was determined usingthe AOAC Official Method on Analysis 920.151 (1990) intriplicate on pieces of flesh and seed. Flesh of 10 GreenPlums were cut into small pieces and the seeds separatedand broken up and approximately 0.5 g samples of fleshand seed were placed in a vacuum drying oven (Series RVT,Heraeus, Berlin, Germany) at pressure 200 Torr and tem-perature 70 �C, until constant weight.

Proximate analysis: Proximate analysis was performed onthe freeze dried powders of both the flesh and seed of theGreen Plum at Symbio Alliance, Eight Mile Plains, Queens-land, Australia, a National Association of Testing Authori-ties (NATA) accredited laboratory that complies withISO/IEC 17025:2005. The analysis were done according tothe NATA approved in house methods or AOAC methods:protein by AOAC method 990.03 (AOAC, 1997); fat byAOAC method 991.36 (AOAC, 1999); saturated, mono-unsaturated, poly-unsaturated and trans fat by in housemethod CFH068.2; moisture by AOAC 934.01 (AOAC,1990), ash by AOAC method 923.03 (AOAC, 2000); totalsugar by CFH001.1, total dietary fibre by CF057; availablecarbohydrate by CF029.1; energy by CF030.1; crude fibreby AOAC 962.09 (AOAC, 1990) and dry matter by inhouse method CF006.1.

Minerals/Trace Elements/Metals: The methods used foranalysis of minerals and trace elements/metals have previ-ously been described in Carter et al.18 The analysis was car-ried out in duplicate using a 7700 inductively coupledplasma (ICP)-MS (Agilent Technologies Australia Pty. Ltd.,VIC, Australia) and ICP-OES (Agilent TechnologiesAustralia Pty. Ltd., VIC, Australia) after microwave diges-tion (MarsXpress, CEM Corporation, Matthews, NC, USA).

Vitamin C: Vitamin C was tested on the freeze dried fleshpowder by Symbio Alliance using the AOAC OfficialMethod 967.21 on HPLC.

Folate: Folate in the freeze-dried Green Plum flesh andseed powders were analysed in triplicate as outlined by Ringl-ing and Rychlik19 using a UHPLC-MS/MS (Shimadzu, Rydal-mere, NSW, Australia) with a Raptor ARC-18 column(Restek, Bellefonte, PA, USA). The derivatives pteroylmono-glutamic acid (PteGlu), tetrahydrofolate (THF),5-methyltetrahydrofolate (5mTHF), 5-formyltetrahydrofolate(5fTHF) and 10-formyl-pteroylglutamic acid (10f PteGlu)were measured by stable isotope dilution assays.

Data analysis and calculations were done using MicrosoftExcel software, version 2013 (Microsoft Corporation, Red-mond, WA, USA). Arithmetic means � SDs are shown intext and tables.

Results

Ten whole Green Plums were measured with the follow-ing characteristics: length 15.07 � 1.14 mm, width11.54 � 0.90 mm and depth 9.14 � 0.96 mm.

Ten whole Green Plums had an average weight of0.75 � 0.13 g. When the flesh was removed from the stoneit separated easily by hand, with only a few small shards offlesh clinging to the seed. The flesh of each Green Plumaveraged 0.39 � 0.09 g and the seed 0.22 � 0.04 g, theseed was 29.7% of the whole fruit by weight and the pulp/seed ratio was 1.75.

The outer skin colour was measured as L* 43.99 � 2.62,C* 29.02 � 2.41 and h 105.36 � 2.12 (n = 10). The innerflesh was L* 49.70 � 2.62, C* 35.83 � 1.86 and h105.28 � 1.67 (n = 10). The Green Plum fruits are a dullcolour with a yellow/green hue. The hue of the skin and theinside flesh is very similar, however, the flesh appears to beslightly more vivid than the skin and lighter in brightness.20

Both the outer skin and inner flesh were fairly uniform incolour across all 10 Green Plums tested.

Moisture content of the Green Plum flesh was around79% and of the seed around 35%.

The results of the proximate analysis on the lyophilisedpowders, as measured by Symbio Alliance, are shown inTable 1.

The minerals/trace elements and metal results of the fleshand seed powders are shown in Table 2.

Five different vitamers of folate were measured in theGreen Plum flesh and seed powders with results in Table 3.Vitamin C was measured in the Green Plum flesh and theresult is in Table 3.

Discussion

The proximate results indicate that both the flesh and seedcan be used as a source of total dietary fibre. The total die-tary fibre % daily intake for a 100 g serve DW of an8700 kJ diet is 184% for the flesh and 292% for the seed(Symbio Alliance nutrition information report received withresults). The protein in the flesh is 26% of the daily intake

Table 1 Proximate results of the freeze dried Green Plumflesh and seed

Proximate (g/100 g DW) Flesh Seed

Protein 12.8 3.0Fat 2.5 1.8Saturated fat 0.5 0.6Mono-unsaturated fat 0.7 0.3Poly-unsaturated fat 1.2 0.8Trans fat <0.01 <0.01Moisture (air) 2.2 <0.1Ash 6.1 0.8Total sugar 2.9 <0.10Dietary fibre (total) 55.1 87.7Avail carbohydrate 21.3 6.7Energy (kJ/100 g DW) 1112 934Crude fibre 7.7 55.3Dry matter 97.8 100.0

Buchanania obovata: An Australian Indigenous food

© 2018 Dietitians Association of Australia 529

per 100 g serve DW. Current remote Indigenous diets arelow in protein contributing only 12.5–14.2% of energy,lower than the recommended 15–25%.13 Through dietdiversification the Green Plum can contribute to improve-ment in protein intake but is not a replacement for tradi-tional sources of proteins.

The results in Table 2 show the Green Plum can be usedas a dietary source of potassium and other macro-mineralsand contains a variety of trace elements essential to thehuman diet. Compared to recommended daily intakes theGreen Plum is a good source of many nutrients, particularlythe flesh which is high in calcium, potassium, magnesium,phosphorous, zinc, selenium, copper and manganese.21

The seed has lower amounts of most nutrients than theflesh but contains higher amounts of iron.

The study by Medley and Bollhöfer9 on the uptake ofradium by the B. obovata tree did a metals and trace ele-ments analysis of flesh and seeds from seven different treesin seven different sites. Their results are similar to those inthis study, with the potassium in the flesh slightly lowerthan in this study (average 1660.0 vs 2274.7 mg/100 gDW). The group II metals, calcium, barium, strontium andmagnesium were all higher in their study in the flesh andseed, except for magnesium which gave similar results tothese. Other Australian native foods also have high-mineralcontent. Of those studied by Konczak and Roulle,22 theGreen Plum flesh has higher levels of calcium, magnesium,phosphorous and selenium, and only the Quandong(3456.2 mg/100 g DW) has higher levels of potassium.

Vitamin C in the Green Plum flesh (170 mg/100 g DW)is similar to that of the mango (220 mg/100 g DW)(reported by the USDA ARS and calculated to DW usingmoisture content),23 but is lower than the Kakadu plumwhich has levels of up to 22 490 � 5290 mg/100 g DW.6

The flesh is much higher in folate then any of theAustralian native fruit tested by Konczak et al. including theAustralian Desert Lime (420 μg/100 g DW) and the seedhas similar levels to the Kakadu Plum, Riberry and LemonAspen (around 110 μg/100 g DW).8 The flesh has muchhigher folate levels than most of the foods measured usingthe same method by Ringling and Rychlik,19 and the resultsare higher than all of the fruit results compiled by Sainiet al. from USDA data.24 However, some Asian green vege-tables such as tatsoi (231 μg/100 g FW), and English spin-ach (213 μg/100 g FW), have higher levels than the GreenPlum flesh (161 μg/100 g FW).25 The low amounts of Pte-Glu found compared to the other vitamers is consistentwith it not generally occurring naturally in food.26

Folate is a B vitamin that is synthesised in plants and isessential for human health due to its role in synthesising,repairing and methylating DNA.24 Folate deficiency is asso-ciated with neural tube defects24 non-alcoholic fatty liverdisease and steatohepatitis,27 and it is a risk factor for vas-cular disease.28 The cells in the human body are unable tosynthesise folate so there must be sufficient dietary intake.24

Folate produced in plants is preferred over synthetic folicacid,24 which may have negative side effects in humans.29

The Green Plum is very high in 5mTHF which is the bioac-tive folate form in the body.24

There are challenges for Indigenous communities as thecurrent food supply discourages healthy eating. Indigenouscommunities are remote and healthy food is relativelyexpensive compared to unhealthy choices.30 The price dif-ference of healthy food in remote Northern Territory can be

Table 2 Minerals/trace elements and metal analysis of theGreen Plum flesh and seed freeze dried powders

Element (mg/100 g DW) Flesh Seed

Macro-mineralsSodium 188.1 55.6Potassium 2274.7 285.1Calcium 426.0 46.8Magnesium 570.5 49.5Phosphorus 216.8 45.5

Trace elementsSelenium 0.022 0.0044Zinc 2.6 0.38Iron 3.8 8.15Manganese 1.5 0.32Copper 0.54 0.72Nickel 0.082 0.050Cobalt 0.018 0.0030Chromium 0.0038 0.067Vanadium 0.0011 0.0031Boron 1.4 0.23Molybdenum 0.013 0.0074Strontium 2.8 0.37Barium 1.1 0.17Antimony <0.001 <0.001Silver <0.001 <0.001Aluminium 0.59 0.36

Heavy metalsMercury <0.001 0.01209Lead 0.014 0.0043Cadmium <0.001 <0.001Arsenic <0.001 0.0090

Table 3 Vitamers in the Green Plum flesh and seed freezedried powders

Vitamer Flesh Seed

Folate Type (μg/100 g DW)PteGlu ND 2.9 � 0.8THF 28.6 � 0.4 12.3 � 1.25mTHF 505.8 � 13.4 59.6 � 1.55fTHF 211.3 � 0.5 29.6 � 1.410f PteGlu 6.7 � 0.5 5.1 � 0.2Total Folate 752.4 � 12.5 109.5 � 0.4Vitamin C (mg/100 g

DW)170 NA

ND, not detected; NA, not available. Folate derivatives measured:PteGlu, pteroylmonoglutamic acid; THF, tetrahydrofolate; 5mTHF,5-methyltetrahydrofolate; 5fTHF, 5-formyltetrahydrofolate; 10fPteGlu, 10-formyl-pteroylglutamic acid, all reported as μg/100 gDW PteGlu Equivalents of means � SD (n = 3).

S. A. Fyfe et al.

530 © 2018 Dietitians Association of Australia

60% higher than in urban Australia.31 The diets of manyremote Aboriginal people are characterised by high fat andhigh-sugar processed foods that are cheap and have a ‘long-life’ including bread and tinned meat.32

Remote Australian Aboriginal communities have verypoor dietary nutrition profiles, and many key nutrientscome from poor-quality nutrient-fortified processedfoods.13 Lee et al. found that in two remote Aboriginalcommunities over 26 years there was an increase in theintake of folate from fortified bread, and an increase of nia-cin and thiamin from fortified breakfast cereals.30 Brimble-combe et al. studied the diets of three remote NorthernTerritory Aboriginal communities and found that fortifiedwhite bread was a major source of fibre, protein, folate,sodium, calcium, potassium and other micronutrients inthe diet in all of the communities. The overall diets in allcommunities were insufficient in fibre, potassium, magne-sium and calcium for the weighted estimated averagerequirement due to the consumption of fortified processedfoods and the minimal intake of traditional foods.13

The nutrients found in Australian Indigenous fruitscould be used as a method of combatting dietary deficien-cies and increasing the diversity and nutrition in the diet.Adding or increasing the availability of Green Plum in thefood supply chain could increase the natural intake offolate, fibre, potassium and magnesium that are insufficientin the current diet.13 Traditional foods are culturally impor-tant in Indigenous communities and are based on valuesand tastes.33 They are seen as a way of achieving a balancedlife, and the decreasing use of them is associated with dete-riorating health in Indigenous cultures.34 Australian nativetraditional foods could be part of the response needed inthe preventive and curative efforts required for health gainin the Indigenous population.15 They provide the ‘balance’and ‘freshness’ that Indigenous Australians associate with‘good food’.32 Increasing the availability of supply, and edu-cating on the nutrients available in the Green Plum andother Australian native foods could lead to an increasein the nutrition and health in remote Indigenous communi-ties, and possibly a replacement of bio-fortified foods.They can be a source of income for remote communitiesfrom the wild harvesting and processing of them incommunities,35 enabling them to be further used, and theirhealth and nutritional properties more widely available forAboriginal communities and non-Indigenous peoples ofAustralia. This work will be useful for commercialisation ofthe Green Plum, for use by dietitians and as a basis for fur-ther studies. As the fruit was picked only from one geo-graphical location for this initial study, further studiesshould be done to assess variation in nutrients based onlocation and other environmental and genetic factors.

In conclusion, this initial study indicates the Australiannative Green Plum has promising nutritional compositionin both its flesh and seed. It is high in fibre, folate andpotassium and contains essential minerals and trace ele-ments which can contribute to diet diversification and havea beneficial impact on health in Aboriginal communities.However, further testing of the Green Plum is warranted to

elucidate the full nutritional profile and dietary value forboth the Aboriginal and wider communities.

Funding source

This project was funded by the University of Queensland,Brisbane, Australia.

Conflict of interest

The authors declare that they have no conflict of interest.

Authorship

SAF was responsible for sample preparation, physical andchemical analysis, data interpretation and the writing of themanuscript. MN supplied editorial assistance. UT wasresponsible for analysis of Minerals/Trace Elements andMetals. EMB was responsible for the Folate analysis. YS wasresponsible for project concept, design and supervision. Allauthors are in agreement with the manuscript and declarethat the content has not been published elsewhere. A con-ference abstract was submitted to Tropical Agriculture 2017and the results were presented at this meeting.

References

1 McMahon G. Bushtucker in the Top End. Department of PrimaryIndustry Fisheries and Mines, ed. Northern Territory,Australia: Northern Territory Government, 2005.

2 Head L. Country and garden : ethnobotany, archaeobotanyand Aboriginal landscapes near the Keep River, northwesternAustralia. J Soc Archaeol 2002; 2: 173–96.

3 Barr A, Chapman J, Smith N et al. Traditional Aboriginal medicinesin the Northern Territory of Australia, by Aboriginal Communities ofthe Northern Territories. Darwin, NT: Conservation Commissionof the Northern Territory of Australia, 1993.

4 Özçelik B, Aslan M, Orhan I, Karaoglu T. Antibacterial, anti-fungal, and antiviral activities of the lipophylic extracts of Pista-cia vera. Microbiol Res 2005; 160: 159–64.

5 Williams DJ, Edwards D, Pun S et al. Organic acids in Kakaduplum (Terminalia ferdinandiana): the good (ellagic), the bad(oxalic) and the uncertain (ascorbic). Food Res Int 2016; 89:237–44.

6 Konczak I, Maillot F, Dalar A. Phytochemical divergence in45 accessions of Terminalia ferdinandiana (Kakadu plum). FoodChem 2014; 151: 248–56.

7 Sultanbawa Y. Food preservation and the antimicrobial activityof Australian native plants. In: Sultanbawa Y, Sultanbawa F,eds. Australian Native Plants: Cultivation and Uses in the Healthand Food Industries. Boca Raton, FL: CRC Press, Taylor & Fran-cis Group, 2016; 251–63.

8 Konczak I, Zabaras D, Dunstan M, Aguas P, Roulfe P, Pavan A.Health Benefits of Australian Native Foods: An Evaluation ofHealth-Enhancing Compounds. Barton, Australia: Rural IndustriesResearch and Development Corporation, 2009.

9 Medley P, Bollhöfer A. Influence of group II metals onRadium-226 concentration ratios in the native green plum(Buchanania obovata) from the Alligator Rivers Region, North-ern Territory, Australia. J Environ Radioact 2016; 151: 551–7.

Buchanania obovata: An Australian Indigenous food

© 2018 Dietitians Association of Australia 531

10 Martin S. Bush Tukka Guide. Prahran, Australia: Explore AustraliaPublishing Pty Ltd, 2014.

11 Whitehead PJ, Gorman J, Griffiths AD, Wightman G,Massarella H, Altman J. Feasibility of Small Scale CommercialNative Plant Harvests by Indigenous Communities. Kingston,Australia: Rural Industries Research and Development Corpora-tion, 2006.

12 Levitt D. Plants and People : Aboriginal Uses of Plants on GrooteEylandt. Australian Institute of Aboriginal Studies: Canberra,Australia, 1981.

13 Brimblecombe JK, Ferguson MM, Liberato SC, O’Dea K. Char-acteristics of the community-level diet of Aboriginal people inremote northern Australia. Med J Aust 2013; 198: 380–4.

14 Morris NF, Stewart S, Riley MD, Maguire GP. The IndigenousAustralian Malnutrition Project: the burden and impact of mal-nutrition in Aboriginal Australian and Torres Strait Islanderhospital inpatients, and validation of a malnutrition screeningtool for use in hospitals—study rationale and protocol. Spring-erPlus 2016; 5: 1296.

15 Vos T, Barker B, Begg S, Stanley L, Lopez AD. Burden of diseaseand injury in Aboriginal and Torres Strait Islander Peoples: theIndigenous health gap. Int J Epidemiol 2009; 38: 470–7.

16 Remans R, Wood SA, Saha N, Anderman TL, DeFries RS. Mea-suring nutritional diversity of national food supplies. Glob FoodSec 2014; 3: 174–82.

17 Brand Miller J, James KW, Maggiore PMA. Tables of Composi-tion of Australian Aboriginal Foods. Australian Institute of A,Torres Strait Islander S, ed. Canberra, Australia: AboriginalStudies Press, 1993.

18 Carter J, Tinggi U, Yang X, Fry B. Stable isotope and tracemetal compositions of Australian prawns as a guide to authen-ticity and wholesomeness. Food Chem 2015; 170: 241–8.

19 Ringling C, Rychlik M. Analysis of seven folates in food byLC–MS/MS to improve accuracy of total folate data. Eur FoodRes Technol 2013; 236: 17–28.

20 Konica Minolta Sensing Inc. Precise Color Communication: ColorControl from Perception to Instrumentation. Sakai Osaka, Japan:Konica Minolta Sensing Inc, 2007.

21 National Health and Medical Research Council. Nutrient refer-ence values for Australia and New Zealand Australia: Common-wealth of Australia, 2006. (Available from: https://www.nrv.gov.au/home, accessed 20 February 2017).

22 Konczak I, Roulle P. Nutritional properties of commerciallygrown native Australian fruits: lipophilic antioxidants and min-erals. Food Res Int 2011; 44: 2339–44.

23 United States Department of Agriculture Agricultural ResearchService. Full Report (All Nutrients) 09176, Mangos, raw USA:

United States Department of Agriculture Agricultural ResearchService, 2016. (Available from: https://ndb.nal.usda.gov/ndb/foods/show/2271?n1=%7BQv%3D1%7D&fgcd=&man=&lfacet=&count=&max=50&sort=default&qlookup=Mangos%2C+raw&offset=&format=Full&new=&measureby=&Qv=1&ds=&qt=&qp=&qa=&qn=&q=&ing=, accessed 20 February 2017).

24 Saini RK, Nile SH, Keum Y-S. Folates: chemistry, analysis,occurrence, biofortification and bioavailability. Food Res Int2016; 89: 1–13.

25 Houlihan A, Pyke M, Scheelings P et al. Folate Content of AsianVegetables. Rural Industries Research and Development Corpora-tion. Canberra, Australia: Australian Government, 2011.

26 Delchier N, Herbig AL, Rychlik M, Renard CM. Folates in fruitsand vegetables: contents, processing, and stability. Compr RevFood Sci Food Saf 2016; 15: 506–28.

27 Silva RP, Kelly KB, Al Rajabi A, Jacobs RL. Novel insights oninteractions between folate and lipid metabolism. Biofactors2014; 40: 277–83.

28 Kolb AF, Petrie L. Folate deficiency enhances the inflammatoryresponse of macrophages. Mol Immunol 2013; 54: 164–72.

29 Smith AD, Refsum H, Selhub J, Rosenberg IH. Decision onfolic acid fortification in Europe must consider both risks andbenefits. BMJ 2016; 352: i734.

30 Lee A, Rainow S, Tregenza J et al. Nutrition in remote aborigi-nal communities: lessons from Mai Wiru and the Anangu Pit-jantjatjara Yankunytjatjara lands. Aust N Z J Public Health 2015;40: S81–8.

31 Henryks J, Brimblecombe J. Mapping point-of-purchase influ-encers of food choice in Australian remote indigenous commu-nities: a review of the literature. SAGE Open 2016; 6:2158244016629183.

32 Colles SL, Maypilama E, Brimblecombe J. Food, food choiceand nutrition promotion in a remote Australian Aboriginalcommunity. Aust J Prim Health 2014; 20: 365–72.

33 Islam D, Berkes F. Indigenous peoples’ fisheries and food secu-rity: a case from northern Canada. Food Security 2016; 8:815–26.

34 Brimblecombe J, Maypilama E, Colles S et al. Factors influenc-ing food choice in an Australian aboriginal community. QualHealth Res 2014; 24: 387–400.

35 Gorman J, Courtenay K, Brady C. Production of Terminaliaferdinandiana Excell. (’Kakadu Plum’) in Northern Australia.In: Sultanbawa Y, Sultanbawa F, eds. Australian Native Plants:Cultivation and Uses in the Health and Food Industries. BocaRaton, FL: CRC Press, Taylor & Francis Group, 2016;89–103.

S. A. Fyfe et al.

532 © 2018 Dietitians Association of Australia

ORIGINAL RESEARCH

First Nations students’ perceptions of school nutritionpolicy implementation: A mixed methods study

Christina GILLIES ,1 ALEXANDER RESEARCH COMMITTEE,2 Anna FARMER,1 Katerina MAXIMOVA3

and Noreen D. WILLOWS11Department of Agricultural, Food, and Nutritional Science, Edmonton 2Alexander First Nation Education, Morinvilleand 3School of Public Health, University of Alberta, Edmonton, Alberta, Canada

AbstractAim: School nutrition policies can improve healthy food access for Indigenous First Nations children in Canada. Thisstudy explored First Nations students’ perceptions of a school nutrition policy.Methods: The research was a process evaluation of school nutrition policy implementation using a mixed-methodsdesign. Students in grades 4–12 (n = 94) completed a 17-question survey to capture their perceptions of the policy.Survey data informed an 11-question semi-structured interview guide. Transcripts from interviews with students(n = 20) were analysed using content analysis to identify barriers and facilitators to policy implementation.Results: Key facilitating factors to policy implementation were student support for the policy and taste preferences.Most students (87%) agreed that only healthy foods should be served at school and, in interviews, expressed a pref-erence for healthy food choices. Barriers to policy implementation included foods available at school and lack ofcommunication between students and their teachers and parents. Half (50%) of surveyed students reported that theireating habits at school were average; interviews explained that their diets could be improved by consuming morefruit and vegetables at school. Both surveys and interviews found that communication between students and theirparents and teachers about what they ate and drank at school was low.Conclusions: To support children’s healthy eating at school, the school nutrition policy could provide clear guide-lines on foods permissible in the school, while considering social and environmental barriers to healthy eating. Theinvolvement of First Nations children in the implementation and evaluation of school nutrition policies isrecommended.

Key words: Canada, children, Indigenous, nutrition policy, school, students.

Introduction

In Canada, Indigenous children (First Nations, Métis andInuit) often have poor dietary behaviours putting them at ahigher risk for obesity and obesity-related chronic diseasescompared to their non-Indigenous counterparts.1,2 FirstNations communities face barriers to healthy eating at mul-tiple levels of influence (i.e. community, built environment,social policy), including poverty, food insecurity and geo-graphic isolation.3 Effective strategies to promote healthyeating behaviours among First Nations children are needed

that recognise the multiple factors that constrain theiraccess to healthy food.

A comprehensive school health (CSH) framework (alsoknown as health promoting schools) is a multi-factoredapproach that supports improvements in both health andeducation through improved social and physical environ-ments.4 A school nutrition policy (SNP) is an integral com-ponent of the broader CSH approach as it establishesformal standards for nutrition-related aspects of the schoolenvironment, including foods available, health educationand community and family involvement.5 By implementingSNPs, schools have the opportunity to create environmentsthat enable and encourage healthy eating behaviours, whichmay support lifelong health behaviours and improvedhealth outcomes for First Nations children.6

In 2014, Kipohtakaw Education Centre (KEC) adopted aSNP intended to promote healthy food choices in theschool (Appendix I). KEC is a locally-controlled, band-operated Cree (one group of First Nations peoples) kinder-garten to grade 12 school located on reserve lands. Follow-ing adoption of the SNP, the school’s administrators soughtto explore the SNP implementation process through an

C. Gillies, MA, PhD CandidateA. Farmer, PhD, Associate ProfessorK. Maximova, PhD, Associate ProfessorN.D. Willows, PhD, Associate ProfessorCorrespondence: N.D. Willows, Faculty of Agricultural, Life &Environmental Sciences, University of Alberta, 4-378 Edmonton ClinicHealth Academy, Mailbox #54, 11405 87 Avenue, Edmonton, AB T6G2P5, Canada.Email: noreen.willows@ualberta.ca

Accepted October 2018

© 2018 Dietitians Association of Australia 533

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 533–540 DOI: 10.1111/1747-0080.12499

evaluation that included key stakeholders to identify areasfor improvement. As the main recipients and beneficiariesof SNPs, it is important to understand students’ perceptionsof the policy and of the school food environment. Studentshave the potential to identify their self-perceived barriers tohealthy eating at school and can provide ideas to overcomechallenges associated with SNP implementation; thus, theirparticipation in policy evaluation may result in increasedfeasibility, acceptance, and overall success of SNPs.7 Thepurpose of this study was to explore students’ perceivedbarriers to and facilitators of a school nutrition policy in aFirst Nations community.

Methods

A community-based participatory research (CBPR)approach was adopted in which community members ofthe Alexander Research Committee (ARC), a well-established research committee in Alexander First Nation,Alberta, Canada,8 worked in close collaboration with uni-versity researchers on all stages of the research. The use of aCBPR approach ensured that the research followed culturalprotocols, reflected local context, and worked to addressneeds identified by community educators. Ethics approvalwas obtained from Research Ethics Board 1 at the Univer-sity of Alberta (Pro00051837).

The research was a process evaluation that focused onfinding areas to improve the current SNP implementationprocess.9 The evaluation was conducted using an explana-tory sequential mixed-methods design, which is recognisedas particularly effective in process evaluation research dueto its ability to yield richer detail about implementationthan quantitative and qualitative methods could alone.10 Inthe first phase, a cross-sectional survey was used to capturea broad understanding of students’ perceptions, attitudes,and experiences. Community members and academicresearchers collaboratively developed a 17-question surveyconsisting of 1 open-ended question (‘what is your grade?’)and 16 closed-ended questions, including: 4 dichotomousquestions; 1 Likert scale question; 3 interval scale questions;2 semantic differential scale questions; and 6 frequencyscale questions. The survey collected information on stu-dents’: gender (1 question); grade level (1 question); per-ceptions of the policy and personal eating habits(3 questions); desirability of foods served and sold at school(3 questions); utilisation of the school food programs(3 questions); communication with parents and teachers(4 questions); perception of the healthiness of foods (1 ques-tion); and desire to be served and sold certain foods(1 question).

All students in grades 4–12 (n = 126) were eligible tocomplete the anonymous survey. As per school policy,parental consent was not required because the survey hadbeen developed to inform school programming. Students inkindergarten to grade 3 were not eligible due to time andresource constraints. The survey was read out to studentsin grades 4–6, while students in grades 7–12 completed the

survey on their own with the same researcher (CG) presentto answer questions.

Response frequencies (%) were calculated using Statisti-cal Package for the Social Sciences version 22.0. Commu-nity members and academic partners discussed the resultsof the surveys and identified findings that were of interestand warranted further investigation. Thus, the quantitativeresults were used to develop the qualitative follow-up ques-tions. Academic partners drafted the initial interview guidebased on their expertise in interviewing and brought it backto the community members for approval, resulting in an11-question semi-structured interview guide (Table 1).

All students in grades 4–12 were eligible to be inter-viewed if they had parental consent. To identify students tobe interviewed, each consent form was assigned a number,and numbers were drawn from each grade level usingMicrosoft Excel random selection. Interviews were con-ducted by an experienced interviewer (CG) in a privateroom at the school. To reduce social desirability bias, theinterviewer began each interview by explaining to studentsthat they were not being tested. Students were also told thatthe interview was confidential, and that their responseswould not be shared by name with their teachers, parents,or friends. The interviews were audio-recorded with partici-pants’ consent. The interviewer had no previous formalrelationship with participants; however, she may have beenrecognised by them as the same person responsible for sur-vey administration. All interviewed students received a $25(CAD) gift card for their participation. Interview recruit-ment continued until theoretical saturation, or the point atwhich no new information emerged, was reached.11

All interviews were transcribed verbatim with identifiersremoved and analysed using ATLAS.ti version 8.0. Tran-scripts were analysed by the same researcher who con-ducted the interviews (CG) using content analysis togenerate knowledge based on participants’ perspectives.12

Guided by the approach outlined by Creswell,12 the analy-sis process included: (i) organising and preparing tran-scripts and field notes, (ii) reading each transcript to gainfamiliarity with the data, (iii) line-by-line open coding usingcodes derived from the actual language used in the data(in vivo), (iv) generating themes, (v) describing themes andsupporting them through transcript excerpts, and(vi) making interpretations of the findings.

Facilitators and barriers to SNP implementation wereorganised using the socio-ecological framework3 during thefinal step of analysis. This is a theory-based framework forunderstanding and describing the reciprocal interactionsand interrelationships that occur between an individual’spersonal dimensions and larger environment (e.g. socialand physical environments).3 In this study, it was used tobring together the themes that naturally emerged from thedata and establish the different levels of influence that eachfactor had on SNP implementation. Individual factors referto the influence of students’ knowledge, attitudes, andbeliefs on SNP implementation. Interpersonal factors referto the influence of family and school staff on SNP imple-mentation. Community, home and sociocultural

C. Gillies et al.

534 © 2018 Dietitians Association of Australia

environment factors refer to the influence of the school,community, and household food environments on SNPimplementation.

The researcher who administered and interpreted theinterviews (CG) is an anthropologist with experience usingqualitative methods to explore how people think about foodand health within local contexts.13–15 These characteristicsmay have influenced the interpretation of results and meth-odological decisions, including the use of the socio-ecological framework to organise results. To mitigate theeffects of individual researchers, trustworthiness of qualita-tive data was established through the principles of credibil-ity, transferability, dependability and confirmability.16 Toestablish credibility, academic interpretations of the datawere corroborated by community members on the ARC anda group of community Elders. Descriptions of the study set-ting, methodology, and data interpretations will allow otherresearchers to determine if and how findings may transfer toother school contexts. Dependability was established by theresearcher checking transcripts several times and defining

codes to ensure that they were used consistently. Finally,confirmability was established by maintaining an audit trailincluding minutes of ARC meetings and data analysis files.

Results

In total, 94 students in grades 4–12 completed the survey(response, 75%). Participants were 49% female and 51%male, and 44% were in grades 4–6 and 56% were ingrades 7–12.

Student responses to questions about healthy foodsaligned with national dietary recommendations.17 Forexample, all students responded that leafy greens and fruitwere healthy and that candy and chocolate bars wereunhealthy. When asked their preferred food and drinks tobe served and sold in school, students’ choices aligned withnational dietary recommendations guidelines17 (Table 2).Preferred foods also included traditional foods of their Creeculture, such as berries (e.g. raspberries and saskatoonberries), baked bannock (a quick bread and staple food in

Table 1 Semi-structured interview guide

Question Probes

1 If you could have anything to eat or drink at home, what would youwant? Why would you choose these foods?

Types of foods and drinks child likes tohave for breakfast, lunch, dinner, orcelebrations; taste; social importance

2 Tell me about some food and drinks you think are healthy. Why doyou think they are healthy?

Food preparation; bannock; meat;rabbit/moose/deer/duck/bush chicken;fish; sports and energy drinks

3 Tell me about some food and drinks you think are unhealthy. Why doyou think they are unhealthy?

Food preparation; bannock; meat;rabbit/moose/deer/duck/bush chicken;fish; sports and energy drinks

4 Tell me about the food you usually eat and drink at school. Could theway you eat at school be healthier? Why or why not?

Breakfast; Hot lunch; Canteen; Are foodsbrought from home or outside theschool?

5 Tell me about the food you eat and drink when you are not at school.Could the way you eat and drink at home and other places outside ofschool be healthier? Why or why not?

At home or other places.

6 Think about the last meal you were served at school. Can you describeit for me? Did you like it? Why or why not?

Breakfast; hot lunch; taste

7 If you were in charge of making a healthy breakfast and hot lunch atschool, what would you make? Why would you choose those foods?

Fish; taste; nutritional value

8 Does anyone in your life hunt for moose, deer, rabbit, duck or bushchicken? Do they fish, or pick for berries like saskatoons, chokecherries,gooseberries, blueberries, and raspberries? If so, which of these foods doyou like to eat? Why or why not?

None

9 How often do you talk with your family about what you eat and drinkat school? If often, what do you talk about? If not often, why not?

None

10 When you are at school, do you talk to your teachers about healthyfoods? If so, what do you talk about? If not, why not?

None

11 Think about the classes you have where you learn about health.Would you like to learn more about healthy foods at school? Howwould you like to learn about healthy eating?

General nutrition information; cookingclasses; shopping trips; field trips;visits from a nutritionist; after-schoolactivities

First Nations school policy implementation

© 2018 Dietitians Association of Australia 535

many First Nations communities), and wild game meat(e.g. moose and deer). An exception was fish, which lessthan half of students chose to be served or sold. Althoughhealthy foods were popular, many students still also choseunhealthy foods like sports and energy drinks, French friedpotatoes, and soda pop to be served or sold at school.

Most students agreed (68%) or strongly agreed (19%)with the policy stating that ‘only healthy foods will beserved or sold at KEC’. Fewer than half of students indi-cated that they liked the foods served for breakfast (41%)and lunch (37%), yet most utilised the breakfast (56%) andhot lunch (53%) programs daily. In contrast, most students(72%) liked the healthy snacks offered in the school can-teen (e.g. chocolate milk, nuts, and low-fat snack foodssuch as pretzels) and just over half of students (54%) pur-chased them every day (12%) or a few days eachweek (42%).

Although most students were utilising the school foodprograms and canteen, they perceived their current eatinghabits to be only average or unhealthy. When asked toreflect on what they usually ate every day at KEC, over half(55%) of students responded that their eating habits were

average, unhealthy or very unhealthy, while the remainingstudents (45%) indicated their eating habits were veryhealthy or healthy. Fewer students (34%) considered theireating habits at home and other places outside of school tobe very healthy or healthy. Over half (57%) of studentsresponded that they had asked their parents to buy healthyfoods based on what they had learned about nutrition atschool; however, most students did not communicate regu-larly with their parents or teachers about healthy eating ortheir nutrition at school (Table 3).

Following survey administration, a total of 20 studentsparticipated in the interview, including11 (55%) girls and9 boys (45%) from all grades, except for grade 11. Inter-views ranged from 8 to 27 minutes in length, averaging14 minutes. Facilitators and barriers to SNP implementa-tion are organised at the individual, interpersonal, and com-munity, home, and sociocultural environment levels.

Individual facilitators:Health knowledge: All students provided examples of healthyand unhealthy foods. Students perceived healthy foods anddrinks as being natural, low in sugar, and containing vita-mins and nutrients. Examples of healthy foods includedwild game meat, fruit, and vegetables. In contrast, studentsperceived unhealthy foods as being processed and havinghigh sugar and fat content. Examples of unhealthy foodsincluded candy, soda pop, and pizza. Students also under-stood that food preparation and cooking methods influ-enced the healthiness of foods. For example, one studentresponded that fish is healthy ‘[depending] on how you cook itin and what it has in it’ (Student 9).

Food preferences: Students expressed considerable interest inhaving healthy choices like fruit, vegetables, and wild gameoffered to them at school. For example, when asked whichfoods they would serve for hot lunch, one studentexplained, ‘I think I would make like some salad or likesoup…Because it’s more healthier’ (Student 7). Students alsoliked traditional Cree foods, and most identified berries andmoose meat as their favourites. One student suggested thatthe hot lunch include, ‘[Moose meat]…and put some othervegetables on the side that most kids like. It seems a little morehealthier because the vegetables are healthy and the meat’(Student 15). An exception was fish, which most studentsdid not have an interest in consuming because they did notenjoy the taste. One student was also concerned about thesafety of fish consumption, explaining ‘maybe it’s just mysuperstition, I don’t know if they keep their lake clean’(Student 1).

Many students also expressed enjoying and wanting toconsume foods like French fried potatoes, soda pop, andcandy that they knew were unhealthy. For example, whenasked to describe favourite foods, one student replied,‘pizza and pop…I know they’re not healthy but it’s good, tastesgood’ (Student 4).

Interest in health education: Nearly all students were inter-ested in learning more about healthy foods at school.

Table 2 Survey responses: students’ desire to be servedand sold certain foods

Yes,n (%)

No,n (%)

Would you like the following foodand drinks sold and served atKEC?

Unprocessed, minimally processed orprocessed food

Fruit 88 (100) 0 (0)Berries 86 (98) 2 (2)100% fruit juice like apple and

orange juice84 (97) 3 (3)

Milk 83 (94) 5 (6)Cheese and yoghurt 78 (89) 10 (11)Leafy greens 78 (89) 10 (11)Vegetables 74 (84) 14 (16)Whole grain breads and pasta 69 (79) 18 (21)Whole grain, low sugar breakfast

cereal67 (76) 21 (24)

Low fat meat and chicken 61 (70) 26 (30)Baked bannock 58 (66) 30 (34)Wild game meat 56 (64) 31 (36)Fish (not breaded) 37 (42) 50 (58)Ultra-processed snack food and fried foodFried bannock 53 (61) 34 (39)Low fat snack foods 54 (61) 34 (39)Sports and energy drinks 52 (59) 36 (41)French fries 42 (48) 46 (52)Ice cream 40 (46) 47 (54)Soda pop 37 (43) 49 (57)Potato and nacho chips 33 (37) 55 (63)Cakes and cookies 33 (37) 55 (63)Candy and chocolate bars 25 (29) 62 (71)

C. Gillies et al.

536 © 2018 Dietitians Association of Australia

One student replied, ‘Yes I would, honestly. What I wouldwant to know is how to prepare healthy meals’ (Student 11).Another student specifically expressed interest in learningabout healthy foods in their non-health classes, saying‘Yes because when we go to our health class with [teacher]we mostly talk about social studies…so we never talk abouthealth’ (Student 10). Students were also interested in hav-ing after-school programming such as cooking classesoffered.

Interpersonal barriers:Lack of communication with teachers: Although the SNPencourages all staff to incorporate health education andpositive food messaging into the classroom, students indi-cated that they do not speak with their teachers abouthealthy eating. Students were unsure of or unable to artic-ulate a reason for this lack of communication, althoughone thought it may be due to time, explaining, ‘becausemaybe she is really busy…she has to teach three classes’(Student 13). In one exception, a student discussed havingeffective communication with a teacher who made a con-certed effort to discuss healthy eating and active livingwith students: ‘I talk to one teacher about living like ahealthy lifestyle…we’re trying to get a fitness group, like witha bunch of other kids in high school…he said that would belike an all-around thing, we would like workout and exerciseand then we would also talk about eating healthy…I hope hereally does that though it’s a really good idea and everybodylikes him at school’ (Student 1).

Lack of communication with parents: Students also revealedthat they do not speak with their parents about what theyeat at school, or healthy eating in general. For instance, onestudent responded ‘there’s not a particular reason why but Idon’t usually talk about that. I usually forget’ (Student 17), andanother explained, ‘I just talk about like what I did at schooland what I worked on and not really about the food’ (Student4). Other students indicated that the lack of communicationwas a result of lack of parent involvement or interest. Onestudent responded that they do not talk about what theyeat and drink at school because ‘no one asks’ (Student 18).Another student said, ‘I would talk about it, but it seems likemy parents are busy’ (Student 5), and another studentexplained that parent involvement needed to be improved,saying ‘I think with the parents it would be nice to have the

parents more involved, so they can ask their kids when they gethome “what did you eat today?”’ (Student 1).

Community, home and sociocultural environment barriers:Food served and sold in the school environment: Most studentsbelieved that the quality and variety of foods served andsold in the school could be improved. Specifically, studentsbelieved that the school’s food environment could be madehealthier by serving and selling more fruit and vegetables.One student explained ‘I think we should have more fruit andmore vegetables…I feel like it should just be mandatory to take’(Student 1). One student recommended incorporating morefruit and vegetables into the hot lunch saying, ‘they couldhave like sides with some vegetables or fruits’ (Student 14), andanother student suggested ‘apples, oranges, and bananas forlike little snacks between breaks’ (Student 17).

In addition, students explained that they were served orsold foods and drinks at school that they did not thinkwere healthy. One student perceived the pizza served at hotlunch as being unhealthy, and some students perceivedunhealthy food options being available in the school can-teen. For example, one said, ‘they sell those [processed, fla-voured rice chips], I eat those a lot. My teacher tells me that it’sbad, but I still eat it’ (Student 2). Many students also reportedbringing nutrient-poor processed foods from home withthem as snacks, including granola bars, cookies, and pud-ding. One student felt that unhealthy foods brought byother students from home should be banned, saying, ‘Somefoods should be banned like energy drinks. I see that a few timesin the school’ (Student 5).

Lack of healthy foods at home: Most students believed thattheir eating habits at home and other places outside ofschool could be improved. Specifically, students felt thatthey should be consuming more fruits and vegetables andless convenience and pre-packaged foods. For instance, onestudent responded they should eat ‘less fatty products andmore vegetables and fruits’ (Student 5). In addition, two stu-dents disclosed relying on the food programs at school tosupplement their diets as they did not have access to manyhealthy food choices at home.

Discussion

Very few studies in Canada have explored the factorsinfluencing nutrition policy implementation in First Nations

Table 3 Survey responses: communication with parents and teachers

Never, n (%)A few days eachmonth, n (%)

A few days eachweek, n (%)

Every day,n (%)

How often do you talk with your parents about whatyou eat and drink at KEC?

32 (34) 30 (32) 26 (28) 6 (6)

How often do you ask your teachers about healthyfoods to eat and drink?

44 (47) 26 (28) 20 (21) 4 (4)

How often do your teachers talk with you abouthealthy food choices?

22 (23) 27 (29) 42 (45) 3 (3)

First Nations school policy implementation

© 2018 Dietitians Association of Australia 537

schools. We surveyed and interviewed First Nations stu-dents, and their self-perceived barriers and facilitators inSNP implementation were organised within the context ofthe socio-ecological framework.3 The key individual facili-tating factors identified in this study were student supportand taste preference for healthy foods. Previous researchhas determined both student support for healthy eating andstudent taste preference to be significant factors influencingpolicy implementation.18,19 Although students admitted toenjoying unhealthy foods, they provided many examples ofhealthy foods that they desired to have at school. Echoingfinding in previous studies,20 this suggests that studentswill make healthy choices if the availability of unhealthyfoods is reduced and availability of healthy foods isincreased at school. Elders in the community were unsur-prised by this finding, explaining that most students areaware of what foods are good for them but experience bar-riers that prevent them from consuming healthy foods. Assuch, the SNP may play a significant role in ensuring stu-dents receive nutritious meals that they may otherwise notreceive due to barriers experienced outside of school.

Survey and interview data also demonstrated students’interest in consuming traditional foods of their Cree culture.This is an important facilitator of policy implementation inthis context, as the school has a focus on incorporating edu-cation on traditional foods, as well providing traditionalfoods in the school food programs. However, surveys sug-gested that students were not interested in being served orsold fish. Interviews explained that this finding was due tostudents’ distaste for fish and food safety concerns. Eldersaffirmed this concern, explaining that contamination causedby pollution and environmental degradation has affected fishpopulations as well as traditional wild game such as mooseand deer. Elders suggested that students’ disinterest in fishmay also be due to unfamiliarity and lack of exposure todiverse types of fish, as traditional food sharing practices aredeteriorating in the community and many families cannotaccess fish to prepare at home for their children. Given theseissues, students may benefit from education on environmen-tal contamination and being exposed to foods that they maynot have had an opportunity to try.

Interpersonal barriers to SNP implementation includedthe lack of communication between students and theirteachers and parents. Survey and interview results con-firmed that most students do not speak to their teachersabout healthy foods, despite staff being encouraged toincorporate nutrition education in their classrooms regard-less of the subject being taught. An evaluation completedwith staff from KEC indicated that the policy was imple-mented inconsistently by staff due to their limited knowl-edge of the policy, competing priorities, and perceived roleas nutrition policy facilitators.21 These findings indicate aneed to encourage and support staff to improve the embed-ding of policy-driven nutrition education into the classroomas part of a broader CSH approach. This may requireexplicit policy guidelines that specifically outline staffresponsibilities in addition to the provision of appropriateresources and support.22

Survey and interview findings corroborated that studentsrarely speak with their parents about what they eat anddrink at school. As students have been identified as driversof change in the home environment6,23 and parents controlmost of the food choices at home,7 parent engagementthrough communication with their children is an importantmeans of extending the SNP intentions beyond the schoolenvironment. Lack of communication is also of concernbecause parents may be unaware of the school’s food pro-grams, their child’s eating habits at school, or the SNP ingeneral. Lack of parental awareness may be a reason whychildren were sent to school with snacks that did not com-ply with policy guidelines, potentially undermining success-ful SNP implementation. Addressing this barrier mayinvolve strategies and resources to increase parent aware-ness of the policy and engagement in the SNP implementa-tion process through ongoing consultation. It is importantto consider, however, that parents may be aware of policyguidelines but are unwilling or financially unable to providehealthy food choices for their children to bring to school.

Finally, foods available to students at home and atschool presented a barrier at the community, home, andsociocultural environment level. Surveys indicated thatmost students believed that their eating habits at schoolwere average or unhealthy; interviews explained that thiswas due, in part, to their belief that they should consumemore fruits and vegetables at school. As a SNP providesthe foundation for coordinating other elements of CSH,including food programs, it is essential that guidelines arespecific, consistent, and mutually reinforcing.6 Thesefindings indicate that there may need to be clearer guide-lines surrounding foods served and sold to students inthe school environment. However, ensuring that foodmenus better adhere to SNP guidelines may be challeng-ing, especially within First Nations contexts due tounderlying issues of food access and affordability. Forinstance, access to fresh fruits and vegetables is an ongo-ing barrier for KEC due to its rural location and issueswith finding suitable and affordable fresh produce pro-viders. Furthermore, some students indicated having lim-ited healthy foods at home, which may be due to theirparents being unable to afford healthy foods like fruit, asreported by previous qualitative research at KEC.24 Whenimplementing SNPs in First Nations contexts, it is essen-tial to take into account the local environment and con-sider multiple levels of the socio-ecological framework,including considerations of healthy food availability,accessibility, and cost, local conceptualisations of health,and cultural foods.25,26

A strength to our study lies in the CBPR approach,which facilitated community control and involvement ofcommunity members, including Elders, in the evaluationof the SNP. However, as data collection took place in onlyone community, the results may not be transferable toother First Nations schools. The school in which theresearch took place is in a rural (but not remote) commu-nity with limited access to off-site fast food and conve-nience stores. The school also has a canteen where healthy

C. Gillies et al.

538 © 2018 Dietitians Association of Australia

foods are sold and a kitchen where a hired cook preparesfree breakfast and hot lunch meals. These characteristicsmay in themselves be important facilitators of SNP imple-mentation in this school. Regardless, the barriers and facili-tators identified in this research offer key factors for otherFirst Nations schools to consider when developing andimplementing SNPs.

To our knowledge, this is the first study to involve stu-dents in evaluating a SNP in a First Nations community inCanada. Future research will benefit from continuing toinvolve First Nations students to understand their percep-tions of the impact of SNPs as well as identify barriers andenablers to their success. With this knowledge, it may bepossible to optimise the successful implementation of SNPsto improve food environments and eating behaviours todecrease obesity risk for First Nations children.

Funding source

This research was funded by a grant from PolicyWise forChildren & Families (formerly, Alberta Centre for Child,Family and Community Research). CG received fundingfrom the Canadian Institutes of Health Research (Fundingreference number: DAR—152257), and the Stollery Chil-dren’s Hospital Foundation through the Women and Chil-dren’s Health Research Institute. At study inception, NDWwas a Health Scholar, Alberta Innovates Health Solutions.

Conflicts of interest

The authors have no conflicts of interest to declare.

Authorship

All authors were involved in study design. Quantitative datawere collected by CG and analysed by CG and NDW. Qual-itative data were collected and analysed byCG. Interpretations of all data were made by CG andreviewed by ARC. CG drafted the manuscript. All authorsreviewed and are in agreement with the manuscript anddeclare that the content has not been previously publishedelsewhere. The authors gratefully acknowledge AlexanderFirst Nation and Kipohtakaw Education Centre. Theauthors thank the First Nations students and Elders whoparticipated in this research, as well as school administra-tive staff that assisted in organising data collection.

References

1 Willows ND. Overweight in first nations children: prevalence,implications, and solutions. J Aboriginal Health 2005; 2: 76–86.

2 Pigford AAE, Willows ND. Promoting optimal weights inaboriginal children in Canada through ecological research. In:O’Dea JA, Eriksen M, eds. Childhood Obesity Prevention: Interna-tional Research, Controversies and Interventions. Oxford, UK:Oxford University Press, 2010; 309–20.

3 Willows ND, Hanley AJ, Delormier T. A socioecological frame-work to understand weight-related issues in aboriginal childrenin Canada. Appl Physiol Nutr Metab 2012; 37: 1–13.

4 McIsaac JL, Storey K, Veugelers PJ, Kirk SFL. Applying theoret-ical components to the implementation of health-promotingschools. Health Educ J 2015; 74: 131–43.

5 McKenna ML. Policy options to support healthy eating inschools. Can J Public Health 2010; 101: S14–7.

6 Nelson M, Breda J. School food research: building the evidencebase for policy. Public Health Nutr 2012; 16: 958–67.

7 World Health Organization. School Policy Framework: Imple-mentation of the WHO Global Strategy on Diet, Physical Activityand Health. Geneva, Switzerland: WHO, 2008.

8 Gokiert RJ, Willows ND, Georgis R, Stringer H. Wâhkôhtowin:the governance of good community–academic research rela-tionships to improve the health and well-being of children inAlexander first nation. Int Indigenous Policy J 2017; 8: 1–20.

9 Bowen S. A guide to evaluation in health research. 2012.(Available from: http://www.cihr-irsc.gc.ca/e/45336.html,accessed 17 May 2018).

10 Ivankova NV, Creswell JW, Stick SL. Using mixed-methodssequential explanatory design: from theory to practice. FieldMethods 2006; 18: 3–20.

11 Patton MQ. Qualitative Research & Evaluation Methods, 4th edn.Thousand Oaks, CA: SAGE, 2015.

12 Creswell JW. Research Design: Qualitative, Quantitative, andMixed Methods Approaches, 4th edn. Thousand Oaks, CA:SAGE, 2014.

13 Higginbottom GMA, Vallianatos H, Shankar J, Osswald B,Davey C. Understanding south Asian immigrant women’s foodchoices in the perinatal period. Int J Womens Health Wellness2016; 2: 1–7.

14 Davey C, Vallianatos H. Postpartum food traditions of Bhuta-nese refugee women: a qualitative study. J Int Migrat Integrat2018; 19: 541–53.

15 Higginbottom GMA, Vallianatos H, Shankar J, Safipour J,Davey C. Immigrant women’s food choices in pregnancy: per-spectives from women of Chinese origin in Canada. EthnHealth 2018; 23: 521–41.

16 Lincoln YS, Guba EG. Naturalistic Inquiry. Newbury Park, CA:SAGE, 1985.

17 Health Canada. Eating well with Canada’s food guide: FirstNations, Inuit and Métis. 2007. (Available from: https://www.canada.ca/en/health-canada/services/food-nutrition/reports-publications/eating-well-canada-food-guide-first-nations-inuit-metis.html, accessed 17 May 2018).

18 MacLellan D, Holland A, Taylor J, McKenna M, Hernandez K.Implementing school nutrition policy: student and parent per-spectives. Can J Diet Pract Res 2010; 71: 172–7.

19 Vine MM, Elliott SJ, Raine KD. Exploring implementation of theOntario school food and beverage policy at the secondary-schoollevel: a qualitative study. Can J Diet Pract Res 2014; 75: 118–24.

20 Downs SM, Farmer A, Quintanilha M et al. From paper topractice: barriers to adopting nutrition guidelines in schools.J Nutr Educ Behav 2012; 44: 114–22.

21 Murray K, Research Committee A, Farmer A, Maximova K,Willows ND. It’s huge in first nation culture for us, as a school,to be a role model: facilitators and barriers affecting schoolnutrition policy implementation in Alexander first nation. Int JIndigenous Health 2017; 12: 43–63.

22 Mâsse LC, Naiman D, Naylor PJ. From policy to practice:implementation of physical activity and food policies inschools. Int J Behav Nutr Phys Act 2013; 10: 71.

First Nations school policy implementation

© 2018 Dietitians Association of Australia 539

23 Storey KE, Montemurro G, Flynn J et al. Essential conditionsfor the implementation of comprehensive school health toachieve changes in school culture and improvements in healthbehaviours of students. BMC Public Health 2016; 16: 1133.

24 Genuis SK, Willows N, Nation AF, Jardine C. Through the lensof our cameras: children’s lived experience with food securityin a Canadian Indigenous community. Child Care Health Dev2014; 41: 600–10.

25 Godin K, Stapleton J, Kirkpatrick SI, Hanning RM,Leatherdale ST. Applying systematic review search methods tothe grey literature: a case study examining guidelines for school-based breakfast programs in Canada. Syst Rev 2015; 4: 138.

26 Rice K, Te Hiwi B, Zwarenstein M, Lavallee B, Barre DE,Harris SB. Best practices for the prevention and management ofdiabetes and obesity-related chronic disease among Indigenouspeoples in Canada: a review. Can J Diabetes 2016; 40: 216–25.

APPENDIX

Kipohtakaw Education Centre (KEC)School Nutrition and Physical ActivityPolicy 126

Written by Alexander First Nation Department ofEducation

Implemented March 2014Policy Statement: Kipohtakaw Education Centre will pro-

mote and provide nutritious snacks and meals consistentwith the First Nation, Inuit, and Métis (FNIM) Food Guidewhile promoting nutrition education and daily physicalactivity.Guidelines:1 All Kipohtakaw Education Centre Staff must ensure that

strategies are in place to foster the knowledge, skills andattitudes that promote healthy eating. In fulfilling thisexpectation Kipohtakaw Education Centre staff will:a establish linkages between health education and foodsavailable at the school

b promote nutrition education and positive food mes-sages provided by Alberta Health Services Website andCanadian FNMI food guide

c limit the use of food items as rewards, e.g. no candyfor cleaning desks or finishing work early.

d All school and classroom celebrations will follow theFNMI food guide and Alberta Health Services Guide-lines for healthy living (e.g. talent show, round dance,pow wow, birthday parties, Halloween, meet theteacher, parent teacher interviews, Christmas concert,Christmas parties, career fair, graduation, track andfield, prom, Easter, year-end parties, 100th day ofschool celebration and in addition to any other schoolcelebrations).

e Hot lunch menu and canteen menu to be posted inthe monthly newsletter.

2 Kipohtakaw Education Centre will promote healthy, rea-sonably priced food choices when food is sold or other-wise offered. In fulfilling this expectation, KipohtakawEducation Centre Staff will plan to:a access expertise in the community through partner-ships, programs, referrals, etc.,

b offer foods that are from the FNMI Food Guidec All fundraisers must follow the FNMI Food Guide andAlberta Health Services guidelines for healthy living.

3 Kipohtakaw Education Centre school community willexamine their nutrition practices and provide opportuni-ties, support and encouragement for staff and students toeat healthy foods. In fulfilling this expectation staff maydo things such as:a create their own health and wellness team thatincludes staff, parents and students

b choose healthy fundraising optionsc create an environment where healthy foods are avail-able, affordable and promoted as the best choice

d review options with food suppliers to maximise thenutritional value of the items

e define the frequency of special celebrations in yearlycalendars and ensure that healthy food items are avail-able on those days

f will promote positive food messaging on lunch andsnack items provided by parents (Kipohtakaw Educa-tion Centre staff are not responsible for unhealthy foodchoices brought from home)

C. Gillies et al.

540 © 2018 Dietitians Association of Australia

LETTER TO THE EDITOR

WHAT DISEASE DID GOETHE WITNESSDURING HIS JOURNEY THROUGH THEITALIAN ALPS? WAS IT PELLAGRA ORANOTHER DISEASE OF MALNUTRITION?

To the Editor:In the chapter entitled ‘From Brenner to Verona’ of his liter-ary work ‘Italian Journey’ (Italianische Reise), Goetherecounted to have seen numerous unhealthy people as hedescended from Brenner in September 1786. He particu-larly noticed deterioration in women’s complexions, theirskin was a pale brown, and also children were extremelythin. They were cachectic, only men seemed to be in aslightly better condition. The cause of this visible physicalsuffering was most linked to diet, there was a continuoususe of maize and buckwheat with which the yellow polentaor black polenta was cooked. In these pages, he highlightedthat on the German side of the Alps, people ate polentacooked with butter while in the Italian Tyrol region, peopleate polenta but cooked only in water, sometimes addingcheese and more than often they never ate meat. This typeof diet, according to Goethe, caused obstructions in thedigestive tract, especially in children and women.1 PerhapsGoethe had observed the pellagra disease as some aca-demics claimed?2,3 We are not sure that it was pellagra.Firstly, if Goethe had seen pellagra in the people of Tyrol,he would certainly have observed (and reported in thesepages) the most obvious signs of the disease, that is, des-quamation of the skin of the hands and neck. Secondly,even if the disease had severely affected the population ofNorthern Italy between the eighteenth and nineteenth cen-turies, especially in Veneto and Friuli, there was no trace ofthis in the mountain areas. Moreover, in the same page,Goethe described that the diet of the local people also con-sisted of fruit and green beans boiled in water and seasoned

with garlic and oil. We know that the onset of pellagra isrelated to an almost exclusive diet based on maize. There-fore, it is possible to assume that Goethe had noticed thatmany inhabitants of those valleys suffered from a disease ofmalnutrition.

Funding source

The authors received no funding for this work.

Conflict of interest

The authors have no conflict of interest to declare.

Authorship

All authors contributed equally to this work. This manu-script was read and approved by all of the authors.

Marta Licata ,1 andSilvia Iorio,2

1Centre of Research in Osteoarchaeology andPaleopathology, Department of Biotechnology and Life

Sciences, University of Insubria, Varese and 2Departmentof Molecular Medicine, Unit of History of Medicine,

Sapienza University of Rome, Rome, Italy

References

1 Goethe JW. Italian Journey. Translated by Auden WH and MayerE. London: Penguin, 1970.

2 Leeberg J. Goethe and the mal de rosa. Z Gastroenterol 1983;21: 241–3.

3 Loyd JH. Goethe’s reference to pellagra. JAMA 1912; 59:959–60.

Correspondence: M. Licata, Centre of Research in Osteoarchaeologyand Paleopathology, Department of Biotechnology and Life Sciences,University of Insubria, O. Rossi, 9, 21100 Varese, Italy.Email: marta.licata@uninsubria.it

Accepted June 2018

© 2018 Dietitians Association of Australia 541

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 541 DOI: 10.1111/1747-0080.12458

ERRATUM

Please note the following error:

Nutr Diet 2018; 75: 341–4 Jayne Thirsk, RD, PhD, FDC Senior Director, Dietitians of Canada, Canada

EDITORIAL Knowledge translation

The error was published on the main text, p. 342 and p. 343 in the Reference section: Matthews instead of Mathews.

It should have read:

Matthews et al.21 explore these phenomena in their examination of dietitians’ opinions regarding the prevalence, diagnosisand effective monitoring of refeeding syndrome in 11 countries.

Matthews KL, Palmer MA, Capra SM. Dietitians’ opinions regarding refeeding syndrome, clinical guidelines and extendedscope of practice. Nutr Diet 2018; 75: 397–405.

© 2018 Dietitians Association of Australia542

AA AA

DDD

ieti

tian

s As

soci

atio

nofof

Aus

tral

ia

Nutrition & Dietetics 2018; 75: 542 DOI: 10.1111/1747-0080.12500