the new zealand medical journal...apr 28, 2000 · (e.g. the ‘under sixes’) and waiting lists...

New Zealand Medical Journal 13128 April 2000

Journal of the New Zealand Medical Association

THE NEW ZEALANDTHE NEW ZEALANDTHE NEW ZEALANDTHE NEW ZEALANDTHE NEW ZEALANDMEDICAL JOURNALMEDICAL JOURNALMEDICAL JOURNALMEDICAL JOURNALMEDICAL JOURNAL

ISSN 0028 8446

Twice monthly except December & January

NEWSLETTER(pages 1-6)

Vol 113 No 1108 28 April 2000

INFORMATION FOR AUTHORSFirst page following cover

EDITORIALS133 A turbulent decade: lessons from the ‘health reforms’ Andrew Hornblow,

Pauline Barnett

ORIGINAL ARTICLES135 Internet use amongst New Zealand general practitioners Jason Eberhart-Phillips,

Katherine Hall, G Peter Herbison, Sarah Jenkins, Joanna Lambert, Richard Ng,Martha Nicholson, Lorna Rankin

137 A meta-analysis of 25 hydroxyvitamin D in older people with fracture of the proximal

femur Mark Weatherall140 Glycaemic index of New Zealand foods Tracy Perry, Jim Mann, Kirsty Mehalski,

Cynthia Gayya, Jan Wilson, Caryn Thompson143 Willingness to pay for new chemotherapy for advanced ovarian cancer Jenny Morris,

David Perez146 Trauma form documentation in major trauma James Hamill, Rhondda Paice, Ian Civil

VIEWPOINT148 Right idea, wrong model. Why general practitioners are not succeeding with

preventive care Stuart Foote150 Research in general practice or general practice research Murray Tilyard, Susan Dovey

132 New Zealand Medical Journal 28 April 2000

Editor: Gary NichollsDeputy Editors: Philip Bagshaw, Evan Begg, Peter Moller, Les Toop, Christine WinterbournBiostatistician: Chris Frampton Ethicist: Grant GillettEmeritus: Pat Alley, John Allison, Jim Clayton, Roy Holmes, John NeutzeEditorial Board: George Abbott, Bruce Arroll, Sue Bagshaw, Gil Barbezat, Richard Beasley, Ross Blair,Antony Braithwaite, Stephen Chambers, Garth Cooper, Brett Delahunt, Matt Doogue, Pat Farry,Bruce Foggo, Peter Gluckman, Jane Harding, Andrew Hornblow, Geoffrey Horne, Rod Jackson,Peter Joyce, Martin Kennedy, Graham Le Gros, Diana Lennon, Tony Macknight, Tim Maling, Jim Mann,Colin Mantell, Lynette Murdoch, Bryan Parry, Neil Pearce, David Perez, Anthony Reeve, Ian Reid,Mark Richards, André van Rij, Justin Roake, Peter Roberts, Bridget Robinson, Prudence Scott,Norman Sharpe, David Skegg, Bruce Smaill, Rob Smith, Ian St George, Andy Tie, Ian Town,Colin Tukuitonga, Harvey White

Tables should be on separate sheets with self-explanatorycaptions. Footnote symbols must be used in a set sequence (* † ‡ §

¶ ** †† # etc).

Figures must be glossy prints or high quality computerprintouts. Since these are likely to be reduced in size whenprinted, use large type and approximately twice column size forthe figure.

Conflict of Interest: Contributors to the Journal should let theEditor know whether there is any financial or other conflict ofinterest which may have biased the work. All sources of fundingmust be explicitly stated in the paper and this information will bepublished.

The Journal does not hold itself responsible for statements madeby any contributors. Statements or opinions expressed in theJournal reflect the views of the author(s) and do not reflect officialpolicy of the New Zealand Medical Association unless so stated.

Information for authorsGuidelines for authors are in accordance with the UniformRequirements for Manuscripts submitted to Biomedical Journals.Full details are printed in NZ Med J 1997; 110: 9-17, Med Educ1999; 33: 66-78 and are on the NZ Medical Association website –www.nzma.org.nz. Authors should be aware of the broad generalreadership of the Journal. Brevity and clear expression areessential. The maximum length for a paper is no more than threeprinted pages in the Journal, one page containing around 1100words. The number of references should not exceed 30. Forpapers accepted for publication which exceed three printed pages(around 3,000 words) there will be a page charge of $450 plusGST for each printed page. Letters should not exceed 400 wordsand ten references. Requirements for letters, obituaries andeditorials are on the website. All material submitted to the Journalis assumed to be sent to it exclusively unless otherwise stated. Eachauthor must give a signed personal statement of agreement topublish the paper or letter.The paper: Papers are to be written in English and typewritten indouble spacing on white A4 paper with a 25 mm margin at eachside. Send three copies of the paper. Organise the paper as follows:

Title page – the title should be brief without abbreviations.Authors’ names, with only one first name and no degrees shouldbe accompanied by position and workplace at the time of thestudy. Corresponding author details with phone, fax and emailshould be given, and the text word count noted.

Abstract page – this must not exceed 200 words and shoulddescribe the core of the paper’s message, including essentialnumerical data. Use four headings: Aims, Methods, Results,Conclusions.

Body of the paper – there should be a brief introduction (noheading) followed by sections for Methods, Results, Discussion,Acknowledgements and Correspondence.

References – in the text use superscript numbers for eachreference. Titles of journals are abbreviated according to thestyle used by Index Medicus for articles in journals the format is:Braatvedt GD. Outcome of managing impotence in clinicalpractice. NZ Med J 1999; 112: 272-4. For book chapters theformat is: Marks P. Hypertension. In: Baker J, editor.Cardiovascular disease. 3rd ed. Oxford: Oxford University Press;1998. p567-95. Note all authors where there are four or less; forfive or more authors note only the first three followed by ‘et al’.Personal communications and unpublished data should also becited as such in the text.

AddressesEditorial: All editorial correspondence is sent to Professor Nicholls,c/o Department of Medicine, Christchurch Hospital, PO Box4345 Christchurch, New Zealand. Telephone (03) 364 1116;Facsimile (03) 364 1115; email [email protected]: All correspondence is sent to the AdvertisingManager, Print Advertising, PO Box 27194, Upper WillisStreet, Wellington. Telephone (04) 801-6187; Facsimile (04)801-6261; email [email protected] or 83-91 CaptainSprings Road, PO Box 13 128 Onehunga, Auckland.Telephone (09) 634-7227; Facsimile (09) 634-1929; [email protected]: All correspondence about circulation,subscriptions, change of address and missing numbers is sent toChief Executive Officer, New Zealand Medical Association,PO Box 156, Wellington. Telephone (04) 472-4741; Facsimile(04) 471-0838. email [email protected]: The Journal is published by Southern Colour Print,PO Box 920, Dunedin. Telephone (03) 455-0554;Facsimile (03) 455-0303.Subscriptions: New Zealand – standard mail NZ$255.15,fastpost NZ$272.25 (GST incl); overseas surface mailNZ$280.00, overseas airmail – South Pacific/AustraliaNZ$340.00; America/Asia/India/Europe NZ$420.00; Africa/Middle East NZ$490.00. All subscription enquiries to NZMedical Association, as for Circulation above.

Twice monthly except December & January Copyright New Zealand Medical Association ISSN 0028 8446

Established 1887 - Journal of the New Zealand Medical Association

THE NEW ZEALANDTHE NEW ZEALANDTHE NEW ZEALANDTHE NEW ZEALANDTHE NEW ZEALANDMEDICAL JOURNALMEDICAL JOURNALMEDICAL JOURNALMEDICAL JOURNALMEDICAL JOURNAL


28 April 2000Volume 113

No 1108THE NEW ZEALANDMEDICAL JOURNAL

A turbulent decade: lessons from the ‘health reforms’

Andrew Hornblow, Dean, Christchurch School of Medicine; Pauline Barnett, Senior Lecturer, Department of PublicHealth and General Practice, Christchurch School of Medicine, Christchurch.

For the New Zealand health system, the turbulent decade ofthe 1990s has taught painful lessons which must not be lostas a new government leads us beyond the so-called ‘healthreforms’.

The 1991 Green and White Paper1 heralded a major shiftin both the organisation of health care and the underlyingideology which governed health policy. Without healthsector consultation or public mandate, the era of imposedmarket-oriented reforms had begun. The stated goals of thereforms could hardly be disputed. These were to improveaccess to an effective and affordable health system,encourage efficiency and innovation, reduce waiting times,widen choice of services, enhance the work environment forhealth professionals, improve illness/injury prevention andhealth promotion, and increase sensitivity of the healthsystem to changing health needs. However, accompanyingthe advocacy for these laudable goals, were repeated andlargely unsubstantiated accusations of gross inefficiency,poor targeting and provider capture. The moral high groundof universally accepted goals, combined with the zeal of thereformers served as the basis for refocusing health servicesfrom a public service to a commercial ethos.

The structural changes proposed in 1991 were radical: thepurchaser-provider split and competitive contracting managed byfour Regional Health Authorities (RHAs); Crown HealthEnterprises (CHEs) structured with corporate boards; anindependent Public Health Commission; a Core ServicesCommittee to ration; health care plans to allow alternativeprivate purchasing arrangements; and social insurance funding.By the time of implementation in 1993, social insurance fundinghad been explicitly abandoned and health care plans faced anuncertain future.

By mid-1996, three years into the reforms and with anelection pending, it was clear that the reformers’expectations were ill-founded. Higher charges limited accessto general practitioners (GPs). Waiting lists had increased byup to 50%. The confident prediction of 20-30% savings hadnot been realised with CHE’s experiencing persistentdeficits, around $179m in 1995/96 alone. RHA costs hadincreased by 40%. Hospital user charges had beenintroduced and withdrawn after public outrage. Contractnegotiations had proved costly, contentious, and slow.Thirteen of 23 CHE CEO’s and six Board Chairs hadresigned. The Public Health Commission had beenabolished, following conflict with the Ministry and pressurefrom the tobacco and liquor industries and other vestedinterests. Only 35% of public health targets were expectedto be achieved, according to Ministry of Health predictions.The Core Services Committee, having failed to define the‘core’ of health services, was transformed into the NationalHealth Committee, with a different task. Alternative health

care plans had quietly been dropped. Community concernwas growing, encouraged by opposition to the reforms fromdoctors, nurses and other health professionals. This,expressed through groups such as the Public HealthCoalition, the professional colleges, the Public HealthAssociation, the Nurses’ Organisation, and others, combinedwith mounting public concern, developed intounprecedented and organised anti-reform advocacy, againstwhich the government’s costly public relations campaignproved ineffectual.

The Coalition Government of December 1996 was relianton New Zealand First, which gained balance-of-power statuson the strength of its campaign against both the governmentand the health reforms. The structural arrangements of 1993were significantly modified in ways now familiar to us all.Competitive contracting was replaced by principles of co-operation and collaboration. The four RHAs werecentralised as the Health Funding Authority. The 23 CHEswere subsequently rehabilitated as Hospital and HealthServices. Resource allocation became more pragmatic, forexample, in primary care there was a retreat from targeting(e.g. the ‘under sixes’) and waiting lists and other target areas(e.g. mental health) received new funding.

The National government of the early 1990s, and theCoalition government which followed, did not formally evaluatethe reforms, although a number of useful monitoring reportswere published2-4 and other reviews have appeared.5-7 However,failure to assess systematically the impact of the changes was alost opportunity. For a country seen in the past as a ‘sociallaboratory’, embarking again on bold and controversial changesto major social structures, the impact of the reforms would proveto be of considerable international interest. Despite the lack offormal evaluation, with a change of government it is now timelyto assess the legacy of the previous decade.

Retention of the purchaser/provider split has maintainedthe essentially adversarial contract negotiating environment.In hospital management the appointment of CHE directorsand senior staff from outside the health sector was intendedto strengthen management and commercial expertise. Evenwith such expertise, the government’s own officials advisedin 1996 that “the pace of performance improvement sincethe reforms seems to have weakened”.3 There was a climateof mistrust between boards, management and professionalstaff. This was much worse in some areas than others, mostnotably in the spectacularly dysfunctional situation inChristchurch where the enquiry undertaken by the Healthand Disability Commissioner8 blamed patient deaths onsystem changes and inadequacies. In the last three yearsmanagement-clinician relationships have undoubtedlyimproved country-wide, but much has yet to be done torebuild trust and establish genuine partnership at all levels.

EDITORIAL


In primary care the establishment of independentpractitioner associations and other primary careorganisations has brought new energy to primary care andpromises much for innovation and expansion of services.4However, with only 50% of general practice services nowgovernment funded there is significant unmet need,particularly among Maori and low income groups, and inrural communities. In a five country comparison,9 only theUS is perceived to have worse access to primary medicalcare, a sad judgement on the health reforms’ ability todeliver affordable and accessible care. In public health, withoverall immunisation coverage estimated to be only about70%, comparable to that of many third world countries, theimplications for increased prevalence of preventableinfectious diseases are alarming.

The health reformers’ expectations that a market model wouldencourage flexibility and innovation, and widen choice, havebeen only partly realised. In some high priority areas, such ascommunity mental health and Maori health, there are now morefunds, more appropriate services and a wider range of providers.In some rural areas community trusts have been formed whichprovide viable alternative services. But waiting lists haveextended, not reduced, and the introduction of a surgical points-based booking system, while intended to be rational and fair, hasproved problematic. Private health insurance, still only 6% ofoverall health expenditure, cannot solve the difficult problem ofthe demand for surgical services outstripping supply. Theproportion of New Zealanders now holding health insurance hasdropped from 50% to 30%, with the clear and inevitablesocioeconomic bias, 63% of above average earners being insured,and only 29% of those with below average income. The publiccomponent of total health expenditure has declined through the1990s to 77%,10 again raising questions about equity and who inthe community is bearing these costs.

The lessons of the health reforms of the 1990s should be setin a cultural context, both of New Zealand culture and theculture of the health sector. New Zealanders take pride in aculture which emphasises equity and social justice. Theideologically driven imposition of market-oriented reformswas an exercise, not of bold leadership but of politicalarrogance and a rejection of established community values.Within the health sector, the reforms polarised clinical andcommercial cultures. The colonisation of one culture byanother brings challenge to and changes in power structures,decision-making, resource allocation, social structures,concepts and language, and dominant values and beliefs(Table 1). Where restructuring fails it is often because issuesof ‘organisational culture’ are ignored.

Table 1. Clinical and commercial cultures in conflict mode.

Clinical Commercial

Desired outcome cure, care profit, growthValues compassion, excellence competition, efficiencyLanguage health sciences business, managementDecision-making collegial hierarchicalKnowledge base health status, pathology, market forces, commercial

treatment, patient risk risk managementmanagement

The lessons of the health reforms have been painful ones, andmust not be lost. A fundamental lesson is that a market approachto the delivery of health care has major limitations. The ultimategoal of a public health system is the provision of care asequitably, effectively and efficiently as possible, not the profitablesale of commodities. A second lesson is that optimal functioningof a sector as complex as health requires a close partnershipbetween those with technical expertise (doctors, nurses and otherhealth professionals), and managers, and the confidence of thewider community. A third lesson is that attainment of desiredhealth status outcomes requires strategic planning, wideconsultation, and co-ordinated action within and beyond thehealth sector. The de-emphasis during the reforms on publichealth and preventive measures must be reversed, with urgentstrengthening of population-based strategies targeted towardpreventable illness and injury, and towards those most in need. Afourth lesson is that our public health system is seen as a nationaltaonga, and that erosion of the sense of ownership andconfidence becomes a shared loss for the community as a whole.

The new Labour-led coalition has promised to restore thepublic’s faith in the public health system. As the hold of marketideology over the public sector diminishes, the reforms of the1990s become part of our history. The challenge remains thatof maximising health gains and effective health care deliverythrough best use of limited resources.1. Upton S. Your health and the public health. Wellington: Ministry of Health; 1991.2. Ministry of Health. Purchasing for your health, 1994/5; 1995/6; 1996/7. Ministry of Health;

Wellington.3. Crown Company Monitoring and Advisory Unit. Crown Health Enterprises: briefing to the

incoming minister. Wellington: CCMAU: 1996.4. National Health Committee. Review of primary health care in New Zealand. Wellington:

National Health Committee; 1999.5. Hornblow A. New Zealand’s health reforms: a clash of cultures. BMJ 1997; 314: 1892-4.6. Ashton T. The health reforms: to market and back? In: Boston J, Dalzeil P, St John S. Redesigning

the welfare state in New Zealand. Auckland: Oxford University Press; 1999. p134-53.7. Barnett P, Barnett JR. Reform and change in health service provision. In: Davis P, Dew K (eds).

Health and society in Aotearoa New Zealand. Auckland: Oxford University Press; 1999. p219-34.8. Stent R. Canterbury Health Ltd: A Report by the Health and Disability Commissioner;

April 1988.9. Donelon K, Blendon RJ, Schoen C et al. The cost of health system change: public discontent

in five nations. Health Aff 1999; 18: 206-16.10. Ministry of Health. Health expenditure trends in New Zealand, 1988-98. Wellington;

Ministry of Health, 1999.

ABC of NutritionEditor. S Truswell. BMJ Publishing. Contains 121 pages. ISBN 0 7279 1233 X.Price not stated.

What to do with the GP who has been taught little or no nutrition at medical schooland is faced with treating patients with a wide variety of nutrition related conditions?Write an ABC of Nutrition. This challenge was given to Dr Stewart Truswell in themid 1980s by Dr Stephen Lock, Editor of the BMJ and led to publication of the firstedition of the ABC of Nutrition in 1986. Today’s GP has little more formal nutritiontraining yet is confronted by a bewildering array of information about nutrition fadsand controversies that makes it difficult to decipher what is sound and informed. Thethird and revised edition of the ABC of Nutrition by Dr Truswell is a timely update.

ABC books often have to sacrifice in the interest of space too much of theimportant for the most important, but this is not the case here. The major nutritionrelated diseases and deficiencies are covered: heart disease, cancer (a little too briefly),hypertension, diabetes, obesity, and nutritional anaemias, but so are osteoporosis,dental caries, and gallstones to name a few. The role of nutrition in normal growthand development (pregnancy, infancy, children and adolescents) and in maintaininghealth in young and old adults is dealt with expertly. There is an excellent chapter thatdescribes food sensitivity and short chapters on nutritional support and therapeuticdiets. The chapter on food poisoning–contributed by Patrick G Wall–is highlyinformative and includes several concise reference tables listing the incubation period,duration of symptoms, and clinical features of the most common food pathogens.

The manual has relevant tables that provide a balance of complementary andadditional detail. The use of text boxes is exploited to introduce definitions,important diagnostic criteria, guidelines for nutrition treatment, and to discusstopical issues (eg. sports nutrition). Each chapter ends with a list of ten to fifteenreferences to the best supporting evidence and review material.

The reader is provided with several UK addresses where additional nutritiondocuments (eg. Coeliac Society’s list of gluten free manufactured products) can beobtained. Unfortunately, the documents are relevant for those living in the UK.Most of the documents have New Zealand and Australian equivalents and futureeditions of ABC of Nutrition could provide in an appendix the relevant addresseswhere the documents can be obtained.

Dr Truswell has called upon his vast experience as a practitioner and researcherto distil and present the essence of nutrition knowledge and practice. I often foundmyself admiring the way complicated debates that have filled volumes werecondensed into less than one page yet retained enough of their essential elements toallow the reader to appreciate the debate and understand the ‘bottom line’.

Health professionals including nurses, doctors, dentists, and midwives would behard pressed to find a better ABC book to guide and inform their nutritionpractices. Dietitians and nutritionists with more formal nutrition training willappreciate Dr Truswell’s informed, balanced, and concise presentation of theessentials of nutrition. I highly recommend this book.

Dr M Skeaff,Dunedin.

Book Review


The astonishing growth of the Internet and the World WideWeb has recently made vast quantities of health-relatedinformation readily available to health professionals andpatients. With the click of a mouse it is now possible foranyone with minimal computer skills to download scientificpapers, case reports, personal accounts and idle chatter onvirtually any medical topic. More than 10 000 Web sites arealready devoted to medical issues and many more are createdevery week.1 People searching for health information maygenerate as much as 40% of Internet traffic according to arecent estimate.2

Sources of online medical information range fromelectronic editions of peer-reviewed medical journals, tosearchable scientific databases such as MEDLINE, tomember-only digests of current medical intelligence, toInternet list servers that forward messages among subscriberssharing similar interests, to various electronic discussiongroups devoted to every imaginable health-related subject.The promises of the new technology for the practice ofmedicine are many, but they are diminished by concernsabout the accuracy of the information available3,4 and otherserious drawbacks. For example, many clinicians worry thatelectronic communications about specific patients might notbe kept confidential.5 Others are concerned about medicolegalliability arising from advice provided over the Internet.6

New Zealanders have embraced the Internet in recentyears. A census of Internet service providers in early 1999found that more than 315 000 accounts had already beenestablished here.7 Considering that multiple users aretypically connected through a single account, it is likely thatthe Internet is now an everyday part of life for tens ofthousands of New Zealanders. We wanted to assess howextensively New Zealand general practitioners are using theInternet, and to examine how they believe the newtechnology is affecting their practice of medicine.

MethodsDuring the first week of December 1998 we sent a postal questionnaire toall 259 general practitioners in the Otago/Southland region who wereknown to the Department of General Practice at the Dunedin School ofMedicine. The confidential questionnaire, which was pilot tested,included 21 pre-coded items on the doctor’s use of the Internet, thedoctor’s impressions of patient use of online resources, and demographic

data. Respondents were also asked two open-ended questions; one abouthow the Internet had changed the ways they practice medicine, and oneabout how the Internet had affected their relationships with patients.

GPs were asked to return the questionnaire in a reply paid envelope.Those who had not responded ten days after the survey was posted weresent a second questionnaire and a reminder letter. 20% of those who hadnot responded to either mailing after four weeks were randomly selectedto undergo a standardised telephone interview in a separate study toascertain the effects of bias due to non-response. The telephoneinterviews, conducted among the GPs who could be reached in up to fourattempts, included selected questions from the postal survey.

Analysis was limited to questionnaires returned within four weeks byGPs currently in practice. Those who reported using the Internet at leastmonthly for any reason were defined as users. Quantitative data wereanalysed by calculating simple frequencies. For analysis of the two open-ended questions, five of the authors, working as a group, assigned theresponses to categories that captured the essence of the comments. Twoother authors then repeated the qualitative analysis, to validate the initialallocation of responses into the selected categories. The study had theapproval of the Otago Ethics Committee.

ResultsA total of 184 questionnaires (71%) were returned, ofwhich 168 (65%) were sent in by GPs currently in medicalpractice. Of those 168 respondents, 121 (72%) said theyhad used the Internet at least once, and 114 (68%) saidthey use it at least on a monthly basis. A total of 55% ofrespondents said they had access to the Internet at home,and 36% said they had access at work. 30 respondents(18%) had access both at home and at work.

Internet users did not differ significantly by gender, with73% of male respondents and 60% of female respondentsreporting use. Similarly, Internet use did not varysignificantly between those who trained in New Zealand andthose who trained overseas, or between those who primarilyserve urban or rural populations. In contrast, Internet usewas significantly less common among older GPs. Only five(39%) of the respondents 60 years of age or older said theyuse the Internet, compared to 71% of those under 60 yearsof age (chi-squared = 6.04, df = 1, p = 0.014).

Of those who use the Internet, 92 (81%) said they use itfor work-related tasks, including updating medicalknowledge, finding information for patients andcommunicating with colleagues. But only 30% of those GPssaid they use the Internet at least weekly to update their

Internet use amongst New Zealand general practitioners

Jason Eberhart-Phillips, Senior Lecturer, Department of Preventive and Social Medicine; Katherine Hall, ClinicalLecturer, Department of General Practice; G Peter Herbison, Senior Research Fellow, Department of Preventiveand Social Medicine; Sarah Jenkins, Trainee Intern; Joanna Lambert, Trainee Intern; Richard Ng, Trainee Intern;Martha Nicholson, Trainee Intern; Lorna Rankin, Trainee Intern, Dunedin School of Medicine, Dunedin.

Aims. To assess how extensively New Zealand doctors areusing medical information on the Internet, and to examinehow the new technology is affecting their practice ofmedicine.Methods. All general practitioners (GPs) known to beworking in Otago and Southland were asked to complete apostal questionnaire regarding their use of the Internet andtheir impressions of patient use of online medical resources.Results. Of 259 questionnaires mailed out, 168 (65%)were returned by GPs currently in practice. Of those, 114(68%) said they used the Internet at least monthly. A total

of 71% of GPs had patients who indicated they hadsought medical information from the Internet. Nearlyhalf of respondents expressed concerns that the Internetcould have unwelcome effects on the doctor-patientrelationship.Conclusions. Internet use among New Zealand doctorsand patients is widespread, and is likely to have significantimpact on medical practice now and in the future. Whilethe potential benefits of the new technology are numerous,the Internet may become a source of conflict betweendoctors and patients.

Abstract

NZ Med J 2000; 113: 135-7

ORIGINAL ARTICLES

135 28 April 2000


knowledge or to obtain information for patients. About 38%of users said they communicate with colleagues via theInternet at least weekly, but just nine users (8%) said theyhad ever communicated with patients via the Internet.

Adult medicine is the area of medical science about whichthe respondents said they most commonly approach theInternet for information. Three-quarters of those who use theInternet said they most often pick the Web sites they visit by‘surfing’ from one site to the next. About 45% said they had‘bookmarked’ at least one frequently used medical sitethrough their Internet browser. Only ten of the Internet users(9%) said they were involved in a Web site of their own.

Among all the GPs who responded, 71% said they hadpatients who indicated that they had sought medicalinformation from the Internet. Of those, 50% saw at leasttwo patients per month who mentioned the Internet, and6% saw five or more such patients. According to thedoctors, patients had sought online information on most ofthe same areas of medical science as clinicians, except thatdoctors are more likely to look up Web sites pertaining todermatology. Only 17% of the GPs thought that medicalinformation on the Internet is too difficult for patients tounderstand, while 26% feared that the informationavailable is not accurate enough.

In response to the open-ended questions, 42 respondents(25%) said the Internet had already changed the ways theypractice medicine, nearly two-thirds of whom viewed thechanges positively. A total of 108 respondents (64%)indicated that the Internet had affected the doctor-patientrelationship, or would do so in the future. This comparedto 5% who believed the Internet would have no effect, and30% who offered no opinion.

In all, just 26 respondents (15%) said the Internet wouldimprove their relationships with patients. By contrast, 82(49%) expressed some anxiety or equivocation about theInternet’s potential effects on the doctor-patientrelationship. Many of their concerns were related toincorrect or inadequate online medical information (citedby 45% of the 82 respondents with concerns), resourceimplications arising from patients requesting particulartherapies (21%), and the undermining of trust betweendoctors and patients (27%).

In the survey of non-responders, nine of the fourteenrandomly chosen GPs could be contacted and agreed toparticipate in the telephone interview. These individuals didnot differ significantly from the respondents to the postalsurvey in terms of their demographic characteristics or theiruse of the Internet.

DiscussionIn the practice of medicine, the potential benefits oftoday’s cyber revolution for improving knowledge andsharing information are numerous. At a time when thescientific evidence for medical decision-making isexpanding rapidly, a computer connected to the Internetmay become as essential for clinicians as the stethoscope,the sphygmomanometer or the telephone.8 There is noquestion that the Internet today offers healthprofessionals an easily accessible and abundant resourceregarding current knowledge and accepted practice atrelatively little cost.9 Unlike traditional medical datasources, this medium is inherently interactive, enablinglong-distance consultations with specialists and others atthe forefront of medical scholarship. As more doctors takeadvantage of the medical resources on the Internet, fewerpatients may settle for care from more traditional office-based practices that shun electronic medicine.

Our survey found that Internet use among Otago/

Southland GPs and their patients is now widespread.There is no reason to believe that medical use of theInternet would differ in other regions of New Zealand.Ready access to current information, especially regardingrare conditions, easier construction of patient handouts,and improved communication with colleagues, are amongthe Internet’s benefits for doctors that were confirmed inthis survey. But despite all the potential gains, doctors’attitudes towards this new technology remain equivocal.Many of their concerns relate to the accuracy of theinformation at hand, resource implications, and worriesabout the corrosive effects of the Internet on the doctor-patient relationship.

Perhaps the biggest barrier to uptake of the Internet inmedical practice is the scarcity of quality control over theinformation available. Anyone, including those with notraining in medicine, can set up a Web site and offer medicalopinions without submitting to scientific editing and peerreview. It is just as easy to broadcast medical misinformationon the Internet as it is to disseminate scientifically validclinical views. For example, a five-month review ofconsecutive messages sent to an online bulletin board for thediscussion of painful hand and arm conditions found that 89%of messages that contained medical information wereauthored by persons with no medical training. Approximatelyone-third of the information provided was judged to beunconventional.3 In another example, a study that examinedonly online articles written by qualified medical authors foundthat just 20% of articles for lay audiences, on the managementof acute diarrhoea in children, complied with AmericanAcademy of Pediatrics guidelines.4

With so much inaccurate or misleading informationavailable online, patients connected to the Internet couldend up well informed yet misinformed. As one GP in oursurvey put it, “Patients often perceive Internet informationhas credibility just because it comes out of a computer.” GPswith Internet-savvy patients may have to redefine their role,becoming ‘brokers and interpreters of information,’ requiredto explain and place into context the abundant informationtheir patients are now bringing in. This new role couldconsume considerable time resources, as a few GPs alreadymentioned in our survey. Other resource issues they citedinvolve dealing with requests for diagnostic tests ortreatments that are either not available in New Zealand orare difficult to obtain anywhere.

It has been suggested that “Medicine along with the otherprofessions, is distinctive in that the specialisation is in sheerinformation.”10 The Internet challenges such definitions ofprofessionalism, in that the ownership of medical knowledgeis not exclusively, or even predominantly, in the domain ofthe medical profession. As one GP wrote in this survey,“[The Internet] is no different from other sources ofinformation accessed by patients in the past - eg peopleresearching in the library etc. - however [it] is moreaccessible”.

Such accessibility to medical knowledge raises thelikelihood of having patients who are more informedabout their conditions than the doctor. Clearly somedoctors are undaunted about such a situation,appreciating a more informed patient. As one stated inour survey, “I always thought information sharing a veryimportant part of medical practice that has not been givenenough time.” Another respondent asserted that a highlyinformed patient is only a threat if the doctor-patientrelationship is based on the biomedical model and is not“patient-centered”.

However, a substantial fraction of GPs in this surveyexpressed varying degrees of anxiety and concern that the

28 April 2000 136


Internet might damage the therapeutic alliance betweendoctor and patient. In the more strongly wordedstatements, there was an anxiety that the Internet woulddiminish trust between doctor and patient. These doctorsdid not argue that ‘blind faith’ was a required element inthis relationship, but some could cite difficulties that hadarisen already because of Internet use by patients. Forexample, one recalled encounters marked by “a reluctanceor even refusal to accept even well tried/provenmedication for conditions because they [the patients] wantto carry out further research into the matter concerned”.

For the restoration and preservation of trust in thedoctor-patient relationship, the surgeon and philosopherMiles Little has said that the public require “confidence inthe knowledge, training and skills of medicalpractitioners... [and assurance that] the medical professionsought the truth in every possible way”.11 This requirementplaces great demands on practitioners if ‘confidence inknowledge’ means that they remain always better informedthan their patients. Given that this is impossible, GPs mustconsider other ways of positively utilising patients’Internet-acquired knowledge. This might include theinterpretative role, mentioned previously. Such aninterpretative role allows the GP to demonstrate sincerelya seeking of the truth in aiding the patient to understandthe context of the information obtained online from this‘large and cheap medical dictionary’.

Perhaps the explosive growth of medical information viathe Internet has shattered once and for all the fantasy ofthe all-knowing doctor. Doctors preparing to practice inthe 21st century may find that patients are becoming lessinterested in GPs who impart ‘book knowledge’, andmore interested in clinicians who will work together with

them to make sense of the extensive information thepatients already have. Doctors may find their bestapproach is open disclosure, accepting the limitations intheir knowledge and getting on with the task of findingthe best solutions for their patients’ problems. As one GPin this survey noted, “I always treat these situations withcomplete transparency admitting I don’t know something,if I don’t, and then this ends up working more as apartnership/team”. Uncertainty has never been easy forthe medical profession.12 Paradoxically, the rise of theInternet, with the increased accessibility of informationfor doctors and patients alike, makes dealing withuncertainty due to the limits of one’s knowledge an evenmore important consultative skill.

Correspondence. Dr Jason Eberhart-Phillips, Department of Preventiveand Social Medicine, PO Box 913, Dunedin.

1. Ferguson T. Digital doctoring - opportunities and challenges in electronic patient-physiciancommunication. JAMA 1998; 280: 1361-2.

2. Huntley AC. The need to know. Patients, e-mail and the Internet. Arch Dermatol 1999; 135;198-9.

3. Culver JD, Geff F, Frumkin H. Medical information on the Internet: a study of an electronicbulletin board. J Gen Intern Med 1997; 12: 466-70.

4. McClung HJ, Murray RD, Heitlinger LA. The Internet as a source for current patientinformation. Pediatrics 1998; 101: e2.

5. Rind DM, Kohane IS, Szolovits P et al. Maintaining the confidentiality of medical recordsshared over the Internet and the World Wide Web. Ann Intern Med 1997; 127: 138-41.

6. De Ville KA. Internet list servers and pediatrics: newly emerging legal and clinical practiceissues II. Pediatrics 1996; 98: 453-4.

7. ISP numbers grow. New Zealand Net Guide March 1999, p. 10.8. Mandl KD, Kohane IS, Brandt AM. Electronic patient-physician communication: problems

and promise. Ann Intern Med 1998; 129: 495-500.9. Klemenz B, McSherry D. Obtaining medical information from the Internet. J R Coll

Physicians Lond 1997; 31: 410-3.10. Arrow KJ, Calabresi G, Pellegrino ED. Government decision making and the preciousness of

life. In: Tancredi LR, (editor). Ethics health care: papers of the conference on health care andchanging values, 27-29 November 1973. Washington: National Academy of Sciences, 1974.

11. Little JM. Trust me - I’m a doctor. Paper presented at Medical Forum, Dunedin School ofMedicine, March 1997.

12. Katz J. Why doctors don’t disclose uncertainty. Hastings Cent Report 1984; 14: 35-44.

A meta-analysis of 25 hydroxyvitamin D in older people with fracture of the

proximal femur

Mark Weatherall, Senior Lecturer, Department of Medicine, Wellington School of Medicine Wellington.

Aims. To perform a meta-analysis on the published studiesof serum levels of 25 hydroxyvitamin D in older peoplewith fractures of the proximal femur compared to controlgroups.Methods. A ‘Medline’ literature search using key words‘vitamin D’ and ‘hip fractures’ for the years January 1966 toJune 1999, seeking only papers published in English andavailable from New Zealand medical libraries, wasperformed. Bibliographies of identified papers were alsosearched. Studies which compared 25 hydroxyvitamin Dlevels in people with a fracture of the proximal femur to anolder control group were eligible for inclusion. 30 studieswere identified and 28 papers could be found in New

AbstractZealand. The method of weighted Z statistics was used inthe meta-analysis.Results. The pooled reduction in serum 25hydroxyvitamin D for the fracture group compared to thecontrols was 0.66 of a standard deviation with a 95%confidence interval of 0.74 to 0.59.Conclusions. Although there may be publication bias inthis meta-analysis and there was some evidence ofheterogeneity in the studies, there is very good evidencethat older people with fracture of the proximal femur havereduced levels of vitamin D compared to controls. Olderpeople with fracture of the proximal femur should betreated with vitamin D.

NZ Med J 2000; 113: 137-40

Fracture of the proximal femur, so called hip fracture, is amajor public health problem in New Zealand. It is the mostcommon serious fracture in older people. A recent report tothe National Health Committee1 stated that over 3000people, mostly over the age of 65 years, suffered a fracture of

the proximal femur in 1994. They had an average length ofstay in hospital of about three weeks. New Zealand datasuggests that 20% of those previously resident in their ownhomes do not return home and as many as 50% of thoseresident in an institution do not return to their previous

137 28 April 2000


level of care. Of those who do return to their own home alarge proportion never regain the degree of independence intheir instrumental activities of daily living they had prior tothe fracture.2 It is now increasingly recognised that vitaminD status may play a crucial role in osteoporosis and perhapsin tendency to fall.3 The purpose of this paper is to subjectthe published literature, in English, concerning 25hydroxyvitamin D levels in people with a hip fracture, tometa-analysis.

MethodsA ‘Medline’ literature search based on the data base from January 1966 toJune 1999 was carried out using the words ‘vitamin D’ and ‘hip fractures’as both subjects and text words. Bibliographies of identified papers werealso reviewed. Papers were eligible for inclusion if they measured serum25 hydroxyvitamin D levels in a group of people shortly after fracture ofthe proximal femur and compared these levels to a group of elderlycontrols and if they could be obtained from New Zealand medicallibraries. If young control levels were included in the paper only theelderly control data were used. Data extracted from the papers were ofthe number of subjects, and mean serum 25 hydroxyvitamin D levels andtheir standard deviation, in each of the groups. In some papers thisinvolved calculating the standard deviations from the standard error ofthe mean. In some papers the 25 hydroxyvitamin D levels were presentedin different time periods of the year and these results were pooled to giveone total figure for the analysis. For purposes of interpretation all levelsare given here as nmol/L although for the method of meta-analysis usedthe scale of the measurement is not important. The method of meta-analysis used was the method of ‘Weighted Z statistics’.4 The statisticalprogramme Minitab 10.2 (Minitab Inc. 1994) was used.

Results98 papers were identified by the ‘Medline’ search. Of these, eightpapers were eligible for inclusion. A further 22 studies wereidentified from the bibliographies of these studies.5-34 Two paperswere not available in New Zealand.33,34 Both studies measured 25hydroxyvitamin D in people with hip fracture and in controlsubjects, and found lower 25 hydroxyvitamin D levels in thefracture group. However, the actual means and standarddeviations were not described in the abstracts available on‘Medline’. For the study of Lund et al,5 means and standarddeviations were estimated from examination of a graph of thedata where lines were drawn to illustrate these parameters but noactual numerical data were given in the paper. In another study,6the fracture group data were pooled from the two time periodsgiven. For study by Wooton et al, the 25 hydroxyvitamin D datawere presented as log transformed, but it was not stated if thiswas natural or base 10 logarithm.9 Base 10 logarithm wasassumed. For the study of Knorring et al, data were presented ina table as mean and standard error, however when the datapoints from an accompanying table were analysed it was clear thedata for 25 hydroxyvitamin D had been written as mean andstandard deviation. Data were also pooled across time periods forthis study.10 For the study of Hoikka et al, data were pooled formales and females.11 In the study of Lund et al, the mean andstandard deviations for the groups were calculated fromestimated raw values presented as dot plots.12 In three studies,fracture data were pooled for trochanteric and cervical fracturetypes.13,16,23 Control groups were pooled in one study,15 andcontrol data were pooled across the two control groups in twoother studies.22,32 Mean and standard deviations for the groupswere calculated from estimated raw values presented as dot plotsin the study of Pun et al.26 The timing of blood samples in thefracture group was pre-operative in thirteen studies,5,10,15-17,24-30

post operative in six studies 9,18,20,23,31,32 and not stated in theothers.

Overall there were 1384 people in the fracture group and1572 people in the control group for the papers for whichdata were available.

Table 1 shows study details and standardised differences in25 hydroxyvitamin D levels. The standardised difference is

expressed as the number of standard deviations betweenfracture and control groups together with the 95%confidence interval for the difference. In the table ‘Not’means the proportion of subjects who were female was notstated and ‘Match’ means the authors stated age and/orgender matched controls were used but did not specify themean age or proportion who were female. In the first twostudies in the table only the age ranges were given.5,6 In theother papers the mean age in the groups is cited. Figure 1shows a plot of the data in Table 1 with the linesrepresenting the 95% confidence intervals for each of thestudies together with the pooled confidence interval. Theestimate of the standardised difference between the twogroups was -0.66 standard deviations with a 95% confidenceinterval of -0.74 to -0.59.

A formal test of heterogeneity gave a chi-squared statisticof 123.14 on 27 degrees of freedom, p<0.0001, suggestingthat the studies were not homogenous.

There were twelve studies where subjects with a fractureof the proximal femur were compared to an inpatient controlgroup.9,10,15,18,20-24,28,31,3 The estimate of the standardiseddifferences based on these studies was -0.45 standarddeviations with a 95% confidence interval of -0.57 to -0.34.

DiscussionThere are two major weaknesses with this meta-analysis. First,it may be subject to publication bias. The search was confinedto studies published in English identified on ‘Medline’, andthat were available in New Zealand medical libraries and withsearches of the bibliographies of identified articles. No effortwas made to identify study information in other formats suchas theses or abstracts of conferences. The two studiesidentified on bibliographic review, but which could not beobtained from New Zealand medical libraries, both reported asignificantly lower 25 hydroxyvitamin D level in the fracturegroup, consistent with the findings of the meta-analysis. Thesecond, and more serious, weakness was the strong evidenceof heterogeneity in the studies on formal testing. This is likelyto be due to the many differences in the studies regardingselection of an appropriate control group. In some studies thecontrol groups were merely stated as elderly, in others theywere people attending outpatient clinics or people hospitalisedfor other reasons. If the control groups might have had less ormore sunlight exposure than the subjects with hip fracturethen this could partly explain the heterogeneity of the resultssince vitamin D status varies with sunlight exposure. If thismeta-analysis is thought of as supporting the contention thatlow serum 25 hydroxyvitamin D has a causative role in hipfracture then this heterogeneity suggests that low 25hydroxyvitamin D levels in the hip fracture group may be afunction of which control groups were selected for thecomparison.

Based on recently published New Zealand data35 thestandard deviation for serum 25 hydroxyvitamin Dmeasurements in 92 adults aged 55-64 years was 29 nmol/L.From the result of the present pooled analysis it can becalculated that people with fracture of the proximal femurmight be expected to have serum 25 hydroxyvitamin D levels18 nmol/L less on average than other elderly adults.

Other evidence to support the view that a low level ofvitamin D may be causative in hip fracture is that a lowlevel of vitamin D is associated with increased levels ofparathyroid hormone as the body maintains serum calciumby mobilising the mineral content of bone. Severeprolonged vitamin D deficiency gives rise to the bonemineralization disorder, osteomalacia. Osteomalacia isuncommon in New Zealand.36 A less severe vitamin Ddeficiency, however, may contribute to osteoporosis.3 It is

28 April 2000 138


now well established that house bound individualsincluding people residing in rest homes have vitamin Dlevels below the accepted normal range.3,37-40 In a recentNew Zealand study institutional care residents made up42% of those presenting with fracture of the proximalfemur.41 People resident in institutional care seem to havean increased risk of fracture of the proximal femur with acrude risk ten times that of the general population,42

although after adjusting for confounding factors there is anodds ratio of 2.2 for fracture of the hip for those resident inresidential care.43 A recent large controlled study of resthome residents has shown that calcium and vitamin Dsupplementation reduces the rate of osteoporotic fractureover a two year period44 although another large study ofcommunity dwelling older people showed no benefit ofvitamin D supplementation for fracture prevention.45 Thesame group that performed the latter study had previouslydemonstrated an increase in bone mineral density in older

women given vitamin D compared to a placebo group.46 Apossible explanation of the association between low 25hydroxyvitamin D levels with fracture of the proximalfemur may be that the fracture group had a number ofpeople normally resident in institutional care, who are nowfairly well recognised to be at risk for vitamin D deficiency,and that in this group vitamin D deficiency, is associatedwith hip fracture. There is evidence that medical inpatientshave a high prevalence of low serum 25 hydroxyvitamin D47

and this may be an explanation for the reduced levels of 25hydroxyvitainin D in this meta-analysis. However, a sub-group analysis of the papers in which the control subjectswere inpatients also showed that 25 hydroxyvitamin Dlevels were lower in hip fracture groups compared tocontrols.

The implications for practice are that people whopresent with fracture of the proximal femur seem to be atrisk of vitamin D deficiency and strong considerationshould be given to either measuring 25 hydroxyvitainin Din this group or treating them with vitamin D, such as iscontained in most multivitamin preparations. Higher dosecholecalciferol is also available. It also seems reasonable,based on increasing evidence, that people in residentialcare or who are house bound should have vitamin Dsupplements. A relatively cheap method is use of vitaminD containing multivitamin preparations. As the ultimateaction of the active metabolite of vitamin D is to enhancecalcium absorption it also seems reasonable, in those atrisk of vitamin D deficiency, to also supplement theircalcium intake. Calcitriol, the active metabolite of vitaminD, is available as a pharmaceutical preparation. It isexpensive and its use does not replenish the body stores ofvitamin D so this preparation is probably not suitable forprophylaxis of vitamin D deficiency.

Correspondence. Dr Mark Weatherall, Department of Medicine,Wellington School of Medicine, Private Bag 7343, Wellington. Fax (04) 3895427; Email: [email protected]

Table 1. Study details and standardised differences in serum 25 hydroxyvitamin D levels.

Study Control Fracture Location Standardised difference(95% Cl)

N Age (years) % Female N Age % Female

5 41 60-95 Not 67 60-95 Not Denmark 0.26 (-0.13,0.65)6 7 64-85 43 39 60+ 87 England -0.86 (-1.69,-0.04)7 22 76.2 Not 22 74.7 Not Israel 0.21 (-0.38, 0.80)8 76 79.4 100 98 80.2 100 England -0.74 (-1.04, -0.43)9 29 74.5 76 42 75.9 86 England -0.55 (-1.03,-0.07)10 41 78 78 58 77 81 Finland -0.66 (-1.07, -0.26)11 58 Match Match 58 74.6 72 Finland -1.26 (-1.66,-0.86)12 34 Match Match 21 72.4 80 Denmark -0.55 (- 1.11, 0)13 95 Match 100 75 78.7 00 England -0.48 (-0.79, -0.18)14 50 71.5 100 67 77.9 100 Australia -1.11 (-1.5, -0.72)15 46 65.6 39 82 76.3 73 Finland -0.96 (-1.34, -0.58)16 18 Match Match 36 72.5 61 Israel -0.88 (-1.46,-0.29)17 25 73.8 100 40 77.1 100 Finland -1.93 (-2.54, -1.33)18 10 69.4 100 10 67.5 100 England -0.11 (-0.99, 0.77)19 74 75.6 73 125 75.9 67 Holland -1.22 (-1.53, -0.91)20 23 82 100 23 80 100 England -0.04 (-0.62, 0.54)21 49 72 100 30 78 100 Finland -0.06 (-0.52, 0.39)22 39 69.7 100 29 74.2 100 Denmark -0.24 (-0.73, 0.23)23 27 51 Not 15 77 Not Israel -0.35 (-0.99, 0.27)24 20 73.3 100 41 77.4 100 England -0.68 (-1.23, -0.14)25 24 76.8 71 30 80.1 81 Finland -0.49 (-1.04, 0.05)26 28 71.2 100 69 78.1 100 Hong Kong -0.53 (-0.98, -0.08)27 61 78.6 Not 61 78.9 Not Hong Kong -1.22 (-1.62, -0.84)28 68 82.5 69 57 83.9 91 France -0.55 (-0.91, -0.19)29 117 77.7 100 117 79.2 100 Belgium -1.01 (-1.28, -0.74)30 40 73.1 0 40 72.4 0 Belgium -1.03 (-1.5, -0.57)31 180 79.9 76 179 80.2 76 Switzerland -0.31 (-0.52, -0.1)32 82 77.9 0 41 79.6 0 Australia -0.65 (-1.03, -0.27)

Pooled -0.66 (-0.74,-0.59)

See text for explanation of Not and Match.

Figure 1. Plot of 95% confidence intervals for standardised differencein serum 25 hydroxyvitamin D levels from 28 studies.

139 28 April 2000


The concept of glycaemic index (GI) of carbohydrate-containing foods was first introduced in the 1970’s byJenkins and colleagues.1 The index represents the glycaemicresponse to a food containing a standard amount ofcarbohydrate expressed as a percentage of the glycaemicresponse to a similar glucose load. The suggestion that lowGI foods may be particularly appropriate for people withdiabetes was initially treated with cynicism, but there is nowconsiderable evidence that long term glycaemic control asassessed by measurement of glycated haemoglobin can beappreciably reduced by replacing high GI carbohydrate-containing foods with those with a lower glycaemic index.2-5

Total and low density lipoprotein cholesterol are alsoreduced suggesting potential additional benefit in terms of

cardiovascular risk reduction.6-9 Furthermore, there is nowadditional evidence from prospective epidemiological studiesthat diets with a high glycaemic load increase the risk ofdeveloping diabetes and conversely that diets rich in low GIcarbohydrate-containing foods are protective.10,11 A recentexpert consultation of the Food and AgricultureOrganisation (FAO) and the World Health Organisation(WHO) has endorsed the potential benefit which accruesfrom the use of low GI foods in diabetes and suggested thatGI might be more widely used to guide populations withregard to the most appropriate choices of carbohydrate-containing foods.12 Many tabulations containing GI valueshave been published but the composition of some foods isinfluenced by geographic location, soil and climatic

Glycaemic index of New Zealand foods

Tracy Perry, Lecturer and Researcher; Jim Mann, Head of Department; Kirsty Mehalski, Research Student;Cynthia Gayya, Research Student; Jan Wilson, Research Student, Department of Human Nutrition, University ofOtago; Caryn Thompson, Statistician, Department of Mathematics and Statistics, University of Otago, Dunedin.

Aim. To determine the glycaemic index values to a rangeof foods that are unique to New Zealand, and those thatare grown and/or manufactured locally.Methods. We determined the glycaemic index of 28carbohydrate foods in both healthy subjects and those withtype 2 diabetes. Venous blood samples were collected overtwo to three hours, and the incremental area under the bloodglucose curve was used to calculate glycaemic index values.

Results and Conclusions. This study has identified theglycaemic index values for a range of New Zealand foodswhich will be clinically useful in the nutritionalmanagement of individuals with diabetes mellitus. Peoplewith diabetes are recommended to choose foods with a lowglycaemic index which is associated with optimal bloodglucose control and lipid levels.

Abstract

NZ Med J 2000; 113: 140-2

1. Sainsbury R, Richards R. Prevention of osteoporosis in older populations. Wellington:National Health Committee; 1997 June. ISBN 0478104650.

2. Weatherall M. One year follow up of patients with fracture of the proximal femur. NZ Med J1994; 107: 308-9.

3. Gloth FM, Tobin JD. Vitamin D deficiency in older people. J Am Geriatr Soc 1995; 43: 822-8.4. D’Agostino RB, Weintraub M. Meta-analysis: a method for synthesizing research. Clin

Pharm Ther 1995; 58: 605-16.5. Lund B, Sorenson OH, Christensen AB. 25 hydroxycholecalciferol and fractures of the

proximal femur. Lancet 1975; ii: 300-2.6. Brown IRF, Bakowska A, Millard PH. Vitamin D status of patients with femoral neck

fractures. Age Ageing 1976; 5: 127-31.7. Weisman Y, Salama R, Harell A, Edelstein S. Serum 24,25 dihydroxyvitamin D and 25

hydroxyvitamin D concentrations in femoral neck fracture. BMJ 1978; ii: 1196-7.8. Baker MR, McDonnell H, Peacock M, Nordin BEC. Plasma 25 hydroxyvitamin D

concentrations in patients with fractures of the femoral neck. BMJ 1979; i: 589.9. Wooton R, Brereton PJ, Clark BMB et al. Fractured neck of femur in the elderly: an attempt

to identify patients at risk. Clin Sci (Colch) 1979; 57: 93-101.10. Knorring JV, Slatis P, Weber TH, Helenius T. Serum levels of 25 hydroxyvitainin D, 24,25

dihydroxyvitamin D and parathyroid hormone in patients with femoral neck fracture insouthern Finland. Clin Endocrinol 1982; 17: 189-94.

11. Hoikka V, Alhava EM, Savolainen K, Parviainen M. Osteomalacia in fractures of theproximal femur. Acta Orthop Scand 1982; 53: 255-60.

12. Lund B, Sorneson OH, Lund B et al. Vitamin D metabolism and osteomalacia in patientswith fracture of the proximal femur. Acta Orthop Scand 1982; 53: 251-4.

13. Lawton JO, Baker MR, Dickson RA. Femoral neck fractures-two populations. Lancet1983; ii: 70-2.

14. Morris HA, Morrison GW, Burr M et al. Vitamin D levels and femoral neck fractures inelderly South Australian women. Med J Aust 1984; 140: 519-21.

15. Harju E, Puranen J, Lahti R. High incidence of low serum vitamin D concentration inpatients with hip fracture. Arch Orthop Trauma Surg 1985; 103: 408-16.

16. Meller Y, Kestenbaum RS, Shany S et al. Parathormone, Calcitonin, and Vitamin D metabolitesduring normal fracture healing in geriatric patients. Clin Orthop 1985; 199: 273-9.

17. Punnonen R, Salmi J, Tuimala R et al. Vitamin D deficiency in women with femoral neckfracture. Maturitas 1986; 8: 291-5.

18. Stevenson JC, Allen PR, Abeyasekera G, Hill PA. Osteoporosis in hip fracture: changes incalcium regulating hormones. Eur J Clin Invest 1986; 16: 357-60.

19. Lips P, van Ginkel FC, Jongen MJM et al. Determinants of vitamin D status in patients withhip fracture and in elderly control subjects. Am J Clin Nutr 1987; 46: 1005-10.

20. Hordon LD, Peacock M. Vitamin D metabolism in women with femoral neck fracture. BoneMiner 1987; 2: 413-26.

21. Harma M, Parvianinen M, Koskinen T et al. Bone density, histomorphometry andbiochemistry in patients with fractures of the hip or spine. Ann Clin Res 1987; 19:378-82.

22. Hartwell D, Riis BJ, Christiansen C et al. Bone metabolism and bone status in osteoporoticpatients. Acta Med Scand 1987; 222: 453-8.

23. Alkalay D, Shany S, Dekel S. Serum and bone vitamin D metabolites in elective patients andpatients after fracture. J Bone Joint Surg (Br) 1989; 71-B: 85-7.

24. Cooper C, Mclaren M, Wood PJ et al. Indices of calcium metabolism in women with hipfractures. Bone Miner 1989; 5: 193-200.

25. Lamberg-Allardt C, von Knorring J, Slatis P, Holmstrom T. Vitamin D status andconcentrations of serum vitamin D metabolites and osteocalcin in elderly patients withfemoral fracture: a follow up study. Eur J Clin Nutr 1989; 43: 355-61.

26. Pun KK, Wong FH, Lau P et al. Vitamin D status among patients with fractured neck offemur in Hong Kong. Bone 1990; 11: 365-8.

27. Macdonald D, Lau E, Chan ELP et al. Serum intact parathyroid hormone levels in elderlychinese females with hip fracture. Calcif Tissue Int 1992; 51: 412-4.

28. Benhamou CL, Tourliere D, Gauvain JB et al. Calciotropic hormones in elderly people withand without hip fracture. Osteoporos Int 1995; 5: 103-7.

29. Boonen S, Broos P, Verbeke G et al. Calciotropic hormones and markers of boneremodelling in age related (type II) femoral neck osteoporosis: Alterations consistent withsecondary hyperparathyroidism induces bone resorption. J Gerontol 1997; 52A; M286-93.

30. Boonen S, Vanderschuren D, Cheng XG et al. Age related (type II) femoral neckosteoporosis in men: biochemical evidence for both hypovitaminosis D and androgendeficiency induced bone resorption. J Bone Miner Res 1997; 12: 2119-26.

31. Thiebaud D, Burchardt P, Constanza M et al. Importance of albumin 25(OH)-vitamin D andIGFBP-3 as risk factors in elderly women and men with hip fracture. Osteoporos Int 1997; 7:457-62.

32. Diamond T, Smerdely P, Kormas N et al. Hip fracture in elderly men: the importance ofsubclinical vitamin D deficiency and hypogonadism. Med J Aust 1998; 169: 138-41.

33. Lips S, Bouillon R, Jongen MJ et al. The effect of trauma on serum concentrations of vitaminD metabolites in patients with hip fracture. Bone 1985; 6: 63-7.

34. Lau EM, Woo J, Swaminathan R et al. Plasma 25-hydroxyvitamin D concentration inpatients with hip fracture. Gerontology 1989; 35: 198-204.

35. Scragg R, Holdaway I, Singh V et al. Serum hydroxyvitamin D3 is related to physical activityand ethnicity but not obesity in a multicultural workforce. Aust NZ J Med 1995; 25: 218-23.

36. Borrie MJ, Campbell AJ, Caradoc-Davies TH. Osteomalacia in the elderly: investigation anddiagnosis. NZ Med J 1985; 98: 989-91.

37. Fraser DR. Vitamin D. Lancet 1995; 345: 104-7.38. Gloth FM, Gundberg CM, Hollis BW et al. Vitamin D deficiency in homebound elderly

persons. JAMA 1995; 274: 1683-6.39. Liu BA, Gordon M, Labranche JM et al. Seasonal prevalence of vitamin D deficiency in

institutionalised older adults. J Am Geriatr Soc 1997; 45: 598-603.40. Stein MS. Low serum 25 hydroxyvitamin D: Prevalence in a large Australian nursing home.

In: Australian Society for Geriatric Medicine Annual Scientific Meeting Handbook; 1995June 19-20; Queensland Australia: 1995 p 62.

41. Weatherall M. Medical profile of patients with fracture of the proximal femur. The risks ofhip fracture in older people from private homes and institutions. Age Ageing 1996: 25: 381-5.

42. Butler M, Norton R, Lee-Joe T et al. The risks of hip fracture in older people from privatehomes and institutions. Age Ageing 1996; 25: 381-5.

43. Norton R, Campbell AJ, Reid IR et al. Residential status and risk of hip fracture. Age Ageing1999; 28: 135-9.

44. Chapuy MC, Arlot ME, Duboeuf F et al. Vitamin D3 and calcium to prevent hip fractures inelderly women. N Engl J Med 1992; 327: 1637-42.

45. Lips P, Graafinans WS, Ooms ME et al. Vitamin D supplementation and fracture incidencein elderly persons. Ann Intern Med 1996; 124: 400-6.

46. Ooms ME, Roos JC, Bezemer PD et al. Prevention of bone loss by Vitamin Dsupplementation in elderly women: A randomized double blind trial. J Clin EndocrinolMetab 1995; 80: 1052-8.

47. Thomas MK, Lloyd Jones DM, Thadhani RI et al. Hypovitaminosis D in medical inpatients.N Engl J Med 1998; 338: 777-83.

28 April 2000 140


conditions and processing procedures.13 Furthermore, somefoods are available only in certain countries. The FAO/WHO expert consultation strongly recommended themeasurement of GI of local foods in individual countries.We report here, for the first time, the GI values ofcarbohydrate-containing foods unique to New Zealand andthe Pacific Islands as well as some commonly eaten foods inwhich GI may differ from internationally published data.We have also compared the results obtained in healthyindividuals with those in people with diabetes, who are likelyto benefit most from this information.

MethodsStudy Design. The Southern Regional Health Authority EthicsCommittee (Otago) approved the study and informed consent wasobtained from all participants. Groups of eight to fifteen healthy subjectsdrawn from a pool of 47 subjects (29 females and eighteen males), andfourteen people with type 2 diabetes (five females, nine males), seventreated by diet alone and seven on oral hypoglycaemic medication,participated in the study: all subjects were Caucasian. The former groupwas younger (average (SD) age of 31 (11.1) years) and leaner (body massindex (BMI) 25 (4.6) kg/m2) than those with diabetes (age 52 (8.7) years,BMI 30 (4.8) kg/m2). The mean (SD) fasting blood glucose was 5.1 (0.48)in the healthy volunteers and 8.4 (2.41) mmol/L in those with diabetes.All foods were tested in healthy volunteers and five foods were also testedin those with type 2 diabetes.

The volunteers attended the clinic for testing at weekly intervals in themorning after a ten to twelve hour overnight fast. A cannula was insertedinto an antecubital vein of the forearm. Three millilitre fasting bloodsamples were taken as soon as the cannula was secure and then again fiveminutes later before ingestion of test food. Further samples were taken atfifteen, 30, 45, 60, 90 and 120 minutes after consumption of the test mealor glucose solution. Additional samples were taken at 150 and 180minutes from those with type 2 diabetes. The reference solution (glucose)and test foods were ingested within ten minutes.

Blood was collected in 5mL vacutainers containing EDTA(K3)(Vacutainer, Becton Dickinson, Rutherford, New Jersey), held on crushedice for two hours then centrifuged for ten minutes at 2500G at 4ºC.Plasma was stored at -18ºC until analysis. The plasma glucoseconcentration was analysed in duplicate using the enzymatic UV test withhexokinase (Glucose HK Unikit III, Roche). All glucose analyses wereperformed on the Cobas II Fara autoanalyser (Roche Diagnostica, Basle,Switzerland). The interassay coefficient of variation for the study groupwas 5.8% and for those with type 2 diabetes mellitus 6.0%.Foods tested. All foods were tested in random order. An exception wasyams due to seasonal availability. The breads tested were Bürgen Dark/Swiss Rye, Bürgen Mixed Grain, Country Fare Kibbled Konini, Nature’sFresh® Fibre White™ and Quality Bakers™ Molenberg® bread. Allbreads were stored frozen and thawed before service. Various starchyvegetables were also tested, including green banana, kumara, potato(‘Nardine’ variety), sweetcorn (‘Honey and Pearl’ variety), taro and yams.The vegetables were either boiled or microwaved in water. Rice and pastafoods tested were Mother Earth Egg Fettuccine, Uncle Bens® WhiteLong Grain rice and Maggi Two Minute Noodles which were cooked inboiling water according to package instructions. The following breakfastcereals were served with water only: Sanitarium Natural Muesli;Sanitarium Muesli Lite; Sanitarium Berry Fruity Bix; KelloggsCornflakes; Hubbards Porridge; Hubbards Fruitful Lite; and HubbardsThank Goodness. The Mother Earth Apricot Bar; Griffins Golden Fruitbiscuit; and Malt meal wafer were served from the package. The Novartismedical nutritional products are drinks which were served directly fromthe tetrapacks: Resource Diabetic (swiss chocolate flavour); Resourcethickened beverage (orange juice, honey consistency); Resource thickenedbeverage (orange juice, nectar consistency); and Resource fruit beverage(peach flavour).

Most meals were prepared in a manner commonly used in the NewZealand diet, but without any traditional accompaniments (for examplethe green banana was served without coconut cream). Each test mealcontained 50 g of available carbohydrate and portion size was based onThe New Zealand Food Composition Tables14 or nutritional informationprovided with the product. On at least three occasions, each subject wastested with the reference solution of 50 g of dextrose glucose powderdissolved in 400 mL of water. All meals were served with additional waterto bring the water content of each test meal to 410 mL (equivalent to theglucose test) to minimise the possible differences in the rate of gastricemptying, thereby eliminating meal volume as a variable.Statistical Analyses. The incremental area under the blood glucoseresponse curve (AUC), ignoring the area below the fasting level wascalculated geometrically using the method described by Wolever andJenkins.15 For each subject, the AUC for each test food was expressed as a

percentage of the mean AUC for the glucose tests taken by the samesubject, and the resulting values meaned to give the actual food GI. Thewide range in individual responses to different foods has been clearlydocumented,16,17 and the main purpose of GI determinations is to producerankings which in turn provide a guide to the most appropriate choices ofcarbohydrate-containing foods. Thus published tables do not usuallyinclude values for the statistical significance of differences between foods.However, comparisons were made for those foods tested in healthyvolunteers and volunteers with diabetes using the Mann-Whitney U Test.

ResultsValues for the GI of various foods tested in healthyindividuals are given in Table 1. The comparison betweenthe results for healthy individuals and for those with diabetesare presented in Table 2: there were no statisticallysignificant differences between the two sets of values.

Table 1. Gycaemic index values of foods tested in healthy individuals.

Food: Number Mean SEM

BreadsBürgen Mixed Grain 10 45 11.8Bürgen Rye 9 55 12.1Fibre White 14 77 9.7Konini 14 68 12.4Molenberg 15 75 9.9

Starchy VegetablesCorn 9 37 11.5Green Banana 8 38 9.6Kumara 9 77 11.5Potato 8 70 17.4Taro 9 56 12.1Yams 14 35 5.0

Rice and PastaFettucine 14 47 5.9White Rice 14 56 7.12 minute noodles 15 48 7.6

Breakfast CerealsHubbards Porridge 10 57.5 8.8Hubbards Fruitful Lite 9 60.5 20.0Hubbards Thank Goodness 11 65.3 17.9Kelloggs Cornflakes 11 72 16.2Sanitarium Berry Fruity Bix 10 113.2 9.9Sanitarium Muesli Lite 10 53.7 12.2Sanitarium Muesli Natural 10 56.7 9.3

Sweet biscuits and fruit barGolden Fruit Biscuit 10 77 25.0Maltmeal wafer 10 49.5 9.8Apricot Bar 15 50 7.9Novartis Nutritional ProductsResource Diabetic 11 16.0 3.7Thickened Beverage -orange juice honey 11 47.2 8.8Thickened Beverage -orange juice nector 11 53.6 7.3Resource Fruit Beverage 11 40.2 8.4

Table 2. Glycaemic index values for selected foods in healthy individualsand those with type 2 diabetes.

Food Number Mean SEM

Type 2 diabetesBürgen Mixed Grain 13 69 6.1Bürgen Rye 14 74 5.5Molenburg 14 84 8.1Kumara 14 78 5.5Yams 13 25 3.7

Healthy IndividualsBürgen Mixed Grain 10 45 11.7Bürgen Rye 9 55 12.1Molenburg 15 75 10.1Kumara 9 77 11.5Yams 14 35 5.0

There was no significant difference (of mean GI values) between those with type2 diabetes and healthy individuals using the Mann Whitney U test.

141 28 April 2000


DiscussionThe GI of a food was initially believed to be determined bythe dietary fibre content, especially the soluble componentof dietary fibre.17 While carbohydrate-containing foodswhich are rich in soluble fibre do indeed have a lowglycaemic index, it is now clear that there are many otherdeterminants of the glycaemic response following ingestionof carbohydrate.17 A low glycaemic index is observed when afood is rich in fat as well as carbohydrate or when fructosecontributes significantly to the total carbohydrate content.Thus fruits high in glucose (eg grapes) will have a higher GIthan those high in fructose (eg dates, apple). The need toavoid high intakes of fat, especially saturated fat, underlinethe fact that GI cannot be used as the sole indicator of themost appropriate food choices for people with diabetes.Food form is regarded as a particularly importantdeterminant of GI. Intact cell walls tend to reduce GI.17

Thus whole grains are associated with lower GI than whenthe cereal is extensively processed and the cellular structuredestroyed. Whole fruits will have a lower GI than pureedforms despite a similar content of total sugars.17 While it issometimes possible, therefore, to utilize internationallypublished tables for certain foods (eg breakfast cerealsmanufactured by international companies with standardizedproduction procedures) it is imperative to gather data forlocally grown, produced and consumed food and products.Measurements of glycaemic index must be undertaken understrictly controlled conditions and in adequate numbers ofsubjects in order to avoid producing misleading results. Thebroadly comparable data for those with type 2 diabetes andhealthy individuals are reassuring, confirmatory of theresults of other groups17,18 and suggest that findings relevantto the treatment of diabetes may be extrapolated fromstudies of non-diabetic individuals. These data represent thefirst attempt to gather such GI information regarding foodsconsumed in New Zealand.

As has been clearly demonstrated in other studies, NewZealand breads appear to have a wide range of glycaemicindices which cannot readily be predicted from informationprovided on the product label. However, there are at least twopossible explanations for the satisfactory (relatively low) GI ofBürgen mixed grain bread. Product labelling suggests a longerfermentation time leading to increased production of organicacids which in turn may lower glycaemic response.19

Furthermore the higher content of kibbled or whole grainswith their intact cell walls may also lower GI.19-21

In regard to starchy vegetables, values for sweet corn, yamand potato are comparable with those published in theinternational literature confirming both sweet corn and yamswith their low GI are particularly good food choices.4 Of thethree uniquely Pacific food choices tested green banana wasfound to have a very low GI as was taro, probably because ofthe high proportion of resistant starch (not digested in thesmall intestine) in both these foods. Kumara (Ipomoeabatatus), not tested previously, appears to have a relativelyhigh GI comparable with that of potato.

Regarding breakfast cereals, and other foods, it wasreassuring that the GI of Kelloggs Cornflakes, Uncle Ben’sparboiled rice, pasta, and instant noodles was virtuallyidentical to that reported from other countries.4 However,

the wide range of values for the small group of localbreakfast cereals, biscuits and fruit bar were not predictablefrom the food labels which underscores the need for clearguidance for people with diabetes. The reported low GIvalues for the Novartis medical nutritional products isclinically useful for those patients with diabetes. Theseproducts are recommended for use in the followingconditions: swallowing problems (Resource thickenedbeverage); for maintenance of energy requirements(Resource Diabetic); and for those requiring a high protein,fat free alternative to milk supplements (Resource fruitbeverage).

In summary we provide for the first time informationregarding the GI of some unique New Zealand foods whichare staples in the diet of Pacific people consuming atraditional diet as well as some commonly eatencommercially available foods. The findings confirm that it isappropriate to use the GI of some commercially availablefoods produced according to international specifications toprovide guidance regarding food choices. It is essentialhowever, to generate New Zealand data for locally producedbread, cereals, pasta, snack foods, as well as indigenoustraditional foods and locally grown fruit and vegetables.

Acknowledgements. We thank Margaret Waldron and Deborah Nicholsonfor specialist nursing and research assistance during this study. Financialsupport was from the Health Research Council, a Laurenson Award,Novartis Nutrition Australasia Pty Ltd and Hubbards Foods Ltd.

Correspondence. Dr Tracy Perry, Department of Human Nutrition,University of Otago, PO Box 56, Dunedin.

1. Jenkins DJ, Wolever TM, Taylor RH et al. Glycemic index of foods: a physiological basis forcarbohydrate exchange. Am J Clin Nutr 1981; 34: 362-6.

2. Crapo PA, Reaven G, Olefsky J, Alto P. Postprandial plasma-glucose and -insulin response todifferent complex carbohydrates. Diabetes 1977; 26: 1178-83.

3. Jenkins DJ, Wolever TM, Jenkins AL et al. The glycaemic index of foods tested in diabeticpatients: A new basis for carbohydrate exchange favouring the use of legumes. Diabetologia1983; 24: 257-64.

4. Foster-Powell K, Brand-Miller J. International tables of glycaemic index. Am J Clin Nutr1995; 62: 871S-93S.

5. Wolever TM, Katzman-Relle L, Jenkins AL et al. Glycaemic index of 102 complexcarbohydrate foods in patients with diabetes. Nutr Res 1994; 14: 651-9.

6. Jenkins DJ, Wolever TM, Buckley GB et al. Low-glycaemic index starchy foods in thediabetic diet. Am J Clin Nutr 1988; 48: 248-54.

7. Fontvielle AM, Acosta M, Rizkalla SW et al. A moderate switch from high to low glycaemic-index foods for 3 weeks improves the metabolic control of Type 1 (IDDM) diabetic subjects.Diab Nutr Metab 1988; 1: 139-43.

8. Wolever TM, Jenkins DJ, Vuksan V et al. Beneficial effect of low-glycaemic index diet inoverweight NIDDM subjects. Diabetes Care 1992; 15: 562-4.

9. Brand Miller JC. Importance of glycemic index in diabetes. Am J Clin Nutr 1994; 59 (suppl):747S-52S.

10. Salmeron J, Jenkins DJ, Ascherio A et al. Dietary fibre, glycemic load, and risk of NIDDM inMen. Diabetes Care 1997; 20: 545-50.

11. Salmeron J, Manson JE, Stampfer MJ et al. Dietary fibre, glycemic load, and risk of NIDDMin women. JAMA 1997; 227: 472-7.

12. FAO/WHO. Carbohydrates in human nutrition. Report of a Joint FAO/WHO ExpertConsultation. Rome, Italy; 1998.

13. Wolever TM, Jenkins DJ, Jenkins AL, Josse RG. The glycemic index: methodology andclinical implications. Am J Clin Nutr 1991; 54: 846-54.

14. Burlingame BA, Milligan GC, Spriggs TW, Athar N. The Concise New Zealand FoodComposition Tables. New Zealand Institute for Crop and Food Research. 3rd ed.Palmerston North and Ministry of Health, Wellington; 1997.

15. Wolever TM, Jenkins DJ. The use of the glycaemic index in predicting the blood glucoseresponse to mixed meals. Am J Clin Nutr 1986; 43: 167-72.

16. Wolever TM, Csima A, Jenkins DJ et al. The glycaemic index: variation between subjects andpredictive difference. J Am Coll Nutr 1989; 3: 235-47.

17. Wolever TM. The Glycemic Index. Aspects of Some Vitamins. Minerals and enzymes inhealth and disease. Bourne GH editor. World Rev Nutr Diet 1990; 62: 120-85.

18. Indar-Brown K, Norenberg C, Madar Z. Glycemic and insulinemic responses after ingestionof ethnic foods by NIDDM and healthy subjects. Am J Clin Nutr 1992; 55: 89-95.

19. Liljeberg H, Björk I. Bioavailability of starch in bread products. Postprandial glucose andinsulin responses in healthy subjects and in vitro resistant starch content. Eur J Clin Nutr1994; 48: 151-63.

20. Brand JC, Foster KA, Crossman S, Truswell AS. The glycaemic and insulin indices ofrealistic meals and rye breads tested in healthy subjects. Diab Nutr Metab 1990; 3: 137-42.

21. Liljeberg H, Granfeldt Y, Björk I. Metabolic responses to starch in bread containing intactkernels versus milled flour. Eur J Clin Nutr 1992; 46: 561-75.

28 April 2000 142


Worldwide, ovarian cancer is the seventh commonest cancerin women. In 1994 in New Zealand there were 258 newregistrations of ovarian cancer which represents 4% of allfemale cancers registered. In that year there were 161 deathsfrom ovarian cancer representing 4% of all female deathsfrom cancer. 42% of cases occurred in those aged less than60 years.1 The majority of women present with widespreaddisease and less than one-third of these survive five years.2

Treatment for advanced ovarian cancer is usually surgery toreduce the size of the tumour followed by a course ofcombination chemotherapy. Widely used chemotherapyregimens for ovarian cancer include carboplatin orcisplatinum and cyclophosphamide. Patients who undergooptimal debulking surgery followed by chemotherapy of thistype have approximately 30% probability of surviving fouryears.3

A relatively new drug, paclitaxel, has been combined withcisplatinum resulting in improved survival for women withadvanced ovarian cancer. The study reported by Maguire et al4showed a fourteen month increase in median survival forwomen treated with paclitaxel and cisplatin compared to thetraditional regimen of cisplatin and cyclophosphamide.Another study5 reported improved survival of ten months. Suchresults have led to the adoption of paclitaxel and cisplatinum asthe ‘standard’ first line chemotherapy regimen for advancedovarian cancer in several countries, and also in a limitednumber of centres in New Zealand. Of interest is the associatedcost of treatment and how decisions are made regardingbenefits measured in terms of months of survival comparedwith actual costs of the drug therapy. Two studies have come todiffering conclusions regarding the cost-effectiveness ofpaclitaxel. One concluded that paclitaxel is cost-effective “undergenerally accepted criteria” with a total (inpatient andoutpatient) incremental cost per year of life gained ofUS$41 042,6 whereas the second indicated a total incrementalcost per year of life gained of Canadian $32 213 and concludedthat “...paclitaxel comes at a high price...”.7 It is clear that in therealm of managed care there is difficulty in agreeing whethercost-effectiveness ratios are indeed cost-effective in practice. Inone cancer centre in New Zealand, the average cost per patientof six cycles of paclitaxel and cisplatinum was $15 681compared with an average cost per patient of $886 for six cyclesof cyclophosphamide and carboplatin. The differential averagecost per patient was, therefore, $14 795 (M Jeffery, personal

communication). This has funding implications as the averagechemotherapy cost of treating the 50-60 patients diagnosedeach year in the South Island with the paclitaxel regimen is inthe region of $900 000 compared with an average cost of$50 000 for the cyclophosphamide and carboplatin regimen.One of the key questions to consider is how much benefit interms of months of extra survival justifies the cost ofintroducing a new treatment?

The aim of this study was first, to investigate how muchindividuals would be willing to pay for a new form ofchemotherapy given various survival and drug cost scenarios;second, to investigate whether results were influenced by thework perspectives of individuals who worked in differentareas of health care; and third, to pilot a questionnairedesigned to collect such data with the aim of undertaking asimilar project on a wider scale.

MethodsDesign. The study was approved by Southern Regional Health EthicsCommittee (Otago). It was cross-sectional and designed to test thehypothesis that different groups of individuals associated with the healthcare system had different views on cost/benefit issues associated with thetreatment of patients with advanced ovarian cancer.Sample. A convenience sample consisted of five groups: (i) managersworking for a Crown Health Enterprise (CHE) (n=82); (ii) managersworking for a purchasing authority (n=70); (iii) senior doctors (n=126);(iv) trained nurses working on general medical and general surgical wards(n=94); (v) Trained nurses working on oncology wards (n=19). Themanagers working for the CHE, senior doctors and trained nurses wereall based in the same hospital which was one of two hospitals providingtertiary cancer services for the purchasing authority.Questionnaire. There were three parts to the questionnaire designed forthis study. The first included two questions which used rating scales toquantify (i) how much an individual would be willing to pay for the newchemotherapy treatment given so many extra months of life; and (ii) theminimum number of months of extra life that would be hoped for given aspecified cost (to the individual) of the new chemotherapy. The twoquestions in the second part were similar but related to how muchrespondents believed a public health system should be prepared to spendfor the new chemotherapy treatment for ten people (the average numbertreated per year in the study centre); and the third part containedsociodemographic questions. The questionnaire was prefaced by a briefdescription of ovarian cancer and the effects of the standardchemotherapy treatment and the new chemotherapy regimen.

Rating scales with fixed anchors were selected rather than open-endedquestions to reflect the nature of the issue being explored. Thirteenmonths is the median number of additional months survival which can behoped for three years following treatment with the paclitaxel regimen,and NZ$16 000 is the average cost for six cycles of the new regimen in

Willingness to pay for new chemotherapy for advanced ovarian cancer

Jenny Morris, Senior Lecturer, Institute of Health Studies, University of Plymouth, Pool, Cornwall, England;David Perez, Senior Lecturer, Department of Medicine, University of Otago, Dunedin.

Aims. To investigate whether health care professionals haddiffering attitudes about cost and survival issues in regardto the treatment of advanced ovarian cancer.Methods. A questionnaire designed for the study wasmailed to hospital doctors, hospital nurses, Crown HealthEnterprise (CHE) managers and health authority managerswho were working in the same geographical area. A total of391 questionnaires were mailed and 186 were completedreflecting an overall response rate of 47%. Replies werereceived from 46% of the health authority managers, 61%of the CHE managers, 59% of the doctors and 27% ofnurses.

Abstract

Results. Results indicated a threshold for how muchindividuals were willing to pay. Group differences wereevident with the nurses and CHE managers willing tospend less than the other two groups. In addition, thosewith experience of cancer treatment were willing to spendless than those with no such experience.Conclusions. There are differences in attitudes to costsand benefits of treatment between individuals working indifferent areas of health care. Health sector workers valuetreatments which extend life expectancy without necessarilybeing curative.

NZ Med J 2000; 113: 143-6

143 28 April 2000


the treatment centre where this study was undertaken. The questionnaire,which is available from the authors, was pre-tested successfully on aconvenience sample of academics working for the Otago Universitymedical school and a consecutive series of seven patients receivingchemotherapy for breast cancer on an outpatient basis.Procedure. Lists of individuals working in the groups identified wereobtained from the relevant departments. Between June and November1997, explanatory letters, copies of the questionnaire and reply paidenvelopes were sent to all individuals inviting them to participate. Onereminder letter and a further copy of the questionnaire was sent to non-responders two weeks later.Data Analysis. Data were analysed using the SPSS statistical package forWindows.8 Differences in data between groups were analysed using theKruskal-Wallis analysis of variance for non-normally distributed data.Data were also analysed using the Mann-Whitney test for non-normallydistributed data to establish whether experience of cancer treatmentinfluenced the results.

ResultsResponse rate. Completed replies were received from 32(46%) of the managers who worked for the purchasingauthority, 50 (61%) of the CHE managers (an additionalfour managers refused to participate with no reasons stated),75 (59%) of the senior doctors (a further doctor refused toparticipated with no reason provided, and a further fourquestionnaires were discarded due to missing information),22 (23%) of the general nurses (an additional three refusedto complete the questionnaire) and seven (37%) of theoncology nurses. Due to the small number of nurses whoreplied their data have been combined.Characteristics of the sample. It can be seen (Table 1) thatthere were more females in all groups except the hospitaldoctors, more individuals in the older age group in thegroups of CHE managers and doctors, and fewer individualswith experience of cancer treatment in the group of healthauthority managers. Although the numbers are small, fewerof the nurses were married or in de facto relationships. It isnot possible to comment on the extent to which respondersdiffered from non-responders as this information is notknown.

Table 1. Characteristics of the sample (% in parentheses).

CHE Health Hospital Hospital χ2 Totalmanagers authority doctors nurses (n=186)

(n=50) managers (n=75) (n=29)(n=32)

Sex Male 17 (35) 10(31) 62 (83) 3 (10) 59 92 (50) Female 32 (65) 22(69) 13 (17) 26 (90) p<0.0001 93 (50)

Age Group 18-40 17 (35) 19(59) 24 (34) 14 (48) n.s. 74 (41) 41-60 31 (65) 13(41) 47 (66) 15 (52) 106 (59)

Marital Status Single 5 (10) 9(28) 1 (1) 10 (34) 25 (14) Married/de facto 38 (78) 22(69) 69 (93) 13 (45) 37 142 (77) Divorced/ separated/ 6 (12) 1 (3) 4 (5) 6 (21) p<0.0001 17 (9) widowed

Cancer Treatment* Yes 29 (59) 11(34) 38 (51) 16 (55) n.s. 94 (51) No 20 (41) 21(66) 37 (49) 13 (45) 91 (49)

*The question asked whether the individual or any members of their family hadbeen treated for cancer. Due to missing data, the total numbers do not alwaysequate with subgroup totals.

Costs and benefits. Results regarding the amount of moneyindividuals would be willing to pay for the newchemotherapy for themselves or a member of their family ifit was not available in the public health system, and howmuch they believed the public health system should spendfor courses of chemotherapy treatment for ten patients, areshown in Table 2. Willingness to pay themselves revealed

statistically significant differences between groups for eachsurvival category (three, eight, fourteen and 25 months extralife gained). Doctors were willing to pay more for the newchemotherapy for each scenario. The hospital nurses mostresembled the CHE managers in terms of the medianresults, these two groups being willing to pay less than theother two groups. The median was zero dollars for these twogroups when only three extra months of life might begained. As might be expected, there was an increase in theamount individuals were willing to pay as the number ofextra months of life to be gained from the new treatmentincreased. This was consistent for all groups.

Table 2. Median amount individuals willing to pay for newchemotherapy and amount public health system shouldspend for treatment for ten patients (interquartile rangein parentheses).

Extra months of life per patient to begained

3 8 14 25Individual willing to pay

(scale NZ$0-25 000)CHE manager 0 (4) 3 (8) 9 (9) 12 (15)Health authority managers 3 (6) 6 (9) 12 (12) 18 (13)Hospital doctors 3 (12) 8 (13) 14 (15) 24 (10)Hospital nurses 0 (3)* 3 (8)† 6 (7)† 12 (18)†

Public health system should spend(scale NZ$0-150 000)

CHE managers 20 (40) 40 (62) 65 (72) 120(100)Health authority managers 15 (40) 40 (40) 80 (60) 125 (35)Hospital doctors 20 (40) 60 (40) 105 (47) 150 (37)Hospital nurses 20 (40) 60 (60) 80 (75) 150 (50)

Differences between groups statistically significant: *p<0.01; †p<0.001.

Conversely, there were no significant differences betweengroups when respondents were asked to consider this issuefrom the perspective of the public health system.

Table 3 gives the median minimum number of extramonths of life per patient which would be consideredworthwhile given specified costs to both the individualand to the public health system for the newchemotherapy. With regard to costs to the individual,differences between groups were significant for the lowerand higher cost scenarios ($500, $17 500, $25 000) andwhen treatment was free. Of note are data from thenurses who would expect a median increase in survivalwhen treatment was free similar to that when treatmentwould cost individuals $10 000 or $17 500. Furthermore,for most of the cost scenarios, the median minimumnumber of extra months of life considered worthwhile islower for the group of doctors. Results from the hospitaldoctors suggest a link between the dollar amount and themedian number of months of extra life hoped for, but thisis not the case for the other groups. Results for the costscenario of $25 000 suggest a ceiling effect although theinterquartile range for the group of doctors shows a greatdeal of variability within this group.

Data from the question addressing costs to the publichealth system also revealed significant differences betweengroups (Table 3). Trends were similar to those describedabove with the median minimum number of extra months oflife considered worthwhile being lowest for the group ofdoctors.Experience of cancer treatment. Individuals withoutfamily experience of cancer treatment (n=94) were willing topay more for the new treatment than the others, (n=91,Table 4). There were no significant differences between thegroups in terms of the minimum number of extra months oflife considered worthwhile given different cost scenarios.

28 April 2000 144


Table 3. Median minimum number of extra months of life per patientconsidered worthwhile (interquartile range in parentheses).

CHE Health Hospital Hospitalmanagers authority doctors nurses

managers

If cost individual:Chemotherapy free 6 (9) 3 (5) 3 (3) 15 (22)†$500 6 (9) 3 (2) 3 (6) 6 (17)*$3000 9 (11) 6 (9) 4 (9) 7 (17)$10 000 14 (13) 12 (11) 10 (12) 16 (14)$17 500 24 (8) 18 (12) 15 (15) 18 (10)*$25 000 25 (2) 25 (4) 24 (13) 25 (7)*If cost public health system(to treat 10 patients):$20 000 6 (7) 3 (3) 3 (3) 6 (9)$50 000 12 (6) 8 (6) 6 (8) 12 (11)*$85 000 12 (6) 12 (6) 9 (7) 15 (8)†$95 000 15 (11) 13 (8) 12 (6) 18 (10)†$110 000 20 (13) 18 (12) 15 (9) 21 (8)*$150 000 25 (8) 24 (11) 19 (13) 25 (5)*

Differences between groups statistically significant: *p<0.01; †p<0.001. Scalerange 0-25 months.

Table 4. Experience of cancer treatment and impact on cost-benefitattitudes.

Experience of cancer No experience of cancertreatment treatment

Median amount in NZ$x1000Individuals willing to pay (interquartile range in

parentheses)3 extra months of life 0 (6) 3 (6)8 extra months of life 6 (8) 7 (9)*14 extra months of life 9 (8) 12 (17)*25 extra months of life 15 (18) 21 (13)*

Median minimum number extra months of lifeconsidered worthwhile (interquartile range in

parentheses)Chemotherapy Free 3 (15) 3 (10)$500 3 (9) 3 (10)$3000 6 (12) 6 (9)$10 000 12 (15) 12 (17)$17 500 18 (14) 18 (13)$25 000 25 (13) 25 (8)

*p<0.01

DiscussionEvaluation of new, expensive chemotherapy drugs poses adilemma for most countries and health systems. This relatesto the modest improvements in clinical outcomes versussubstantial expenditure involved. At present there is nocoherent mechanism in the New Zealand public healthsystem to evaluate the cost effectiveness of drugs such aspaclitaxel. Decisions are made locally in the six cancercentres based on various criteria. Consequently women inNew Zealand with ovarian cancer have variable access topaclitaxel. In the Waikato the drug has been available for anumber of years but it is not funded in Auckland. In 1995the Medicines Policy Committee at Auckland Hospitalpublished its reasons for not funding paclitaxel and theseprincipally related to a lack of randomised studiesdemonstrating improved survival.9 Randomised studies havesubsequently demonstrated a survival advantage yet the drugremains unavailable in the Auckland public health system.

Pharmaco-economic analyses for the use of paclitaxel inovarian cancer have not been performed in New Zealand.Those performed overseas have reached differingconclusions although the majority report the incrementalcosts and associated benefits are in line with costs acceptedin other spheres of medicine.6,7,10,11 Pharmaco-economicanalyses have contributed to the acceptance of publicfunding for paclitaxel in a number of countries includingAustralia where the drug has been available for first line

treatment of advanced ovarian cancer since 1997. Thepresent study does not address pharmaco-economic issuesbut assesses the perceptions and judgements of healthprofessionals concerning levels of effectiveness for paclitaxelin ovarian cancer which justify the inherent costs.

All four groups of health sector workers surveyed indicated awillingness to spend substantial amounts of personal funds toobtain paclitaxel for themselves or partners should they developadvanced ovarian cancer. However, there were clear differencesconcerning thresholds for willingness to pay. For examplehospital managers and nurses would not spend anything formodest extensions of life expectancy. Conversely whenextension of life expectancy was 25 months all groups agreedwith expenditure of $12 000 to $24 000 per woman treated.Since the median increase in life expectancy with paclitaxel isten to thirteen months it is instructive to consider the amountsthe respondents would pay for life extension of fourteenmonths. These amounts varied from a median figure of $6000for nurses to $14 000 for hospital doctors with hospitalmanagers and health authority managers recording $9000 and$12 000 respectively. Considering that the drug costs for acourse of paclitaxel and cisplatinum are approximately$16 000 the hospital doctors and health authority managersresponses are within range of the actual drug costs whereas theamounts nurses and hospital managers would be prepared topay fall well short of drug costs.

Interestingly, when personal funds were the focus, twogroups would not be prepared to spend money on newtreatment when only three extra months of life were likely tobe gained. When this issue was posed in terms of publichealth system funding, all groups felt that money should bespent on treatment even for modest extra gains in survival.

When the reverse question was asked concerning extramonths of life expectancy for a given expenditure to theindividual, hospital doctors recorded the lowest median lifeextension of fifteen months for expenditure of $17 500. Themedian for hospital managers was 24 months whilst that forhealth authority managers and nurses was eighteen months.The hospital doctors’ response was within range of actualoutcomes whereas the other groups had expectations whichwere considerably above this. This discrepancy betweenperceptions of health workers from different sectorsconcerning utility outcomes has been noted by otherinvestigators. Slevin et al12 reported that medical oncologistswere much more likely to accept chemotherapy for relativelyminor benefits than were radiation oncologists or nurses.

Of additional interest in the present study was the responseof those who had some experience of cancer treatment, eitherfor themselves or a family member. Those with experience ofcancer treatment were inclined to spend less of their personalfunds to achieve a specified extension of life expectancycompared to those with no experience. The reason for thisdifference was not solicited but may reflect differingappreciation of the side effects of treatments.

One objective of the study was to pilot the questionnaire.Results indicate that the participants were able to understandwhat was required and to complete the questionnaire in themanner requested. Only three questionnaires had to bediscarded due to large amounts of missing information and,on inspection, the missing data related to questions aboutthe public health system. Only two individuals stated thatthe questionnaire was too difficult. This information,together with the data collected at the pre-pilot stagesuggests that the format of the questionnaire is appropriatefor further work in this area.

A strength of this study is the framing of the question interms of the individual’s willingness to pay for treatment.Willingness to pay for treatment is a measure of the

145 28 April 2000


Trauma form documentation in major trauma

James Hamill, Trauma Fellow; Rhondda Paice, Trauma Patient Care Coordinator; Ian Civil, Director of TraumaServices, Trauma Services, Auckland Hospital, Auckland.

AbstractAims. To examine the impact of a standardised traumaform for documentation in cases of major trauma, aprospective study was undertaken.Methods. Records written by medical staff pertaining tothe assessment and treatment of major trauma patients inthe resuscitation room were scored against a panel ofparameters derived from advanced trauma life supportguidelines. Demographics, aetiology, trauma scores andoutcome data were obtained from a trauma registry.Attitudes of medical staff involved in major trauma to thetrauma form were assessed using a questionnaire.

Results. The trauma form was used in 53 of 69 (76.8%)consecutive cases of major trauma seen over a three monthperiod. No significant differences existed in demographics,aetiology, trauma scores or outcome between form andformless groups. In the form group, a median of 44 of 51(86.3%) relevant information parameters were documentedversus 32 of 51 (62.7%) in the formless group, p<0.0001. Apositive approach to the trauma form was indicated by thequestionnaire results.Conclusion. The use of a standardised form improvesdocumentation in major trauma.

NZ Med J 113; 2000: 146-8

Initial assessment and treatment of major trauma casesinvolves rapid acquisition of substantial amounts ofinformation and performance of multiple procedures.1

Documentation of these events by the attendant EmergencyDepartment (ED), Intensive Care Unit (ICU) and GeneralSurgical (GS) doctors is frequently incomplete. We haveobserved that written records are typically brief, sometimesillegible, and the three notes (ED, ICU and GS) oftencontain duplicated but little complementary information.

To remedy these problems, we designed and introduced atrauma form, (TF) to our ED for use in cases of majortrauma. Medical staff in the ED, ICU and GS wereintroduced to the form and it was made readily available inthe resuscitation room. To examine the impact of the TF onquality of documentation, we prospectively audited its use.

MethodsThe TF was designed by the trauma service in consultation with staff inED and ICU. It consisted of two pages with check boxes, areas for shortwritten responses, a diagram and a space for names and signatures of all

medical attendants at a trauma call. Two further pages provided space fora narrative of events and a summary of injuries. The total paper spaceconsisted of two sides of one A3 sheet. The team leader was responsiblefor completion of the TF and delegation of other team members tocontribute to relevant parts of the form.

Major trauma was defined as cases to which the trauma team werecalled. According to protocol, the trauma team was called for traumapatients who: (a) were unstable; (b) sustained penetrating trauma to thehead, neck or torso; (c) sustained major chest, spine, pelvic or limbtrauma; or (d) had a mechanism of injury associated with a highlikelihood of major trauma.

Over a three month period January to March 1999, records of the initialassessment of major trauma cases were prospectively scored forcompleteness of information against a panel of 51 parameters (Table 1).Parameters were chosen for their relevance to the prediction, detection andtreatment of injuries as per Advanced Trauma Life Support teaching.1 Itwas considered that, with ideal trauma documentation, a note on eachparameter would be obtainable from the medical record. All parameters arerepresented on the TF. In road traffic crash cases, five further parameterswere scored: estimated speed, seatbelt use, occupant position in the vehicle,direction of impact and whether or not the occupant was ejected ortrapped. All records written by medical staff pertaining to the assessmentand treatment of the patient in the resuscitation room were scrutinised andscored. To avoid bias, trauma teams were not made aware of the study.

individual’s subjective assessment of the utility of atreatment. Whilst income may have influenced the results, itis unlikely to be the sole explanation. Salaries for hospitaland health authority managers are similar yet theirwillingness to pay differed.

The study also has weaknesses. First, the questionnaireresponse rate was low, particularly for nurses. One reasonmay be that nurses do not perceive themselves to bedirectly involved with decisions about resource allocationand were not, therefore, motivated to complete thequestionnaire. Second, individuals outside the health sectorwere not surveyed nor were women with ovarian cancer.Slevin et al reported that patients with cancer have muchlower benefit thresholds than well health professionals orthe population at large.12

Despite these limitations the study indicates that healthsector workers value treatments which extend life expectancywithout necessarily being curative. In fact chemotherapy forovarian cancer does offer approximately a ten percent chanceof cure and it is likely that paclitaxel and cisplatinum willmodestly improve the cure rate. The issue of cure was notposed in the questionnaire but it is probable that had thisbeen done the respondents’ benefit thresholds would havefallen. Health sector funders should acknowledge data of this

type when considering allocation of funds rather than fallinginto the trap of considering expensive cancer therapy tooineffective, too toxic or lacking general support.

Correspondence. Dr Jenny Morris, Institute of Health Studies, Universityof Plymouth, Penhaligon Building, Trevenson Lane, Pool, Cornwall, EnglandTR15 3RG. Email: [email protected]

1. New Zealand Health Information Service. Cancer. New registrations and deaths 1992.Ministry Of Health, Wellington; 1997.

2. Kornblith A, Thaler H, Wong G et al. Quality of life of women with ovarian cancer.Gynecol Oncol 1995; 59: 231-42.

3. Cannistra S. Cancer of the ovary. N Engl J Med 1993; 329: 1550-9.4. McGuire W, Hoskins W, Brady M et al. Cyclophosphamide and cisplatin compared with

paclitaxel and cisplatin in patients with Stage III and Stage IV ovarian cancer. N Engl J Med1996; 334: 1-6.

5. Stuart G, Bertelsen K, Mangioni C et al. Updated analysis shows a highly significantimproved overall survival for cisplatin-paclitaxel as first line treatment of advanced ovariancancer: Mature results of the EORTC-GCCG, NOCOVA, NCIC, CTG and ScottishIntergroup Trial. Proceedings of ASCO 1998; 17: 361a.

6. McGuire W, Neugut A, Arikian S et al. Analysis of the cost-effectiveness of paclitaxel asalternative combination therapy for advanced ovarian cancer. J Clin Oncol 1997; 15: 640-5.

7. Elit L, Gafni A, Levine N. Economic and policy implications of adopting paclitaxel as first-linetherapy for advanced ovarian cancer: an Ontario perspective. J Clin Oncol 1997; 15: 632-9.

8. Norusis M. SPSS for Windows. Base System User’s Guide. Release 6.0 SPSS Inc., Chicago, 1993.9. Goh K, Bryant L, Black P. Guidelines for the use of expensive medicines. NZ Med J 1995;

108: 38-40.10. Covens A, Boucher S, Roche K et al. Is paciltaxel and cisplatin a cost-effective first-line

therapy for advanced ovarian carcinoma? Cancer 1996; 77: 2086-91.11. Berger K, Fischer T, Szucs TD. Cost-effective analysis of paclitaxel and cisplatin versus

cyclophosphamide and cisplatin as first-line therapy in advanced ovarian cancer. A Europeanperspective. Eur J Cancer 1998; 34: 1894-1901.

12. Slevin M, Stubbs L, Plant H et al. Attitudes to chemotherapy: comparing the views of patientswith cancer with those of doctors, nurses and general public. BMJ 1990; 300: 1458-60.

28 April 2000 146


Table 1. Items scored.

History Secondary surveyDate of arrival Scalp/skull examinedTime of arrival Maxillofacial itegrityDate of injury Ears examinedTime of injury Nose examinedMechanism of injury Mouth examinedAllergies Cervical spine examinedMedicines Neck soft tissues examinedPast medical history/comorbidity Chest wall injury, or notPrehospital Chest sufficient to detect pneumothorax or haemothoraxScene respiratory rate Abdominal inspectionScene blood pressure Abdominal tenderness or guardingScene Glasgow coma score Rectal examinationPrimary survey Genital or vaginal examinationAirway adequacy Pelvic girdle stable, or notRespiratory rate (or mechanical ventilation) Back and spinal examinationSaO2 Upper limb injury, or notHeart rate Lower limb injury, or notBlood pressure Investigations and planPeripheries well perfused, or not Blood grouped/cross-matchGlasgow coma score Lateral cervical spine radiographPupil size and/or light reflex Chest radiographUpper and lower limb movement present or not Pelvic radiographPatient temperature Documented planOxygen given SignaturesIV lines placed Emergency department doctorNaso- or oro-gastric tube Intensive care doctorUrinary catheter General surgical doctorResuscitation fluids given

Demographic data, aetiology, trauma scores and outcome data wereobtained from the Auckland Hospital Trauma Registry.2

At the completion of the study period a questionnaire was sent to allICU and GS registrars and ED consultants. Responses were sought onthe following statements: (a) the trauma form provides a useful means ofdocumentation; (b) the trauma form is the only medical documentation ofa trauma call needed in most cases; and (c) the percentage of time I usethe trauma form for documenting trauma calls is? The response to eachstatement was marked on a visual analogue scale.3 Each visual analoguescale consisted of a 100 mm long horizontal line, with disagree or 0% atthe left end and agree or 100% at the right end of the line. Each markwas converted to a score between 0 and 100 by measuring in millimetresfrom the left end of the line.

Records in which the TF was used (form group) were compared withthose without its use (formless group). Statistical significance was testedby Fisher’s exact test and the Mann-Whitney U test, with p<0.05considered significant.

ResultsOver a three month period, documentation was scored from69 consecutive major trauma cases. The TF was used in 53cases (76.8%) and no TF in the remaining sixteen. The TFwas the only form of documentation used in eighteen cases(26.1% of all cases, 34% of cases of TF use).

Table 2 presents a comparison of the characteristics of thetwo groups. No significant differences were present indemographic details, aetiology, physiological stability of thepatients on admission (as measured by the revised traumascore), injury severity score or survival.

Table 2. Group comparisons.

Form Formless pn=53 n=16

Age, median (range) 31 (17-83) 31 (17-84) 0.8Sex, male (% male) 43 (81.1) 12 (85.7) 0.9Aetiology, blunt (% blunt) 52 (98.1) 15 (93.7) 1.0RTS, median (range) 7.84 (3.8-7.84) 7.84 (5.97-7.84) 0.9ISS, median (range) 14 (1-66) 15 (1-24) 0.4Survival, n (%) 48 (90.6) 14 (87.5) 0.3

RTS, revised trauma score. ISS, injury severity score.

Documentation scores are presented in Table 3. With TFuse, significantly more information was documented in thehistory, prehospital recordings, primary survey and

secondary survey categories. Documentation of investigationresults, treatment plan and legible signatures were similar inthe two groups. In the form group, 20 (37.7%) were roadtraffic crash victims versus seven (43.7%) in the formlessgroup (p=0.8). In both groups, a median of the four of thefive parameters pertaining to road traffic crashes weredocumented (p=0.6).

Table 3. Number of parameters documented with and without traumaform use: median (range).

Form Formless pn=53 n=16

History, n=8 7 (2-8) 5 (2-8) <0.0001Prehospital, n=3 3 (0-3) 0 (0-3) <0.0001Primary survey, n=15 14 (7-15) 9 (2-12) <0.0001Secondary survey, n=17 16 (6-17) 11 (3-14) <0.0001Investigations & plan, n=5 4 (0-5) 4 (1-5) 0.8Signatures, n=3 2 (0-3) 2 (1-3) 0.1Total, n=51 44 (25-51) 32 (16-41) <0.0001

Of the 23 questionnaires distributed, seventeen were returned.Questionnaire results, presented in Figure 1, suggest apositive attitude to the usefulness of the TF (median 72 mm)but a mixed response to the statement that the TF could bethe only documentation in most cases (median 50 mm).

DiscussionPoor quality medical documentation is well recognised intrauma4-7 and in other fields of medicine.8,9 The presentstudy demonstrates an increased amount of relevant clinicalinformation documented on major trauma cases with the useof a standardised form. This is in agreement with otherstudies.10-13 Kondziolka et al found a significant increase inthe amount of information gathered with the use of astandardised form for cases of neurotrauma.10 Walters andMcNeill examined the effect of separate trauma forms forthe team leader, anaesthetist, general surgeon and eachspeciality surgeon and observed improvement in most itemsof information recorded, especially by the team leader.11

Williams et al found the documentation of four vital signs,respiratory rate, pulse rate, blood pressure and Glasgowcoma score, to be significantly improved after theintroduction of a trauma form.12

0 10 20 30 40 50 60 70 80 90 100

Visual analogue scale (mm)

Useful?

Only notes needed?

Utilisation (% time)

Figure 1. Boxplot of questionnaire results measured from visualanalogue scales (higher scores for positive results). Shown aremedian and interquartile range.

147 28 April 2000


Right idea, wrong model. Why general practitioners are not succeeding with

preventive care

Stuart Foote, General Practitioner, Hastings.

NZ Med J 2000; 113: 148-9

GPs are familiar with the exhortations of interest groups, bethey the National Health Committee, Ministry of Health ortheir own College or IPA, to do better at various aspects oftheir work. Preventive care, opportunistic screening, get thesmear and immunisation rates up. Its obvious, isn’t it, thatprevention is better than cure? Yet the numbers are notbeing delivered and the more the exhortation the moremany GPs switch off. It is not surprising that GP morale isas low as it is when we are being repeatedly told, with thenumbers to back up the accusations, that we are failing.Why are we failing and what is necessary to make a change?

To answer these questions we need to look at the changesbuffeting GPs. Consumerism demands an immediateresponse to perceived needs and wants, available at a timeconvenient to the consumer, with immediate results, noerrors, highest quality and cheap price. Increasingly peoplewant a ‘one stop shop’ with good value for money, thevalues being those of the consumer.

The role of the GP is changing too. Trained in diseasedetection and destruction, GPs increasingly find that their

work involves disease prevention, lifestyle change andmanaging the concerns of the worried well. TheGovernment is encouraging this shift too, not by changingthe system but by contracting groups of GPs to takeresponsibility for the health needs (as defined byGovernment) of the community they care for. “You workout how to do it”, they seem to be saying. So GPs need tolearn new skills and manage on the basis of epidemiologicaldata as well as individual care. This will involve developingsystems of information collection and collation, differentways of delivering services to different sectors in thecommunity and formally assessing health needs.

The knowledge and technology explosion is leaving GPsbehind. The days of being a true generalist have probablyalready gone. None of us can keep fully up to date with it allso there is, most often, someone or some other organisationwho claim they can do it better, slicker and more efficiently.GPs will need to justify their place in the future, using newways of accessing information, better organisation and betterteamwork.

In the present study, the low rate of documentation ofprehospital recordings in the formless group probablyreflects the availability of this information elsewhere (theambulance run sheet). Recordings of vital signs in theresuscitation room are also recorded elsewhere (nursesrecords), but these signs are critical in the primaryassessment of major trauma cases and should also berecorded in the medical record. The TF increases theamount of history, primary and secondary surveyinformation recorded, probably by providing prompts,structure and by reducing the amount of writing required.

Documentation quality is not necessarily associated withthe quality of care provided in medicine.14-16 Fessel and vanBrunt, in assessing quality of care for two conditions,appendicitis and myocardial infarction, concluded thatneither quality nor quantity of recorded data was relevant tooutcome.14 Murphy and Jacobson found the quality of themedical record was negatively, but not significantly, relatedto short and medium term health status indices in asthmapatients.15 Humphries et al found no significant correlationbetween patient outcome and the use of preformatted chartsfor obstetric and gynaecology assessments in the emergencydepartment.16 Our data are consistent with this in that therewas no significant difference in survival observed betweenthe form and formless groups. Quality of documentationdoes, however, have important audit,14 research9 andmedicolegal5 implications.

The questionnaire results from the present study suggeststhe TF is considered a useful resource for major traumadocumentation but does not fully replace traditionaldocumentation methods. Utilisation of the TF was high(76.5%). Although ED, ICU and GS medical staff wereorientated to the form and encouraged to use it, TF useremained voluntary throughout the study period.

The decision against using the TF in sixteen cases appearsto originate from personal preference of the medical staff atthe time. With similar revised trauma scores and injuryseverity scores in the form and formless groups, it appearsthat neither physiological stability nor the severity of injuriesinfluence the decision to use the TF.

In conclusion, the increased amount of relevant clinicalinformation documented and the positive response by userssupports the place of a trauma form for documentation inmajor trauma.

Correspondence. James Hamill, Department of Paediatric Surgery,Christchurch Hospital, Private Bag 4710 Christchurch. Phone (03) 364 0640Fax (03) 364 1584 email: [email protected]

1. American College of Surgeons Committee on Trauma. Advanced trauma life support ®

program for doctors. Chicago, IL: American College of Surgeons; 1997.2. King M, Paice R, Civil I. Trauma data collection using a customised trauma registry: a New

Zealand experience. NZ Med J 1996; 109: 207-9.3. McCormack HM, Horne DJ, Sheather S. Clinical applications of visual analogue scales: a

critical review. Psychol Med 1988; 18: 1007-19.4. Solomons G. Trauma and child abuse. The importance of the medical record. Am J Dis

Child 1980; 134: 503-5.5. Frazier WH, Brand DA. Emergency department trauma care: priorities and documentation.

Surg Clin North Am 1980; 60: 1009-20.6. Macey AC, Tait GR, Rowles J et al. The M1 aircrash disaster documentation: lesions and

prospects? J Bone Joint Surg Br 1990; 72B: 1089.7. Hargarten S, Karlson T. Motor vehicle crashes and seat belts: a study of emergency physician

procedures, charges, and documentation. Ann Emerg Med 1994; 24: 857-60.8. Schoenfeld PS, Baker MD. Documentation in the paediatric emergency department: a review

of resuscitation cases. Ann Emerg Med 1991; 20: 641-3.9. Ross AC, Bishay M. The collection of research data: don’t ask the houseman. J Bone Joint

Surg Br 1994; 76B: (Suppl 1). 46.10. Kondziolka D, Schwartz ML, Walters BC, Mc Neill I. The Sunnybrook Neurotrauma

Assessment Record: improving trauma data collection. J Trauma 1989; 29: 730-5.11. Walters BC, McNeill I. Improving the record of patient assessments in the trauma room. J

Trauma 1990; 30: 398-409.12. Williams HR, Templeton PA, Smith RM. An audit of trauma documentation. Injury 1997;

28: 165-7.13. Ross RT, Hammen PF, Frantz EI et al. Gunshot wounds: evaluating the adequacy of

documentation at a level 1 trauma centre. J Trauma 1998; 45: 151-2.14. Fessel WJ, Van Brunt EE. Assessing quality of care from the medical record. N Engl J Med

1972; 286: 134-8.15. Murphy JG, Jacobson S. Assessing the quality of emergency care: the medical record versus

patient outcome. Ann Emerg Med 1984; 13: 158-65.16. Humphreys T, Shofer FS, Jacobson S et al. Preformatted charts improve documentation in

the emergency department. Ann Emerg Med 1992; 21: 534-40.

VIEWPOINT

28 April 2000 148


Third party payers want comprehensive, high quality‘packages of care’. While they want ‘coordinated care’, theyare not really concerned about who does the co-ordinating.GPs have felt that being the coordinator of care was theirsby right, but that notion is clearly wrong if there is a shortterm dollar to be saved by avoiding the gatekeeper – recentACC changes for instance. GPs will have to deliver suchpackages of care if they want to stay in business, and theywill need to include ways of delivering preventive care.

The current system of primary healthcare delivery isclearly not suited to respond to these pressures and a newone is needed. Equally, the current system is not providingthe preventive healthcare desired.

New Zealand provides a ‘free’ secondary service thatnurtures a belief that healthcare ought to be free. It thenexpects people to pay for services in primary care, evenservices that they may not want or understand why theyshould want them. Selling preventive care is not easy. The‘product’ is information that may lead to a future benefit forthe patient, which he or she will exchange for a currentbenefit (cash), for which there are many competingdemands. Future benefits are heavily discounted so the easeof sale depends on the person’s motivation. Motivation ishigh in the worried well and those recently frightened butoften low in those who heavily discount.

Future benefits are determined by the value that eachindividual places on them. The value that counts is that ofeach person. The values that are used when GPs are beingexhorted to do better depend on the objectives of theorganisation doing the exhortation. For instance, the valueof reducing the cost of diabetic complications in the futurerates highly with health funders when exhorting GPs tomanage diabetes better. The values that usually apply in theGP consultation are the values that the GP assumes thepatient ought to have. Usually, those are the GPs values.Only infrequently are the values of patients themselvesdetermined.

The only value that will make a difference is the patient’s value.

Placing value on future health is a form of saving. Saving,however, is a middle class habit. The very rich don’t have tosave, the poor have nothing to save and only those in themiddle perceive a need to and have something left over tosave. Health conserving behaviour commonly follows thesame pattern. This is the essence of the ‘preaching to theconverted’ that GPs are so familiar with. Related, butdifferent, is the Inverse Care Law, first proposed by J TudorHart1 30 years ago. Market forces reinforce the Inverse CareLaw particularly in areas of preventive care.

And let’s not forget that people buy a GP’s time for anagenda they bring with them-often not a single reason but asaved-up list. Dealing with that in fifteen minutes, let aloneseven, is frequently a challenge. Switching to the doctor’s orthe health service’s agenda during the time paid for by thepatient is frequently just not possible.

What, then, is to be done? Nothing will change unless thesystem changes. If our society really believes that thebenefits of preventive care will be worth more than the cost,then a system of care is needed to ensure that it happens.This will involve taking care of the money question beforethe consultation. Those in greatest need of assistance tochange will not pay for it directly.

In order to cope with up-to-date knowledge and skillrequirements primary healthcare will need to develop truemulti-disciplinary teamwork. Becoming a team member willrequire change for a lot of us doctors enjoying as we do thecomfort of being in charge. Forming effective teams willrequire moving rapidly away from the corner store approachto healthcare to organisations capable of managing thetechnology, knowledge and skills required today.

Changing from mainly care of individuals to managing thehealth of groups of people requires systems of care. Thesestart with education and awareness, lead into screening andcase finding, early intervention then continuing assessmentand review of the system. GPs and their teams can onlyprovide part of this system but that part needs to beorganised consistently and uniformly. Currently it is not.

Current medical training is, I hope, paying more attentionto the sciences of motivation and behaviour change thanpreviously. The existing doctors need to be encouraged toeducate themselves about these things and to use techniquesarising from those sciences in their practices.

The single most needed change is, I think, to move awayfrom the consultation being the unit of compensation for GPs.The long hallowed fee-for-service system in general practicemay well become its nemesis as the needs of the patients andthe health service change. The current system is focussed onattention to acute, episodic illness and encourages quick, highturnover, high resource utilisation practice. Those wanting todo better with the chronically ill, those needing emotional orpsychological support or to spend time with preventivemedicine are heavily disadvantaged. GP teams will need towork out ways of educating and helping groups of patients.Changing the way GPs are paid will shift the emphasis frompaying the doctor towards paying for the services of the team.

GPs will have to decide if they want to remain proprietorsof small businesses selling a product mainly oriented aroundacute episodic illness or become key players in teams thataddress the health needs of the public. The former is abusiness person role, the latter is a servant of the public.

No system will work unless all the participants are valuedand are satisfied and happy in their work. GPs need toensure that their training, experience and responsibility isadequately recognised and that both rewards and re-creationare adequate. Waiting for Governments to provide such acombination may be a long process. GPs themselves mustcome up with the solutions.

Correspondence. Stuart Foote, 403 Nelson Street North, Hastings. Email:[email protected]

1. J Tudor Hart. Three decades of the inverse care law. BMJ 2000; 320: 18-9.

149 28 April 2000


Research in general practice or general practice research

Murray Tilyard, Professor and Chair; Susan Dovey, Senior Research Fellow and Lecturer; Department of GeneralPractice, Dunedin School of Medicine, Unversity of Otago, Dunedin.

NZ Med J 2000; 113: 150-1

‘Research’ has been described as “organised curiosity”.1 It ischaracterised by a systematic approach to finding answers toquestions. Historically, research in general practice has usedthe resources of general practice (patients and doctors) toaddress questions arising from somewhere else, often frompublic health perspectives, from government or from otherexternal agencies. It is research in which general practiceplays a passive but essential role. This type of approach hasshaped common understandings of general practice research.In this paper, we present a case for a different view of generalpractice research. We propose a definition in which thequestions general practitioners ask are of dominantimportance. These questions are important to our patients,because their answers provide the key to better care forpatients. They are important to the health systems withinwhich we work, because their answers can suggest moreeffective ways to deliver health services. They are importantto our own professional growth, because in answering themwe learn more skills, become more reflective practitionersand participate more in the world beyond the practice doors.And they are important to the disciplinary growth of ourprofession because they underpin the fundamental principlesof general practice that set it apart from other medicaldisciplines. These questions define general practice research.They create a different body of knowledge from that builtfrom research in general practice. We consider the reasonsfor doing research, as well as the considerable barriers thatprevent most general practitioners from initiating anyresearch activity, in practice. Finally, we look forward to theprospects for general practice research that may be producedby a changing orientation in both health services andresearch funding agencies.

The traditional justification for general practice research isthat it is an essential part of establishing an academic base.But if that were the only reason for doing it, very little wouldever be done, and eventually the research-based discipline (ofgeneral practice) would disappear. Medical educationbecomes interesting when we challenge facts, and in theprocess our discipline grows.2 It is true that as we publish ourresearch, the work of general practice is exposed to academicreview by peers in other medical disciplines, and thediscipline of general practice becomes better defined. But inreality, we start doing research because we need practicalanswers to practical questions. It is also true that we arebuilding the academic base of general practice by adaptingthe philosophies and research approaches of other disciplinesto meet our needs. However, general practice cannot affordto sacrifice its practical orientation by becoming hung up onpondering the philosophic base of other disciplines. Thismethodological questioning is a characteristic of the socialsciences that is best left to the social sciences, rather thancarried over into general practice along with the methodsborrowed from the social sciences. The academic andphilosophical positions of general practice grow from a verypragmatic base that must be the first concern of generalpractice research.

Every day, general practitioners are faced with newproblems their patients bring. In every consultation a generalpractitioner draws on a body of knowledge built fromindividual experience. Because of the diversity of patients and

problems however, an individual’s experience, althoughnecessary, is often not a sufficient knowledge base fromwhich to recommend the best care. We must shareexperiences, and develop a broader body of knowledge sothat we understand what clues a particular group ofsymptoms give about the possible diagnosis or prognosis, orin deciding what treatment should be used for a patient.This evidence base represents current ‘best practice’. It isaccumulated from previous research and forms the basis of ascientific approach to the practice and delivery of care ingeneral practice.

General practice has embraced the move to evidence basedmedicine but there are problems in its practical applicationin this setting. At the root of these problems is the fact thatby tradition, most of the evidence we have relied upon hasnot come from general practice, with the result that themessages seem sometimes odd, and sometimes they areincorrect. It can be erroneous, for instance, to transferevidence from a hospital-based ENT clinic into the clinicalcare of children coming to see their general practitioner withotitis media. But until we knew who these children were, andwhat was the most effective treatment to provide for otitismedia in general practice, we were judged against thehospital-based standards of care, and found wanting.3 Westill carry the legacy of the mixed messages that researchevidence has provided us on otitis media management.4

Unfortunately, no discussion of research in generalpractice is complete without reference to the barriers to itsundertaking. These are important because they are a majorfault in the system: a cause of much lost valuable knowledge.We find ourselves in a ‘catch-22’ position where we need tohave more research directed toward addressing and resolvingthe barriers to research in general practice. Some practicalobstacles are discussed below.Personality barrier. Doctors choose to go into generalpractice because they prefer to develop working relationshipswith patients and their families rather than adopt a technical ordisease orientated focus to the practice of medicine. Some mayperceive research as interfering with this unique relationshipand therefore avoid it. In general practice, professionals work inrelative isolation from one another, in contrast to their hospital-based colleagues who usually work in clinical teams. Thisindependence can result in professional complacency becausethere is no constant, routine requirement for peer review. Somegeneral practitioners have dismissed the science of generalpractice in favour of the art. This view of the discipline is just asdeprived as those who take the opposite stance. Neither viewfairly represents the way general practice works. Neither view,alone, helps patients.

From time to time we hear it suggested that research andgeneral practice are mutually exclusive activities. It seemsobvious that this suggestion is insupportable. There aremany examples of general practitioner researchers. Somehold a special place in the history of medical research, suchas renowned general practitioners William Pickles andRichard Fry. At first glance Pickles’ classic book“Epidemiology in Country Practice”5 is about, as the titlesuggests, epidemiology rather than general practice research.Epidemiology is concerned with patterns of diseaseoccurrence and the factors that influence these patterns.

28 April 2000 150


This must be an important component of general practiceresearch but “Epidemiology in Country Practice” is morethan descriptive epidemiological research. William Picklesprovides an outstanding example of the accommodation ofthe personality of a doctor inclined towards general practicewith the intellect of a research scientist. He was no freak.There have been many like him since and general practicecontinues today to produce innovative researchers.Time and Resource Barriers. Just as there is a time barrier tousing evidence, shortage of time is also one of the major obstaclesto generating useful evidence through general practice research.Kljakovic has provided a poignant, but realistic, account ofpressures on a general practitioner’s time, and how they inhibitgeneral practice research.6 Juggling the demands of routineclinical practice, after-hours service, continuing education,practice management and often wider level managementresponsibilities in the Independent Practitioner Associations(IPAs), leaves little spare time to pursue research interests. Manygeneral practitioners are faced with the additional burden ofhaving to acquire research skills before important researchquestions can be addressed. In the past, medical undergraduatetraining placed little emphasis on the development of researchskills. Unfortunately, present medical school curricula still fail, ingeneral, to address this shortcoming.Acccess to funding and other resource. This is a problemcommon to those working in general practice research in allparts of the world.7-9 Our research is more often ‘applied’than ‘basic’ and the importance of the application,sometimes escapes funding bodies, who may include fewgeneral practice representatives. Universities world wide arestill locked into past priorities.9 They continue to invest inexpensive equipment to support basic science research.While important, too often this is done at the cost of lostopportunities to invest in people whose ideas, insights, andexperience are more likely to produce research that isneeded and acted upon by those involved in providing healthservices. In addition, there is a common misconceptionamong funders that general practice research is equivalent topublic health research and that methods other thanepidemiological ones are therefore suspect. No doubt thesemisconceptions will evaporate with time and perseveranceon our part, but in the meantime they do continue to createbarriers.Technical barriers. There are also technical barriers toundertaking general practice research. The lack of a careerstructure is an important barrier to developing a research-competent workforce. This year, the Health ResearchCouncil will award a Foxley Fellowship to someone involvedin clinical work or health service management.10 For the firsttime, clinicians such as general practitioners will have accessto funding for sabbatical leave from their practices. This willallow the successful applicant to increase his/her knowledgeof academic research while at the same time bringing to aresearch team knowledge of the practical considerations ofclinical practice – which is crucial to successfulimplementation of research. This is an important, excitingdevelopment.

One of the technical barriers to general practitioners’involvement arises from the ‘generality’ of their profession.Many general practitioners seem to think they need to beexperts in all fields of research such as statistics, studydesign, data collection and analysis, before they areprepared to tackle a research project. Ideally researchapprenticeships should be part of the training of all

aspiring general practitioners, in the same way that clinicalapprenticeships are built into current undergraduate andvocational training. Research is too important to generalpractice medicine and the health of our patients, to allowus to abandon this training to other scientists. Even so, itneeds to be openly acknowledged that general practiceresearch often depends on the skills and experiences ofresearchers trained in other disciplines. We should not beashamed to ask for help. The strength of general practiceresearch is that it weaves a diverse set of techniques from awide range of other disciplines into something much morethan the sum of its parts.

General practice research is research arising from aquestion generated from practice, and it is research whoseanswer(s) will provide insights or inform decisions. It iscritically dependent on the context of general practice. It isnot defined by a single methodological approach. Sometimesit is epidemiology, sometimes it is behavioural science. It isnow generally accepted that research methods of the socialsciences must so often merge with the research methods ofthe clinical sciences in general practice research that thequantitative/qualitative dichotomy expounded in the past nolonger exists.11 General practice research encompasseseconomic evaluation, decision analysis, systematic reviewand mixed method approaches to research. Any method thatprovides effective results for general practitioners to use inthe care of their patients is liable for inclusion in thearmoury of general practice research. It is the role of thegeneralist both in clinical practice and in research to developa broad range of skills.

The future development of general practice research ispromising. General practice researchers and cliniciansparticipated in the Foresight Project sponsored by the Ministryof Science, Research and Technology and we predicted a futurein which general practice research would be a key componentof many research programmes.12 In New Zealand, we aremoving into an era where strategies for public research fundingwill encourage interdisciplinary endeavours. Most generalpractice research is at the applied end of the research spectrum,but it will need to be underpinned by relevant research in thebasic and social sciences. A future in which general practiceresearchers form strategic alliances with researchers from otherdisciplines increases the possibilities for relevant research whichwill be more quickly acted upon than in the past. This can onlywork to the benefit of patients.

Correspondence. Professor Murray Tilyard, Department of GeneralPractice, Dunedin School of Medicine, PO Box 913, Dunedin.

1. Carson N, Silagy C. Research I: General Approach. In: Fry J, Yuen N, editors. Principles andpractice of Primary Care and Family Medicine: Asia-Pacific perspectives. Oxford: RadcliffeMedical Press: 1994.

2. Gayle Stevens G. What kind of research in family practice? In: The Intellectual Basis ofFamily Practice. Tuscon: Winter Publishing Company; 1982. p195-206.

3. Bain J. Childhood otalgia: acute otitis media 2. Justification for antibiotic use in generalpractice. BMJ 1990; 300: 1006-7.

4. Williams RL, Chalmers TC, Stange KC et al. Use of antibiotics in preventing recurrentacute otitis media and in treating otitis media with effusion: a meta-analytic attempt toresolve the brouhaha. JAMA 1993; 270: 1344-51.

5. Pickles W. Epidemiology in Country Practice. 3rd Impression. Exeter: The Royal College ofGeneral Practitioners; 1984. p30-8.

6. Kljakovic M. Flourishing research in general practice. NZ Fam Physician 1998; 25: 35-40.7. Research in General Practice: RCGP Information Sheet 12. London: RCGP; 1996.8. Cooper JR. The importance of practice-based primary care research. Colorado Family

Physician 1995; 3rd Quarter: 12-16.9. Kernick D, Stead J, Dixon M. Moving the research agenda to where it matters. BMJ 1999;

319: 206-7.10. Fellowship for Health Sector Professionals. Newsletter: Health Research Council of New

Zealand 1998; 27: 1.11. Schattner P, Wellard R, McGrath B. Choosing a method to answer the question. Aust Fam

Physician 1998; 27: 266-8.12. Foresight 4 Health: General Practice research review: Report of the General Practice

Research Advisory Group; 1998.

151 28 April 2000

the new zealand medical journal...apr 28, 2000 · (e.g. the ‘under sixes’) and waiting lists...

Documents