Facilitation visits as a method for implementing change in general practice.

What lies beneath the label?

PhD thesis

Tina Drud Due

U N I V E R S I T Y O F C O P E N H A G E N F A C U L T Y O F H E A L T H A N D M E D I C A L S C I E N C E S

This thesis has been submitted to the Graduate School at the Faculty of Health and Medical Sciences, University of Copenhagen March 2016

Facilitation visits as a method for implementing change in general practice.

What lies beneath the label?

PhD thesis

Tina Drud Due, cand.scient.san.publ

The Research Unit for General Practice and Section of General Practice, Department of Public Health, University of Copenhagen, Denmark

Name of Institute: Department of Public Health Name of department: The Research Unit for General Practice and Section of General Practice,

Department of Public Health, University of Copenhagen, Denmark Author: Tina Drud Due (cand.scient.san.publ) Title: Facilitation visits as a method for implementing change in general practice.

What lies beneath the label? Supervisors: Professor Frans Boch Waldorff (MD, PhD) The Research Unit for General Practice and Section of General Practice,

Department of Public Health, University of Copenhagen, Denmark. Research Unit for General Practice, Department of Public Health, University of Southern, Denmark

Thorkil Thorsen (Sociologist, PhD) The Research Unit for General Practice and Section of General Practice,

Department of Public Health, University of Copenhagen, Denmark Marius Brostrøm Kousgaard (cand.scient.pol, PhD) The Research Unit for General Practice and Section of General Practice,

Department of Public Health, University of Copenhagen, Denmark Submitted: 14 March 2016 Public defence: 10 June 2016 Assessment committee: Professor Flemming Bro (MD, PhD) Research Unit for General Practice, Department of Public Health University of Aarhus, Denmark Professor Kate Seers (DSc) Royal College of Nursing Research Institute, Division of Health Sciences,

Warwick Medical School, University of Warwick, UK Professor Jakob Kragstrup (MD, PhD)

Section of General Practice, Department of Public Health University of Copenhagen, Denmark (Chair of the committee) Financing: The Danish Research Foundation for General Practice The Health Foundation The Research Foundation for Primary Care in the Capital Region of Denmark

This thesis is based on the following papers

Paper 1 Due TD, Thorsen T, Waldorff FB, Kousgaard MB. Role enactment of facilitation in primary care – a qualitative study. (Manuscript)

Paper 2 Due TD, Kousgaard MB, Waldorff FB, Thorsen T. Influences of peer facilitation in general practice – a qualitative study. (Manuscript)

Paper 3 Due TD, Thorsen T, Kousgaard MB, Siersma VD, Waldorff FB. The effectiveness of a semi-tailored facilitator-based intervention to optimise chronic care management in general practice: a stepped-wedge randomised controlled trial. BMC family practice. 2014;15:65.

Contents Acknowledgements .............................................................................................................................. 1

Summary .............................................................................................................................................. 3

Dansk resumé (Danish summary) ........................................................................................................ 5

Introduction .......................................................................................................................................... 7

Ambiguity in definitions of facilitation ............................................................................................ 7

The broad spectrum of facilitation interventions ............................................................................. 8

Ambiguity due to related concepts ................................................................................................... 9

Current knowledge about facilitation effectiveness and practice ................................................... 10

The purpose of the thesis ................................................................................................................... 12

The context and content of the facilitation intervention .................................................................... 14

The intervention context ................................................................................................................. 14

The intervention ............................................................................................................................. 15

Method ............................................................................................................................................... 19

The qualitative studies .................................................................................................................... 19

The randomised controlled trial ..................................................................................................... 30

Findings .............................................................................................................................................. 33

Paper 1. Role enactment of facilitation in primary care – a qualitative study ............................... 33

Additional findings of facilitation enactment not presented in paper 1 ......................................... 35

Paper 2. Influences of peer facilitation in general practice – a qualitative study ........................... 38

Paper 3. The effectiveness of a semi-tailored facilitator-based intervention to optimise chronic care management in general practice: a stepped-wedge randomised controlled trial .................... 39

Discussion .......................................................................................................................................... 40

Summary of findings in relation to current facilitation research.................................................... 40

Strengths and limitations of my studies ......................................................................................... 45

A need for discussion and clarification of the facilitation concept ................................................ 48

Challenges of complexity in the evaluation of facilitation interventions ....................................... 50

Implications for research and practice ............................................................................................... 62

Conclusion ......................................................................................................................................... 65

References .......................................................................................................................................... 67

Appendices ......................................................................................................................................... 75

Appendix 1: Selection of definitions of facilitation from the literature ......................................... 76

Appendix 2: Description of the Data Capture Module................................................................... 77

Appendix 3: Paper 1 ....................................................................................................................... 79

Appendix 4: Paper 2 ....................................................................................................................... 97

Appendix 5: Paper 3 ..................................................................................................................... 113

Abbreviations

GP: General practitioner

CDMP: Chronic disease management programmes

DCM: Data Capture Module

COPD: Chronic obstructive pulmonary disease

ICPC diagnosis coding: International classification of primary care diagnosis coding

1

Acknowledgements First of all, I would like thank my team of supervisors. Thank you for giving me the opportunity to undertake this thesis, and the liberty and trust to follow my interests. Thank you for always being available when I needed advice and to discuss my findings. Thank you for creating and maintaining a good and supportive atmosphere.

A special thanks to Frans Boch Waldorff for your positive spirit, for believing in the progress and success of the thesis, and for remembering to reassure me of this. Thank you for introducing me to the field of randomised controlled trials, and for always looking out for my interests and thinking about the bigger picture. A special thanks to Thorkil Thorsen for the numerous interesting discussions and for the ability to always ask questions that fostered my reflections and challenged my understanding. Thank you for always making it a priority to help me, and for being generous with your time whenever I needed it. A special thanks to Marius Brostrøm Kousgaard for your sharpness and thoroughness in the discussions and revisions of the manuscripts and for challenging my understanding. I would also like to thank Volkert Dirk Siersma for statistical contributions and advice, Willy Karlslund for data management of the project, and my student assistant, Anne Katrine Bjørkholt Sørensen, for exhibiting great flexibility in both time and tasks to make data collection a much more smooth process.

Grateful thanks to Gritt Overbeck for all our talks – the scientific discussions about implementation science and our studies, and the non-scientific talks brightening up the day. Thank you for lending an ear during the ups and downs, and thanks for the stories we shared.

Thanks to my colleagues at the Research Unit for General Practice in Copenhagen. I have appreciated the supportive atmosphere and being in a research environment with such varied professional and methodological perspectives. Thanks to all those who supported and encouraged me, especially in the final phase of the thesis.

I would also like to give warm thanks to the participants in my studies. Thank you to the participating practices and facilitators for letting me observe your facilitation visits and for setting aside time for the interviews. Without you this thesis would not have been possible. Thanks to the project initiators and project managers for letting me use the intervention for my thesis, and for the interviews and intervention documents.

I have great appreciation for the people in the Implementation Research Programme at the School of Healthcare Sciences at Bangor University in Wales for kindly hosting me for a research stay during my PhD studies. Thank you to Jo Rycroft-Malone and Chris Burton for having me in your research programme, for the discussions about my studies and the field of facilitation, and for introducing me to new research perspectives. Also thanks to Lynne Williams, Patricia Masterson Algar and Heledd Owen for the discussions about our studies and common interests and challenges. Thanks to everybody in the programme for the warmth and kindness and for making my time with you so enjoyable.

2

Sincere thanks are due to Brendan McCormack and Gill Harvey for taking the time to discuss my findings about facilitation enactment and the field of facilitation.

Thank you to those who have funded my PhD. The Danish Research Foundation for General Practice, The Health Foundation and The Research Foundation for Primary Care in the Capital Region of Denmark. Also a warm thanks to Susanne Reventlow the Research Director of the Research Unit for General Practice in Copenhagen for always having my back financially. It means a lot that I did not have to worry in times of trouble.

Not least I would like to thank my friends and family. I wish to thank my friends for being there, for trying to comprehend what it is that I am doing, and for understanding when I went into hiding during the last period of the thesis. A heartfelt thanks to my mother and father. Thank you for always supporting me and my family. Thank you for showing an interest in the thesis and for always being there to take care of my daughters when we needed help. Most of all, to my husband Lars, thank you so much for your love, support, encouragement, and great patience. Thank you for keeping me in my senses, for being there in the tough times, keeping me grounded, and for making me laugh. Thank you for having shown an interest in, listened to, and discussed my work with me for all these years, and for pulling the extra weight these last six months, taking charge of everything on the home front. Finally, thanks to my wonderful daughters, Josefine and Filippa, for your unconditional love, and your smiles and hugs after a long day’s work.

3

Summary Introduction Various strategies are used to support guideline implementation and practice development in health care. One of these, facilitation, is an active and increasingly widespread strategy. It is a multifaceted intervention, where an external person (most often a health care professional) visits and supports a process of change. However, the field of facilitation is complex and conceptual clarity is lacking. There is no clear and consistent operational definition of facilitation and the content of facilitation interventions varies considerably. Both within and between studies facilitators are described as having multiple roles and are responsible for conducting a variety of different activities. At the same time a facilitation approach that is tailored to the needs of the individual practice or team is perceived as vital.

To contribute to a differentiated and more detailed picture of what may lie beneath the label of facilitation the idea of this thesis was to explore a Danish facilitation intervention in general practice from different perspectives. The three specific research questions were:

1. How is facilitation enacted during facilitation visits? 2. How does facilitation influence the visited practices and how do they value the facilitation? 3. What is the effectiveness of the facilitation intervention?

Methods The facilitation intervention studied in this thesis took place in general practice in the Capital Region of Denmark with the purpose of supporting the implementation of chronic disease management programmes for type-2-diabetes and chronic obstructive pulmonary disease in general practice. Fourteen general practitioners were hired as facilitators and all practices in the region were offered up to three visits of one hour each.

The three research questions were examined by means of two qualitative research studies and a randomised controlled trial. The qualitative data consisted of observations of 30 facilitation visits in 13 practice settings1, focus groups with facilitators, and interviews with project initiators, project managers, facilitators, and the observed practices. I based the study on hermeneutic principles and conducted a thematic analysis. In the randomised controlled trial, 189 general practices were included and of these 96 practices were allocated to receive facilitation visits in 2011 (intervention group) and 93 practices to receive facilitation visits in 2012 after the trial had ended (control group). Data were based on Danish national registers and practice questionnaires.

Findings Facilitation in this intervention was mainly enacted through four facilitator roles: the teacher, the super user, the peer, and the process manager. Facilitators provided factual knowledge either through presentations or a hands-on approach, passed on experience-based knowledge, and ensured 1 It is labelled practice setting because 4 of the 13 were visits involving collaborating practices. Hence, a total of 18 practices were represented in the data.

4

task definition, delegation, and deadlines. In general, facilitators did not enact a coaching-based role to enable collective reflections and internal discussions on practice organisation and implementation during the visits. The study also identified variations in the facilitation visits due to both practice and inter–facilitator variations. In general practices reported that the facilitation visits gave them increased knowledge, awareness, and skills. They also influenced practice motivation by altering the perception of usefulness and manageability of the change areas. In addition to the content of facilitation, an equally important influence was the provision of protected time, an occasion to change, definition and delegation of tasks, and a sense of deadline due to recurrent visits. The intervention’s design, contextual conditions, and the fact that the facilitators were also general practitioners influenced both the facilitation enactment and how practices valued and were influenced by the visits.

The randomised controlled trial showed mixed results with no differences between allocation groups regarding the primary outcome (annual chronic disease check-ups), but differences regarding some of the secondary outcomes (diagnosis coding and patient stratification). Further, there were improvements in both allocations groups in most outcome measures. Hence, the intervention did not add substantially to a change in chronic care management under the given contextual conditions.

Conclusion The thesis demonstrates that facilitation is a complex phenomenon on both a conceptual level, in practice, and as a research object. It is a diverse and multifaceted concept with limited conceptual clarity and much variation between interventions. Facilitation is enacted in a variety of roles, and some facilitator roles are more likely to be enacted than others depending on the design, content, and context of the intervention as well as on the professional background of the facilitators. Further, facilitation influences the change process in diverse ways, where some elements of change are due to the enacted facilitation, while others are more likely to be caused by the visits providing protected time and attention. In general, contextual conditions appear influential regarding facilitation enactment, practices’ experiences of valued facilitation elements, facilitation process, and outcomes. This thesis also demonstrates that there is a need for conceptual discussion and clarification around facilitation. It points to methodological challenges in the study of facilitation, and limitations in the transferability and comparability of the results of facilitation studies to other settings. There is also an issue around the usefulness of systematic reviews, and particularly meta-analyses of facilitation. These issues need to be addressed by both researchers and those who adopt facilitation as an implementation method.

5

Dansk resumé (Danish summary) Introduktion Der anvendes en række strategier til at understøtte implementering af retningslinjer og praksisudvikling inden for sundhedsvæsenet, som eksempelvis bestemmelser, finansielle incitamenter og informationsformidling. I modsætning hertil er der facilitering, som er en mere aktiv og i stigende grad udbredt strategi. Facilitering er en multifacetteret intervention, hvor en ekstern person (oftest med en sundhedsfaglig baggrund) besøger praksis eller teamet og støtter dem gennem en udviklingsproces.

Faciliteringsfeltet er komplekst, og der lader til at mangle konceptuel klarhed. Der er ingen klar og konsistent definition af facilitering, og der er markante variationer i indholdet mellem forskellige facilitatorinterventioner. Facilitatorer beskrives at have mange forskellige roller og mangeartede aktiviteter er beskrevet både inden for enkelte interventioner og på tværs af interventioner. Desuden fremhæves det som afgørende at tilgangen skræddersyes til den enkelte praksis eller team.

For at bidrage til et mere nuanceret billede af hvad der kan være under faciliteringsetiketten var ideen med denne afhandling at undersøge en dansk facilitatorintervention i almen praksis fra forskellige perspektiver. Dette blev gjort via følgende tre forskningsspørgsmål:

1. Hvordan udmøntes facilitering i facilitatorbesøgene? 2. Hvordan påvirkes de besøgte praksis og hvordan vurderer de facilitatorbesøgene? 3. Hvad er effekten af denne danske facilitatorintervention i almen praksis?

Materiale og metode Facilitatorinterventionen blev udført i Region Hovedstaden og havde til formål at understøtte implementeringen af forløbsprogrammer for type-2-diabetes og kronisk obstruktiv lungesygdom i almen praksis. Fjorten praktiserende læger var ansat som facilitatorer på konsulent basis og alle praksis i regionen blev tilbudt tre facilitatorbesøg af en times varighed.

De tre forskningsspørgsmål blev undersøgt via to kvalitative studier og et randomiseret kontrolleret forsøg. Datakilderne til de kvalitative studier var observationer af 30 facilitatorbesøg i 13 forskellige praksislokationer (i alt 18 praksis, da nogle samarbejdspraksis havde fælles besøg), fokusgrupper med facilitatorerne og interview med facilitatorer, de besøgte praksis, projektudviklerne og projektledelsen. De kvalitative studier tog udgangspunkt i en hermeneutisk tilgang og er analyseret via en tematisk analyse. I det randomiserede kontrollerede forsøg blev 189 praksis inkluderet. Heraf blev 96 praksis allokeret til interventionsgruppen og 93 til kontrolgruppen, som efter forsøgets afslutning også modtog facilitatorbesøg. Data var baseret på registre og spørgeskemaer til praksis.

Fund Facilitering blev i denne intervention primært udmøntet i form af fire facilitatorroller: underviseren, superbrugeren, kollegaen og projektlederen. Dette fremgik ved at facilitatorerne videregav faktuel viden enten via præsentationer eller via mere aktivt deltagelse ved klinikkens computer, videregav

6

erfaringsbaseret viden samt søgte at sikre at opgaver blev defineret, uddelegeret og udført. Facilitatorerne udmøntede generelt ikke en også tiltænkt coaching-baseret facilitatorrolle, hvor de under besøgene understøttede interne diskussioner og refleksioner over den besøgte praksis’ organisering og implementering. Studiet identificerede desuden både praksis- og faciliatorafhængige variationer mellem besøgene samt at faciliteringsprocessen var kendetegnet ved både at have praksis- og facilitatorstyrede elementer. De besøgte praksis oplevede generelt at besøgene gav dem øget viden, opmærksomhed og nye kompetencer, samt ændrede deres oplevelse af brugbarheden og håndterbarheden af de områder der skulle ændres. Ud over at de besøgte praksis blev påvirket af besøgenes indhold, virkede det ligeså vigtig at besøgene bidrog med at være en anledning til at fokusere på de valgte emner, at de gav beskyttet tid, sikrede definering og uddelegering af opgaver, samt at de tilbagevendende besøg gav en oplevelse af have en deadline. Interventionens design, kontekst og det at facilitatorerne var kolleger tydede på at have påvirket både måden hvorpå faciliteringen blev udmøntet, og hvordan besøgene påvirkede de besøgte praksis samt disses vurdering heraf.

Det randomiserede kontrollerede forsøg viste blandede resultater. Der var ingen forskel mellem interventionsgruppen og kontrolgruppen ved det primære effektmål (årskontroller), men der var signifikante forskelle ved nogle af de sekundære effektmål (diagnosekodning og stratificering). Der skete desuden stigninger fra baseline til opfølgningstidspunktet for begge grupper ved de fleste effektmål. Interventionen virkede således kun til i begrænset omgang at medføre yderligere ændringer i indsatsen for patienter med kronisk sygdom i den givne kontekst.

Konklusion Afhandlingen viser at facilitering er et komplekst fænomen både på et konceptuelt plan, i praksis og som forskningsobjekt. Det er et multifacetteret koncept med begrænset konceptuel klarhed og heterogenitet på tværs af interventioner. Facilitering udmøntes som en vifte af roller, og nogle facilitatorroller vil sandsynligt oftere blive udmøntet end andre afhængigt af interventions design og kontekst og facilitatorernes professionelle baggrund. Facilitering påvirker desuden forandringsprocesser i praksis på en række forskellige måder, hvor noget hænger sammen med den facilitering der sker på selve facilitatorbesøget (eksempelvis øget viden og motivation) og andet hænger sammen med at afholdelsen af besøgene giver praksis en anledning til at igangsætte ændringer, beskyttet tid og øget fokus. Kontekstuelle betingelser tyder generelt på at have indvirkning på facilitering ved både udmøntningen, praksis’ oplevelser, processen og udbyttet. Afhandlingen viser desuden, at der er behov for diskussion og afklaring af faciliteringskonceptets indhold, at der er metodemæssige udfordringer i hvordan facilitatorinterventioner kan studeres samt at der er begrænsninger i overførbarheden mellem forskellige kontekster. Især sidstnævnte begrænser anvendeligheden af systematisk reviews og især metaanalyser inden for facilitering. Dette er forhold som både forskere og de der anvender facilitatorinterventioner som metode til implementering og praksisudvikling bør forholde sig til.

7

Introduction Various strategies are used to support guideline implementation and practice development in health care e.g. regulations, financial incentives, information dissemination, lectures, audit and feedback, and reminder systems [1]. Compared to these, facilitation is a more active and multifaceted strategy, where an external person (most often a health care professional) visits and supports a process of change2. Facilitation in health care is described in quality improvement and guideline implementation predominately among hospital nurses and in primary care. In primary care, it is increasingly used under the term “practice facilitation”. In this thesis I focus on facilitation in general practice, but the findings and discussions are relevant for facilitation within other health care settings as well. Facilitation within primary care is said to originate from the implementation of the Oxford Model of preventive health, studied by Fullard et al. in the early 1980s [3, 4]. Facilitation in health care also draws upon the field of education, counselling, and organisational learning [2]. As a method of managing change, the use of facilitation is increasingly widespread and has an international reach [2, 5-11]. However, the conceptual clarity of the method appears to be lacking [5, 12, 13] and there are differences in the definitions of facilitation and variations in the content of facilitation interventions. Furthermore, the differences between facilitation and related practice improvement concepts are indistinct.

The following description of the field of facilitation and the diversity within it is based on a wide selection of facilitation studies3 [3-5, 7, 9, 11-72]. My intention is to introduce current knowledge on facilitation and to illustrate the breadth and dimensions of the facilitation concept. My selection of facilitation studies is not exhaustive, but the included studies are sufficiently comprehensive to attain a profound understanding of the field4.

Ambiguity in definitions of facilitation There is no overall and agreed upon definition of facilitation and most studies do not explicitly define it. In current definitions, the content and comprehensiveness vary. Further, the concept entails the process of facilitation as well as the facilitator as a person and as a distinct role. Thus, the research literature describes elements related to facilitators: their skills, characteristics, and roles; and elements related to the facilitation process: its purposes, activities, and outcomes. Examples of definitions and descriptions of facilitation are presented in Appendix 1. In some of the research literature (primarily regarding practice facilitation), a common denominator appears to be that facilitation entails creating awareness of a need for change, often by audit and feedback; assistance in consensus building and goal setting; and support in management of quality improvement activities tailored to local needs. Another trend in the literature explicitly focus on the facilitator

2 In some parts of the facilitation literature, there is a division between internal and external facilitators [2]. This thesis focuses on external facilitation. 3 Unless specific references are indicated, then the descriptions are based on all the included studies. 4 Studies have been included if the authors have labelled it facilitation or called the visitors facilitators. However, some studies not included might deal with interventions with similar content, but where other labels have been applied e.g. educational outreach visits. Further, some studies of facilitation interventions might have been missed in the literature search.

8

enabling change by supporting internal discussion, relationship building, interpersonal processes, critical reflection, capacity building, and empowerment [2, 57, 59, 73].

To integrate the different understandings of facilitation and facilitation activities, the Promoting Action on Research Implementation in Health Services (PARIHS) framework described facilitation as a continuum [59, 60, 74]. At one end of the continuum is a goal oriented and task driven approach focusing on guideline implementation through practical and technical help. At the other end of the continuum a more holistic approach emerges where the facilitator supports practice development through a process of critical reflection and emancipatory and transformative approaches focussing on empowerment, personal development, and contextual issues affecting implementation. According to the authors of the PARIHS framework, a facilitator should use diverse approaches on the continuum when appropriate [2, 59, 74]. The authors have recently elaborated on this perspective and now describe three levels of facilitator: novice, experienced, and expert. The novice facilitator does not have the ability to perform all types of facilitation and needs supervision from more experienced facilitators [2]. Different levels of facilitator experience do not appear to be mentioned in other parts of the research literature on facilitation in health care.

The broad spectrum of facilitation interventions It is evident from current facilitation studies that there are large differences between facilitation interventions. These include differences as to which areas the interventions aim to improve, the duration and number of visits in the interventions, the facilitators’ professional background, and the facilitation activities.

Facilitation is used to implement guidelines for preventive services (e.g. cardiovascular disease, diabetes, and cancer screening), for prescribing of medicines, palliative care, and in chronic care management [4, 7, 25, 26, 28-33, 36-39, 41-44, 46-53, 63, 64, 66, 68, 70]. It is also used for organisational issues e.g. in the use of the maturity matrix, shared care, implementation of professional and practice development plans, as well as learning practice programmes [14, 16-18, 27, 34, 35, 41, 50, 75]. Finally, it is used within practice development in nursing focusing on transformational practice culture development, and empowerment [73]. Across the selected studies the number of facilitation visits varied from 1 to 33 [25, 43] and the time period for a facilitation intervention in each health care setting varied from 3 to 24 months [26, 56]. To illustrate the variations further, two of the interventions in the included studies provided one visit a month for a year [39, 46]; one provided 18 visits in total, with one visit twice a month [49]; and one provided 12 visits in 12 weeks [28]. Yet other interventions provided a couple of visits within three months, or with a year between visits [16, 56]. In several studies, the length of the intervention in the individual settings was not stated. Facilitators were most often nurses or practice assistants, occasionally general practitioners (GPs), and they rarely seemed to have a background in social science or organisational change [6, 9, 16, 22, 25, 28, 29, 35, 56]. Several of the selected facilitation studies did not report the number of facilitators in the intervention, and among the few that did, the mean number was four facilitators (in the range 1-7; except for a study with 16 and a study with 171 facilitators).

9

The combination of facilitation as a multifaceted approach and diversity between interventions results in a considerable latitude in facilitation activities both within and between interventions. In the studies I selected for my research, these activities included: raising awareness of the need for change, audit and feedback, consensus building, goal setting, planning and allocation of responsibility, plan-do-study-act circles, helping practices develop reminder systems and data collections systems for quality improvement, sharing knowledge and information, presentations, providing education and training for staff in the understanding and use of data, presentation of tools, providing advice, technical expertise to support use of data in the quality improvement process, low-cost technical support, experiences of the facilitator, facilitating discussions, hands-on help, promoting effective communication patterns, create a culture receptive to change, motivation, project management, on-going monitoring and evaluation, maintaining continuous improvement capability, assistance with research and gathering research knowledge, and cross pollination of good ideas between practices.

Though several definitions, purposes, and types of activity are mentioned in the facilitation research literature, they are most often superficially described. Thus, it is rarely explicit how these activities are supposed to be or actually are conducted, nor is it stated what the preferred and actual balance is between the different activities and facilitation approaches. Further, although there are variations between both the definition and the intended content of specific interventions, it is difficult to assess which of these differences are of significance to the enacted facilitation. To add to the complexity, there is an increasing emphasis of a need for a tailored approach, whereby both the facilitation approach and the application of guidelines are tailored to the local context and the specific needs of the individual practice or team [5, 13, 24, 26, 28, 63, 76]. According to a systematic review and meta-analysis of practice facilitation, 74% of studies reported that the practice facilitators tailored the intervention to the needs of the practice [6]. However, it is not clarified in the literature what a tailored approach more specifically entails nor how it should be managed or ensured.

Ambiguity due to related concepts A conceptual ambiguity is also seen in the overlap between related concepts like “practice facilitation”, “outreach facilitation”, “practice coaching”, “academic detailing”, “educational outreach visits” and ”practice enhancement assistance”. This does nothing to improve the distinctiveness of the facilitation concept. Some of these terms are occasionally described as synonyms e.g. “practice facilitation” and “practice coaching” [9, 58] or “academic detailing” and “educational outreach visits” [56, 77]. However, there does not seem to be clarity or agreement on whether or how, for example, practice facilitation and educational outreach visits differ. Across studies, there are differences between interventions using the same concept label and similarities in interventions using different concept labels. Further, some of the activities e.g. audit and feedback and knowledge provision are mentioned in both types of intervention. These indistinct borders between the concepts are also evident in systematic reviews and meta-analyses in the field. In 2007, a Cochrane review was conducted of educational outreach visits5 and a systematic review and meta-

5 The review includes studies in health care, but 53 of the 69 studies are with primary care physicians or teams practicing in community settings.

10

analysis of practice facilitation6 was conducted in 2012. In these two papers there is a partial overlap in both the definitions and in some of the included studies, but neither of the papers mentioned how the concepts may relate or how they may differ. In the meta-analysis of practice facilitation, facilitation was defined as “a multifaceted approach that involves skilled individuals who enable others, through a range of intervention components and approaches, to address the challenges in implementing evidence-based care guidelines within the primary care setting” [6]. The Cochrane review of educational outreach visits described this type of intervention as “trained people visit clinicians where they practice and provide them with information to change how they practice. The information given may include feedback about their performance, or may be based on overcoming obstacles to change” [78]. To further illustrate the ambiguity, the Cochrane review of educational outreach visits included 69 studies, while the meta-analysis of practice facilitation included 23 studies, and they had six studies in common.

In a paper by Mold et al, academic detailing and practice facilitation were explicitly described as being different concepts [49]. Contrary to this Van Hoof et al [62] described the concepts as overlapping, but developed side by side in different journals and by different researchers, and they argued for an integration of the two concepts perhaps on a continuum [62]. A potential difference between the concepts could be that educational outreach visits often seem to be of shorter duration and more focused on knowledge provision; whereas practice facilitation often seems focused on supporting the change process through several visits and by the use of more multifaceted tools [61]. Additionally, some distinguish the concepts by the logic of change. In this thinking, educational outreach visits and academic detailing are described as being based on more of a marketing logic by using persuasion and telling the practices about the correct approach; whereas facilitation is described as helping, enabling, and developing a learning process [2, 13, 79]. Finally, there are also several other related concepts described within health care e.g. “change agents”, ”knowledge brokers”, “linking agents”, “intermediaries”, “champions” and “opinion leaders” [13, 80, 81]. The differences between these concepts are not well defined either and there appear to be similarities with the role of facilitators.

Current knowledge about facilitation effectiveness and practice Several randomised controlled trials (RCTs) have been conducted on facilitation within primary care, while none appear to have been conducted within nursing. Some of the RCTs have shown that facilitation interventions have contributed to changes [31, 32, 43, 44, 47, 48, 51, 54, 56], while others have found mixed results [33, 34, 36-38, 41, 42, 46, 49, 50, 52, 53, 55, 66] or have found no effect [39, 64]. The previously mentioned meta-analysis of practice facilitation concluded that there was a moderate, but robust effect on guideline adoption [6]. The Cochrane review of educational outreach visits concluded that there was a small, but consistent effect on prescribing. In other areas the effect was small to modest, but more variable [78]. However, given the diversity in the field of facilitation and educational outreach visits both meta-analyses included heterogeneous studies.

6 In the rest of the thesis “the systematic review and meta-analysis of practice facilitation” is described as “the meta-analysis of practice facilitation”. Further, the Cochrane review of educational outreach visits also included a meta-analysis.

11

Most RCTs in facilitation are of interventions that aimed at implementing concrete, disease specific guidelines and preventive services. Due to the diversity of the diseases and the services in focus, multiple outcomes were measured across the different studies, for instance cancer screening, immunisation, assessment of risk factors, and performed preventive services [31-33, 36-39, 42, 43, 46-48, 52, 56, 64]. Other RCTs have assessed organisational measures [34, 41, 44, 50, 51], or both organisational and preventive service outcome measures [32, 41, 42, 49, 50, 53, 66]. The data sources in these studies were either patient or practice questionnaires or journal audits, and the number of included practices was often quite low (in the selected studies it ranges from 16 to 264, with a mean number of around 65 practices)

Qualitative studies have explored how facilitation is performed [4, 7, 9, 12, 19-22, 27-29, 68, 71]7, challenges and barriers in facilitation and practice change [16, 18, 22, 23, 26, 29, 30, 71], how practices experience the support from facilitation visits [16, 17, 23, 30, 71], and practice experienced outcomes and satisfaction with the visits [16, 17, 19, 25]. The data in qualitative studies mostly consisted of interviews with facilitators [4, 7, 26], practices [15-17, 23, 69], or both [9, 19, 25, 29, 30, 71]. A few studies have gathered facilitators from different facilitation interventions for individual interviews, focus groups, or symposiums [12, 20, 21]. Observations and audio recordings of facilitation visits have been used, but are less common [16, 17, 19, 25, 27, 28, 30, 71]. Other studies have used questionnaires, facilitators’ registrations, and visit reports to explore practiced facilitation by focusing on facilitators’ use of time, conducted activities, and practices’ satisfaction [15, 24, 65].

7 Most of these studies were published after my theses was initiated

12

The purpose of the thesis Based on the descriptions above, the field of facilitation is complex. Facilitation is a multifaceted intervention, there is no clear and consistent operational definition, studies convey a range of facilitator roles and facilitation activities both within and between interventions, and an individual tailored approach is perceived to be vital. Additionally, there is great heterogeneity between interventions and although there are common features in definitions and intended activities across studies, these activities are often superficially described. Combined, this entails a profound possibility for variation between intended and enacted facilitation both within and across studies. It is therefore important to explore how facilitation is actually enacted and how it influences practices in specific interventions in order to understand what lies beneath the label of facilitation.

To get a more comprehensive understanding of facilitation, and to aid the operationalisation of the concept, several researchers have recommended the study of facilitator roles and facilitation activities, communication processes, and how facilitators enable change in practice [5, 9, 19, 20, 27]. There is some existing knowledge of practiced facilitator roles and facilitation activities and practices’ experiences of facilitation visits. However, because these current descriptions are fairly broad, there is still a need for more detailed knowledge of facilitation enactment e.g. how activities are performed and the balance between them. Further, observations used in combination with interviews could ensure a more nuanced picture of the facilitation enactment. Additionally, previous studies have provided some insights into practices’ experiences with facilitation. However, an in-depth exploration of practices’ entire facilitation processes explored through practice interviews and observations could enhance the understanding of how practices are influenced by and value facilitation visits, and also address potentially influential circumstances.

As for the effectiveness of facilitation interventions, various RCTs have been conducted, but only a few assessed the implementation of more complex guidelines. Further, with facilitation interventions being adopted by health planning authorities as implementation methods, effectiveness studies of facilitation interventions in real-life settings are recommended [64].

Therefore, my idea in this thesis is to explore facilitation visits as a method for guideline implementation and practice development in general practice by studying a Danish facilitation intervention in a real-life setting from different perspectives. I will contribute to a differentiated and more detailed picture of facilitation, and thereby provide inputs to a conceptual discussion and into potential areas for improvement that might strengthen facilitation as an implementation method. This overall idea was translated into the following three research questions:

1. How is facilitation enacted during facilitation visits? 2. How does facilitation influence the visited practices and how do they value the facilitation? 3. What is the effectiveness of the facilitation intervention?

13

I explored these three questions by means of two qualitative studies and one RCT8. This thesis was initiated as part of an evaluation of the facilitation intervention, which my supervisors and I were hired to conduct by the Capital Region of Denmark. As researchers, we were not involved in any decisions about intervention content or implementation, and we have designed and conducted the research studies without interference from the intervention developers.

8 In this thesis, the studies are presented and referenced in the order corresponding to the three research questions. However, the paper on the RCT was written and published first.

14

The context and content of the facilitation intervention The intervention context The Danish health care system is organised under a decentralised administrative structure with five regions and 98 municipalities. The regions are in charge of hospitals and GPs [82]. Danish health care is primarily tax financed and GPs serve as the primary care providers and gatekeepers for patients’ referral to specialists and hospitals. GPs are private entrepreneurs mainly financed through the tax financed health care reimbursement scheme. General practice service provision is regulated via collective agreements between the Danish Regions and the Organisation of General Practitioners [82, 83]

With a population of 5.4 million, Denmark has approximately 3600 GPs; hence, a GP has around 1500 patients. Partnership practices make up forty per cent of GP practices, where most commonly 2-3 GPs co-own a practice. The rest of the practices are solo-practices although some of these GPs collaborate and share facilities or practice staff [84]. Practice staff consist of secretaries and nurses, with nurses increasingly performing some of the consultations and chronic disease check-ups. However, only GPs are permitted to perform annual chronic disease check-ups.

About a third of the Danish population suffers from at least one chronic disease and allegedly use 70-80 % of the resources in the Danish health care sector [85]. Chronic disease management programmes (CDMP) based on the Chronic Care Model [86, 87] have been developed in all five regions of Denmark [88, 89]. As guidelines these programmes outline evidence based treatment and a systematic approach to chronic care including a division of tasks between GPs, hospitals, and municipalities for a given chronic disease. The programmes describe the GP’s role as the coordinator of care and outline a systematic, proactive approach with population based patient registration, annual chronic disease check-ups, and stratification of patients into three levels according to risk of complications and the complexity and state of the disease [90, 91]. It is mandatory for all five regions to develop CDMP, and it is mandatory for the hospitals and municipalities to adhere to them. However, it is not mandatory for GPs. Several simultaneous strategies have been carried out to aid the implementation of the CDMP and to improve the care of patients with chronic disease. For instance, the facilitation intervention in this study was one of 18 projects in the Capital Region of Denmark that were granted money by a foundation from the Ministry of Health to improve chronic care management.

General practice in Denmark is fully computerised regarding patient records, prescriptions, referrals, lab results, and hospital discharge letters [82]. Independently of the facilitation intervention studied in this thesis, but shortly after it was initiated, it became mandatory for all Danish GPs to sign-up to the Data Capture Module (DCM) within a two-year period (before April 2013). The DCM is a software program for quality development based on patient data, where data is automatically collected and aggregated from electronic patient records (Appendix 2) [92].

15

The intervention The facilitation intervention that I have studied in this thesis was developed by the Capital Region of Denmark (in the division that administers primary care) and the Regional Unit for Quality Development and Continuing Education in General Practice, and it was carried out from January 2011 to December 2012. The purpose of the intervention was to support the implementation of CDMP for type-2-diabetes and chronic obstructive pulmonary disease (COPD) in general practice9.

The available documents10 contained limited information on the intended facilitation, hence I interviewed the project initiators (a regional leader and the leader from the Regional Unit for Quality Development and Continuing Education in General Practice) and the regional project managers for further insight11. It is explicitly stated when the descriptions are based on these interviews.

The facilitators and the facilitation educational programme Fourteen GPs were hired on a consultancy basis to act as facilitators12. They took part in an educational programme consisting of a one-weekend seminar and 10 three-hour meetings during a four month period. Here they were updated on the central elements of the CDMP for COPD and diabetes and the DCM, and they were introduced to various implementation and facilitation tools such as Plan-Do-Study-Act (PDSA) circles13[93] and the brown paper method14[94]. The facilitators also participated in workgroups where they developed the tools they expected to use during the facilitation visits. There was also a pilot phase where they gained experience of conducting facilitation visits in a few practices. Network meetings were held approximately every third month throughout the intervention period. Here the facilitators discussed their experiences through different exercises, had further presentations with professional updates and more information on the facilitator role, and adjusted their tools.

In the interviews, the project managers described how they had not sought to directly control or standardise the facilitators as to how they conducted their visits i.e. there were no templates or requirements for the structure of the visits. Project managers had introduced the facilitators to

9 In the rest of the thesis type-2-diabetes will described as diabetes 10 These documents entailed: brief descriptions of the purpose, intervention content, the role of the facilitator and the success criteria; a programme with dates and headlines of the facilitators’ education; and PowerPoint slides from the facilitators’ education concerning the act of facilitation. 11 The interviews focused on the project initiators’ and project managers’ description of the intervention purpose and content, the choice of using facilitation as an implementation strategy, the intervention development, their understanding of facilitation and the facilitator role, the education of the facilitators, expected changes, and intervention flexibility. The interviews were conducted about six month after the first facilitation visit. Hence, the interviews do not necessarily illustrate the interviewees’ initial intentions, but their recollection thereof, which may have been influenced by their experience with the implemented intervention. 12 The intervention also included two organisational facilitators. One was a GP with an education in organisational development and the other was an organisational consultant. These are not included in this thesis, because their area of focus was substantially different to the other facilitators. They were predominantly working with practices’ missions, visions, cultural issues, and meeting management. 13 Plan-Do-Study-Act circles. Plan: Plan a change. Do: Try the change. Study: Observe the results. Act: Act on what is learned. 14 The brown paper method is a visual display of a process with big post-its on the wall, where a practice actively focuses on current and future workflows and division of tasks.

16

diverse components relevant for conducting facilitation visits, but the facilitators could use these as they saw fit. Further, they had sought to give facilitators a sense of confidence regarding the content of their role and to provide them with a shared platform and language. Although the project managers described and recommended a certain degree of flexibility, they also regarded a few topics as mandatory. For instance, they expected facilitators to explain that the frame of the visits was the CDMP for diabetes and COPD, to describe the interventions’ success criteria, and to introduce a webpage of municipal chronic care services, to which GPs can refer patients.

Practice visits and the role of the facilitators All practices in the Capital Region of Denmark were offered up to three visits of one hour each and the timespan between the visits was flexible. Visits were free of charge and the practices were compensated for lost income. Within practices it was recommended that all GPs and staff participated in the visits. In the Capital Region of Denmark, there were 762 practices when the intervention was initiated. Of these, 504 signed up for the facilitation intervention and 431 received at least one visit.

The potential topics to cover during the facilitation visits described in the information material to the practices, were:

• Workflow and division of tasks for chronic disease check-ups • Overview of IT-solutions to improve systematisation in chronic care (applied in the intervention

as the DCM) • International Classification of Primary Care (ICPC) diagnosis coding • Patient stratification • Leadership and organisation • Collaboration with municipalities and hospitals • The role of GPs as coordinators of care

Before the visits, each practice was asked to fill out two questionnaires. One was a baseline questionnaire containing items about practice characteristics, use of the DCM, annual chronic disease check-ups for diabetes and COPD, diagnosis coding, and stratification. The other questionnaire focused on the practice’s knowledge of the CDMP, division of tasks in the practice, collaboration with the municipalities, and their suggestions for topics for the visits. The facilitators were intended to use these questionnaires in their preparation for the visits.

During the facilitation visits, the facilitators’ role was to act as a catalyst for change by:

1. Ensuring that the change process was driven by the practices’ motivation, wish for development, and choice of topics.

2. Providing information to the practices about the CDMP. 3. Acting as sparring partners and colleagues rather than experts. 4. Engaging in a dialogue with the practices about goals for development. In the information

material to the practices it was also stated that gathering the whole practice for a facilitation visit was an obvious opportunity to talk about work flows, tasks. and common goals.

17

5. Assisting the practices in defining specific objectives for the visits and in choosing suitable means for achieving them.

6. Providing relevant tools and suggesting courses and other regional consultant programmes if needed.

7. Ensuring the structure of the visits by managing agenda setting and time, and after each visit providing the practice with a standardised visit-report containing the topics discussed, goals agreed, and the between-visit tasks to be completed.

The project initiators and project managers mentioned multiple facilitation activities in the interviews: information dissemination, experience based knowledge from a peer, facilitated internal dialogues, ensuring motivation and a sense of meaningfulness, and meeting management (time management, ensuring everybody had their say, opening and closing the meeting, and getting the practice to commit to intermediary tasks). They described how facilitators were to be flexible in their approach and that the approach was to be tailored to the individual practice.

An idea of a continuum of facilitator roles was emphasised in the interviews and was described as having been highlighted in the education of facilitators. This continuum ranged from the role of teacher or expert at one end to the role of coach at the other end, with the role of a sparring partner in-between (Figure 1). Although the facilitators were not expected to be technical or disease specific experts, it was intended that they mastered most roles on the continuum and switched between roles according to the situation and the needs of the practice. The precise content of the specific facilitator roles in the continuum and the differences between them was not explained in the interviews. PowerPoint slides from the facilitators’ education described a coaching approach as a “helping and focused conversation between two [or more] persons, where one, by using open and focused questions and neutral formulations, gives the other/others the possibility to formulate problems/challenges and create possible solutions”. The project managers said that they had prioritised the coaching role during the facilitation education, because they had presumed this role would be more difficult for the facilitators to enact, since it was likely to be more unfamiliar to them given their background as GPs.

Figure 1. The continuum of facilitator roles intended in the intervention

Intended outcomes The written success criteria were threefold. First, that practices had implemented their chosen goals. Second, that the division of tasks between all GPs and staff groups in the practice related to patients with diabetes and COPD were systematically described and performed. This entailed that the practice should conduct systematic chronic disease check-ups, stratify patients, use ICPC diagnosis

Expert/teacher Sparring Partner Coach

18

coding, and use the DCM (or another IT tool) for patient overview. The third goal was that general practice handled its part of a coordinated and intersectoral collaboration. This involved GPs as coordinators of care, a continuous patient assessment, systematic patient follow-up, a proactive approach, and support for the maintenance of treatment goals. According to the interviews with the project initiators and project managers, the practices should also have a legacy role by the end of the intervention, i.e. they should be able to continue development work after the visits and be better equipped with methods to implement future CDMP.

19

Method The qualitative studies For the two qualitative studies that form part of this thesis, I chose an exploratory approach in both data collection and analysis. This was because the conceptual definitions of facilitation and the previous facilitation studies did not provide sufficient descriptions to qualify for a relevant deductive approach. Hence, I considered that an a priori choice of focus or theory before entering the field would be too arbitrary.

Material The data in the qualitative study consist of:

1. Observations of facilitation visits in 13 practice settings (4 of the 13 were visits involving collaborating practices, hence a total of 18 practices were represented in my data). I followed the practices from their first visit, and there were between one and three visits in each practice setting15, hence 30 visits were observed. Seven practice settings were observed and interviewed in the first year of the intervention; six were observed and interviewed in the second year to give room for potential changes in facilitation or contextual conditions. However, no noticeable differences were found.

2. Interviews with the observed practices twice – once after the first and once after the last visit. I conducted 20 interviews, which lasted for one hour each and took place in the practices.

3. Two one-hour focus groups with nearly all facilitators and individual one-hour interviews with seven of the facilitators whom I had observed at the facilitation visits (some of the observed practices had the same facilitator).

4. Observations of five facilitator network meetings.

Tables 1 and 2 demonstrate the data gathered in the study and, as shown, data collection was not complete for all practice settings.

The practices were sampled strategically to ensure variation in geography, practice size, facilitator, current use of the DCM, and status in CDMP (assessed by the initial questionnaires). I assumed these elements could influence variations in both facilitation enactment and the experienced outcomes of the visits [95]. Since the initiation of the DCM was a dominant theme in most of the practices that I observed in the first year, I purposefully added an additional selection criterion that some of the later practices either should have had the DCM for a while, or explicitly have suggested another topic in the questionnaire. I did this to explore if the preliminary findings were connected to the practices focusing on the set-up of this IT-system.

15 One practice had a fourth facilitation visit, because one of their visits was video recorded by the Region for use in the advertising of the facilitation intervention.

20

Table 1. Participating practices and data material

Practice type

Participants Primary topic Observation visit 1

Interview 1

Observation visit 2

Observation visit 3

Interview 2

Facilitator Comments

1.

Solo

GP, 1-2 nurses, 1 GP in training

DCM (potential and correct set-up) Diagnosis coding (need) Reflective exercise

x

1 GP

x -

1 GP

A

Third visit completed, but not observed.

2. Solo 2-3 GPs, 2 secretaries, 1-2 GPs in training 1 temporary GP

DCM (potential and installation) Diagnosis coding (need) Webpages

x

2 GPs, 2 secretaries, 2 GPs in training

x x

3 GPs 2 GPs in training B

Partnership

3. Partnership 4-5 GPs DCM (potential, installation and correct set-up) Diagnosis coding (need) Brief introduction to new meeting structure

x

4 GPs

x x

4 GPs

C

Was given a fourth visit because the third was video recorded for intervention promotion. The fourth visit was also observed

Partnership

Solo

4.

Partnership

3 GPs, 2 GPs in training, 1 nurse 1 secretary

General questioning of practice and suggestions for COPD detection, testing, registration, and prevention and use of health care centre

x

2 GPs, 2 GPs in training, 1 nurse

D

Declined more visits.

5.

Partnership

2 GPs DCM (potential and installation) Webpages Diagnosis coding

x

2 GPs

x x - E

Partnership ended between 2nd and 3rd visit. The last visits were in both of the new practices. I did not manage to schedule an interview.

6. Solo 1 GP, 2 nurses DCM (potential and installation) x 1 GP,

2 nurses 1 GP C Intended more visits, but they did not contact the facilitator.

7. Solo

2-3 GPs, 3 nurses, 1 secretary, 1 GP in training

New procedures for COPD check-ups DCM (brief introduction, potential and partly correct set-up) Webpages

x

3 GPs, 3 nurses, 1 secretary, 1 GP in training

x

3 GPs 3 nurses 1 GP in training

F

Three hours of facilitation in two visits

Partnership

Continues on the next page

21

X observed visits; - not observed visits or not interviewed; shaded areas are not conducted visit

8. Solo 2 GP, 1 nurse

DCM (potential and correct set-up) Diagnosis coding (need and progress)

x

2 GPs

x - - B

Third visits completed, but not observed. I did not manage to schedule an interview

Solo

9.

Solo

1 GP, 1 nurse

DCM Webpages COPD professional knowledge Annual chronic disease check-ups

x

1 GP, 1 nurse

x

1 GP 1 nurse

G

Three hours of facilitation in two visits

10. Solo

1 GP DCM (potential and correct set-up) Diagnosis coding (need)

x 1 GP

x x 1 GP

H

11.

Solo

1 GP, 1 nurse

DCM (potential) Webpages x

1 GP, 1 nurse x - I

Facilitator stopped and another was not chosen. I did not manage to schedule an interview

12.

Partnership

2-3 GPs, 1-2 GPs in training, 1-2 nurse, 1 secretary, 1 1 healthcare assistant

New procedures for diabetes and COPD check-ups Diabetes/COPD professional knowledge DCM (potential)

x

3 GPs, 1 GP in training, 1 nurse, 1 healthcare assistant

x x - J

I did not manage to schedule an interview

13.

Partnership

2 GP, 1 nurse, 1 secretary, 0-1 GP in training

DCM (potential) Diagnosis coding Procedures for diabetes and COPD

x

2 GP, 1 nurse x - D

Three hours of facilitation in two visits. I did not manage to schedule an interview

22

Table 2. Facilitator characteristics and data material

Facilitator Facilitator gender

Facilitator age Individual interview Focus group Observed

A Female 45-49 + + + B Female 55-59 + + + C Male 40-44 + + + D Male 60-64 + + + E Female 60-64 - + + F Female 40-44 + - + G Female 45-49 + + + H Female 45-49 + + + I Female 45-49 - + + J Female 55-59 - + + K Male 60-64 - + - L Male 50-55 - + - M Female 50-55 - + - N Male 60-64 - - -

+ interviewed, in focus group or observed; - not interviewed, not in focus group or not observed

Observations During the observations of the facilitation visits I wrote extensive notes and the visits were audio recorded. The purpose of the observations was to explore meeting structure, content, dialogue, and interaction between facilitator and practice. The observations were also used to qualify the subsequent practice and facilitator interviews. The initial observations were highly explorative and all apparently interesting observations were written down. As my understanding changed, certain elements were given increased focus at the subsequent observations; however, the explorative approach was maintained ensuring that new perspectives were discovered. The following elements increasingly became the areas of attention in the observations:

• People attending • Physical setting and seating arrangement • Introduction and agenda setting • Topics and time spent on different topics • Meeting structure and use of PowerPoint slides and PC • Dialogue between the facilitator and the practice and between the individual participants • Communication form • Nonverbal expressions and behaviour • Interruptions • The ending of the visits, task agreement, and scheduling of new visits • The handling of agreed upon tasks at subsequent visits

The notes were intended to supplement the audio recordings; hence they were a mixture of the overall content and structure of the visit, nonverbal elements, information I perceived as important to understand the audio recordings, my impressions and thoughts, and potential interview questions. A student assistant who was informed about what to observe and how to take notes, observed five

23

of the visits. After each of these visits I listened to the audio recording and added more notes; we discussed the visit and I asked the student assistant for additional relevant information.

I observed five facilitator network meetings during the intervention period to obtain an insight into the facilitators’ experiences and how the facilitator role and the intervention as a whole were discussed and defined amongst the facilitators and the project managers.

In addition to a final evaluation report to the Region, we were also obliged midway to present our preliminary findings to the facilitators, projects managers, and a project steering committee for their continuous adjustment of the intervention. I was aware that this might have implications for the studied intervention. However, I chose to perceive it as an opportunity to hear their responses to my preliminary findings. Additionally, I did not identify profound changes as a consequence of the presentations. An evaluative network meeting at the end of the intervention period was also observed and audio recorded. At this meeting I presented my findings and the facilitators expressed their thoughts thereof, and there was a session led by the project managers where they discussed their perceptions of facilitator education, the organisation of the intervention, and the visits.

Interviews I performed all interviews, and they were audio recorded. For the facilitators conducting the visits that I had observed, the interview guides were divided into questions regarding the observed practice and the facilitator’s general perceptions and behaviour.

About the observed practice • The preparation of the facilitator and the practice • Their descriptions and experience of the meetings • Their own role at the visits and in the change process • The practice impact from the visits • What they perceived as successful/less successful in the specific practice • How visits differed from other practices and what was similar between practices About their general perception and practice • Their understanding of the facilitator intervention • Their contribution as facilitators • How they were prepared for the role as facilitator • What had influenced their understanding and enactment of facilitation • Variations between facilitators and practices • The implication of being GPs • Thoughts about project design (number of visits, meetings in the practice etc.)

I conducted the interviews with practices as group interviews. The aspiration was that all practice members who were present at the facilitation visit also participated in the interview. However, in most interviews some staff members were absent (Table 1). I posed questions to the group, but when it arose that some participants did not state their thoughts during the interview, I explicitly addressed questions to them. The overall themes in the practice interviews were:

24

In both first and second interview • The visit’s dialogue, content, and structure • The role of the facilitator • Their understanding and assessment of the intervention • The intervention design (number of visits, participation of the whole team, GP facilitators etc.) • Ideas of intervention adjustment • Specific questions related to my impressions at the observed visits in the particular practice In the first interview • The practice’s expectations and preparation • Thoughts about the future process In the second interview • The practice’s process between visits • Experienced outcomes

Interview guides for both practices and facilitators were continuously revised and new questions were added based on my changed understanding, new perspectives, and challenges with specific questions. Throughout the data collection, I increasingly focused on the enacted facilitator roles, the balance between roles, absent roles, the facilitator as a peer, and the contribution of the facilitation visits in the process of change. Given this increased attention, I added questions in both facilitator and practice interviews to foster reflection on their understanding and experience of intended/expected and enacted facilitation and facilitator roles. To aid their reflections, I presented a drawing of the continuum of facilitator roles that were intended in the intervention. I asked the interviewees to pinpoint the role or roles that they had enacted (facilitators) or felt the facilitator had enacted (practices) (Figure 1). I also asked them for synonyms for a facilitator and presented a list that I had composed with suggested facilitator synonyms.

Focus groups with facilitators The facilitators were divided into two focus groups which were concurrently led by me and by one of my supervisors and co-authors. The overall themes in the two focus groups were:

Focus group 1 • Their understanding and description of the facilitator intervention • Their preparation • The visit content • The good visit • Their competences • Their tools • Facilitator variations and unity • The impact for the practices Focus group 2 • Their understanding of facilitation and the roles of the facilitator • If and why some roles were more often applied • What influenced their roles • Whether they had collaborated on a common understanding

25

• Improvements of their performance over time • Their competences • The significance of them being GPs

With focus groups, it is possible to explore similarities and differences in the participants’ views and experiences, which are illuminated more clearly because the participants have the opportunity to reflect upon each other's statements [95]. However, I was aware that participants’ statements in focus groups do not necessarily represent their prior perceptions. Discussions and the group dynamic within the focus group can influence participants’ perceptions and statements. This also means that focus groups can be used to explore both the content of the participants’ statements and the group dynamic in discussions about the given topic [96, 97]. Therefore, as well as using focus groups to explore perceptions and experiences, I also considered how participants discussed and handled differences that emerged during the focus group session. I was able to identify some differences between statements made in individual interviews and those made in the focus groups.

A hermeneutic approach In the qualitative studies, I was influenced by hermeneutics, which focuses on interpretation of the meaning of texts and human experience [98]. My attention was drawn by two central concepts within hermeneutics – the hermeneutic circle and pre-understanding16.

Traditionally the hermeneutic circle entailed that the interpreter shifted between parts of the text and the whole text, where the whole was initially perceived as the entire text, and later also as the texts’ context (for example, a persons’ life and their society) [99]. Hence, the interpreter deepens the understanding by interpreting the parts, the whole, and the relationship between them, and by sifting between these until there is a meaning without inner contradictions [99, 100].

According to later hermeneutic tradition, the hermeneutic circle is not a methodological circle but one that describes continuous movement between the interpreter’s pre-understanding, interpretation, and understanding [99]. Contrary to other philosophies of science and methods of analysis, for instance phenomenology and grounded theory, pre-understanding in hermeneutics is perceived as something that cannot be bracketed (i.e. identified and removed) from the process of interpretation [99]. Pre-understanding is constantly present, both as constructively enabling interpretation and as setting limits to the range of the understanding. Further, central to the hermeneutic approach is that one’s pre-understanding constantly is and actively should be confronted and challenged [99].

Hence, with this influence from hermeneutics, I focused on understanding and interpreting participants’ perceptions, experiences, and behaviour. Throughout the whole process I reflected upon and wrote down in a log file my pre-understanding and how it changed. I challenged my pre-understanding and my changing understanding in the engagement with the field, the continuous reading of facilitation research literature, and discussions with other researchers (e.g. discussions with my supervisors and colleagues, presentations at research conferences, and in presentations and discussions during my research exchange). I strived to be conscious about how my pre-

16 The concept of pre-understandings is often also called preconceptions.

26

understanding and evolving understanding influenced the questions that I posed to the participants and in the analysis. At the same time, I acknowledged that a complete awareness of my pre-understanding and its influence cannot be attained [99].

I used the hermeneutic circle in my continuous attention to movements between my pre-understanding and understanding of the intervention and the field of facilitation. In the analysis, I also used the methodological hermeneutic circle. In the analysis of individual interviews, focus groups, and observations, I focused on the single extracted statements and related them to the context of the entire individual interview, focus group, or observation. In the analysis of facilitation enactment, I related the single coded statements in the facilitator interviews to the entire individual interview, to the individual facilitator’s statements in the focus groups and to his or her observed visits. The individual statements from the focus groups were also analysed in relation to the focus group as a whole. In the analysis of the practices’ experiences, I related statements to the entire interview and to all interviews and observations conducted in each single practice. Further, in both analyses, I related extracted statements and whole interviews, focus groups, and observations across facilitators and practices with each other.

My pre-understanding

Professional background My professional background as a Master of Science in Public Health Science may have influenced my data collection and analysis. As I am not a GP, I may have been able to identify and question implicit understandings in the practice culture and have a more critical perspective. A possible limitation of not being a GP is a potentially more limited understanding of the culture in general practice, and the reasons for or advantages of certain ways of organising or conducting change processes. Further, my profession may have influenced the participants’ degree of openness in the interviews. However, whether my status as an “outsider” meant that participants were more open or more guarded about speaking their minds, because they did not need to consider judgement from a peer, is uncertain.

I have had prior experience of conducting evaluations of municipal and intersectoral preventive and chronic care initiatives, and empirical and theoretical knowledge of implementation processes. This means that prior to the study I already had an understanding of motivation, skills, social norms, the organisational frame, and contextual conditions as influential factors in the process of change, and an understanding that knowledge alone is rarely enough to ensure change. This might have caused me to be more attentive towards these areas and more sceptical of a simple change process.

Assumptions As part of my pre-understanding at the onset of data collection, I understood facilitation as entailing a range of potential activities and facilitator roles. I was inspired by the understanding of facilitation as a continuum of roles, and its potential to embrace the various roles described in the material of this intervention, and in the facilitation research literature. However, I expected a facilitator primarily to be a person who stimulated and managed an internal dialogue by ensuring reflection and the development of new procedures. Nevertheless, I did consider it likely that it might be

27

challenging for the GP facilitators to handle this role, due to their professional background. I had anticipated that I would witness internal developmental discussions during the facilitation visits, e.g. of new procedures of chronic care and division of work, and therefore I was quite surprised when I witnessed a very different type of facilitation at the observed visits. This has most likely influenced my continued attention towards the diverse facilitator roles, the balance between them, and the practices’ experiences thereof. I was aware of continuously considering that although the facilitation approach was different from what I had anticipated, it might not conflict with the practices’ expectations or wishes. Further, regarding practice expectations, I was initially sceptical as to whether the practices had sufficient knowledge of the intervention or understanding of the concept of facilitation, because the information material provided to them was rather vague.

My initial unit of analysis was the practice; hence, the facilitation processes within practices, and how facilitation influenced them to change. This was because I considered that the facilitation process and the level of change would very likely be influenced by potential differences between the practices and the tailored approaches of the facilitated intervention. Hence, I strategically selected and recruited different kinds of practice rather than making a selection based on the aim of following specific and diverse facilitators. Prior to the data collection, I had not considered the possibility of profound differences between facilitators. However, I became more attentive to inter-facilitator variations in the focus groups. I also became aware that some facilitators who conducted visits in two different observed practices, appeared to act in a very similar way across practices compared to other facilitators.

Further, early in the data collection, I was interested in exploring the influence of the facilitators being the peers of the people they were facilitating. This was influenced by the peer aspect seemingly playing a central role in this intervention and also in other facilitation interventions in Denmark [16, 19]. On the one hand, I considered it likely that practices would perceive the peer aspect as beneficial and that it would reinforce the process. On the other hand, I also considered the possibility that peer facilitators could pose a risk of being too close, making it more difficult for the facilitator to be critical and to challenge the practices’ perspectives.

Before the data collection I was interested in the processes initiated by the practices due to facilitation visits, the visits’ influence in a learning process and how the enacted facilitation and contextual conditions shaped the attained changes. However, this proved less relevant and attainable given the enacted facilitation and limited practice processes, and due to limitations in the ability to identify patterns in the data in relation to attained changes. Further, my research exchange led to an introduction to realistic evaluation and an interest in mechanisms and contextual influences. This steered my attention towards exploring by which means the visits influenced the practices and the impact of contextual conditions.

As for the effectiveness of the intervention, I was sceptical as to whether there would be an effect and whether any potential effect would be measurable. I was aware that this scepticism might make me excessively attentive towards disadvantages in both observations and interviews. When I came across elements that made me question or be critical of the facilitation visits and the practices’

28

statements, I took great care to question myself to determine what influenced this understanding. On the other hand, my occasional scepticism also meant that I also questioned and challenged other aspects too, for instance the immediate positive assessment of the intervention by facilitators and practices.

Qualitative analysis I used thematic analysis inspired by Braun and Clarke’s description and step-by-step guide [101]. In thematic analysis, the aim is to identify, analyse and report repeated patterns of meaning within the data. The approach can be either inductive or deductive and my analyses were inductive, where I identified themes linked to the data. However, in the analysis for the first paper (Appendix 3), I was influenced by my pre-understanding of potential facilitator roles and facilitation activities, and a continuum hereof. In the analysis of the practice interviews for the second paper (Appendix 4), I was influenced by an interest in how the facilitation visits influenced the practices, how they valued the visits, and the influence of contextual conditions.

Braun and Clarke offered a set of guidelines for conducting a thematic analysis and I will present my analysis in accordance with these steps (Figure 3). Thematic analysis is not a linear process, and I moved back and forth between the phases.

Figure 2. Phases of thematic analysis by Braun and Clarke [101]

Phase1 After each observed visit, interview, and focus group, I wrote down spontaneous reflections for the refinement of the interview guides and the later analysis. In this phase, the interviews and focus groups were transcribed verbatim, predominately by a student assistant following a transcription manual that I had written, and I proof-read all transcripts and checked them against the audio recordings. I listened to each audio recording of the observations and added more notes if needed. I read the entire data before the coding process and wrote down initial notes (thoughts, potential codes, and themes) which were linked to the individual piece of material and from across the entire dataset.

29

Phase 2 NVivo qualitative data analysis software was used in the coding and theme constructing process for the interviews and focus groups. As recommended by Braun and Clarke, I systematically went through the entire dataset and paid equal attention to all parts. Each interview was coded and, as Braun and Clarke described, the length of a data extract (coded chunk of data) varied with sometimes several different codes applied to individual extracts. I coded the interviews with the facilitators and the focus groups inductively sentence by sentence, with short sentence-specific codes describing the content. In the analysis of the practice interviews, I selected a list of overarching codes after detailed reading of all the interviews. These were “Preparation and expectations”, “Facilitator contribution”, “Attitudes towards enacted facilitation”, “Perceived changes”, “Between visits”, “Contextual influencing conditions”, “Perceived deficiencies”, and “Attitude towards the design”.

In the analysis of facilitation enactment for the first paper, the observation notes together with the audio recordings were analysed in relation to the interview-based themes. The observations provided both supplementary details of the enactment of facilitation and insights not revealed in the interviews. In the analysis for the second paper, I used the observations to identify task completion, concrete changes, and elements potentially limiting the facilitation impact based on practices’ descriptions during the visits of their processes and challenges.

Phase 3 In this phase of re-focusing the analysis, the task is to sort the different codes into potential overarching themes, and to identify the relationships between codes, themes, and different levels of themes. In this phase, I grouped the sentence-specific codes in the facilitator interviews and focus groups into themes and sub-themes. In the practice interviews, I recoded the content in each of the overarching codes into more detailed codes, and thereafter organised them into themes and subthemes.

Phase 4 In this phase of theme refinement, I compared the content of the different themes. I identified their distinctiveness and divided or merged themes accordingly. For example, this was the case in the identification of enacted facilitator roles where some roles were split (a role of teacher and a role of super user) and others were merged (a role of motivator was split and then merged in part into a role of teacher and in part into a role of peer). As Braun and Clarke described in this phase, I ensured an overview of codes and themes, and constructed maps of themes and subthemes. For this I used the NVivo program. However, the overviews were not exclusively in phase 4, but were created and refined throughout the analysis.

Phase 5 In this phase one has to go back to the data extracts for each theme and organise them into a coherent and internally consistent account with an accompanying narrative. In both analyses, I wrote a coherent text of the identified themes. I did this by reading the grouped coded data extracts and initial codes and relating them within and across the interview and focus group contexts. Each

30

theme narrative also included the analysed data from the observations. In case of questions of doubt or puzzlement the individual interviews were re-read and observation notes and audio-recordings were re-examined. These documents of combined narratives were read by and discussed with my supervisors (co-authors). We discussed the analysis, and the focus, content, and structure of the papers over a period of time. This process also lead me to re-read extracts of the different codes and themes, whole interviews and observation notes, and to re-listen to audio recordings.

Phase 6 Not all themes described in the narratives from phase 5 were included in the final analysis because of the need for focus in the research papers. Some of the omitted findings are presented in this thesis as additional findings to paper one. In this final phase, the papers were written, illustrative extracts were chosen for quotations, and related facilitation research literature was identified and compared.

The randomised controlled trial For the effectiveness study (Appendix 5), a stepped-wedge, RCT was used, where an intervention is rolled out to groups sequentially and data is collected at baseline and before new groups receive the intervention. Groups that have not yet received the intervention thereby act as control groups [102]. In this study there were only two groups, unlike other stepped-wedge studies where several sequential groups are included [102]. This design was chosen because it was considered important to conduct an RCT given the high likelihood for contextual influences in this field. At the same time a pure control group was not an option because the intervention was offered to all practices in the region.

Outcome measures and data The primary outcome was: • Change in the number of annual chronic disease check-ups per 100 patients affiliated with the

practice as per the Danish National Health Service Register (DNHSR) [103].

The secondary outcomes were: • Reduction in the number of practices with few (less than 1%) annual chronic disease check-ups

per 100 patients affiliated with the practice (DNHSR). • Change in the number of spirometry tests per 100 patients affiliated with the practice (DNHSR). • Change in the number of annual chronic disease check-ups for diabetes and COPD (self-

reported). • Sign-up to the DCM (register based). • Changes in the use of ICPC diagnosis coding for diabetes and COPD (self-reported). • Changes in the use of stratification of patients with diabetes and COPD (self-reported).

The power calculation was based on data from the DNHSR on the use of annual chronic disease check-ups and two parameters were calculated:

• To increase the number of annual chronic disease check-ups from 1.25 to 2.0 consultations per 100 patients per quarter.

31

Randomisation

• To reduce the number of practices with less than 1% annual chronic disease check-ups per 100 patients per quarter from 40% to 20%.

At power of 80% and a significance level of 5% respectively 128 and 163 practices were needed for the two parameters. The estimated dropout rate was 10%. Of the 762 eligible practices, 189 practices were consecutively included as they signed up for facilitation visits and completed the baseline questionnaire. The practices were randomly allocated to facilitation visits in 2011 (intervention group) or to facilitation visits in 2012 (control group). They were stratified by practice type (solo or partnership practice) and by geographical location.

Data were retrieved from the DNHSR (for the primary outcome and some of the secondary outcomes), the Danish Quality Unit of General Practice (for sign-up to the DCM) [92], and from practice questionnaires. The baseline questionnaire that the practices filled out before the visits was used in the RCT for baseline data and a follow-up questionnaire was used for assessment of change. The visit reports filled out by the facilitators after the visits provided data on the number of visits to the practices. The baseline questionnaires were collected before randomisation, and registry data were collected for each practice in the three months up to randomisation. The follow-up registry data were collected a year later in the three months equivalent to the baseline period. The follow-up questionnaires were sent out on 15 February 2012. Thus, the follow-up was 12 months and the intervention 9 months. The timetable of the data collection is presented in Figure 3.

Figure 3. The timetable in the RCT

Analysis The trial flow of the RCT is illustrated in Figure 4. In the statistical analysis, differences in the use of annual chronic disease check-ups and spirometry between allocation groups were assessed using t-tests. Differences between allocation groups regarding the rest of the outcome measures were assessed using chi-squared tests. All differences were assessed at baseline and at follow-up. The difference in sign-up rate to the DCM over time was visualised in a Kaplan-Meier plot and analysed with a log-rank test. All statistical analyses were performed using SAS version 9.2 (SAS Institute Inc., Cary, NC).

2010 2011 2012 Dec. Jan. Feb. Mar. Apr. May - Nov. Dec. Jan. Feb. Mar.

Baseline

Registry data

Baseline Questionnaire

Intervention group - facilitator visits

Delayed intervention group - no visits

Follow-up Registry data

Follow-up Questionnaire

32

Figure 4. The trial flow of the RCT

ANAL

YSIS

FO

LLO

W U

P AL

LOCA

TED

ENRO

LLM

ENT

Assessed for eligibility (n = 762)

Excluded (n = 573)

• Not meeting inclusion criteria (n = 23) • Did not sign-up for the intervention or did not answer the baseline questionnaire before the end of randomisation (n = 550)

Randomised (n = 189)

Facilitator visits Delayed intervention group

Allocated for intervention (n = 96) • Received allocated intervention (n = 87) • Did not receive allocated intervention (n = 9) Reason(s):

Did not receive any visits (n = 9)

Allocated for intervention (n = 93) • Received allocated intervention (n = 80) • Did not receive allocated intervention (n = 13) Reason(s):

Received visits (n = 13)

Lost to follow up (n = 7) Reason(s):

Retirement (n = 2) Not responded questionnaire (n = 5)

Lost to follow up (n = 11) Reason(s):

Retirement (n = 4) Not responded questionnaire (n = 7)

Analysed Registry data: (n=94) Questionnaire data:(n=89)

Excluded from analysis (n = 0)

Analysed Registry data: (n=89) Questionnaire data: (n=82)

Excluded from analysis (n = 0)

33

Findings Paper 1. Role enactment of facilitation in primary care – a qualitative study (Appendix 3)

Purpose The purpose of the study was to explore the enactment of facilitation in specific roles carried out by external peer facilitators as they sought to support the uptake of CDMP in general practice.

Method The data were observations of 30 facilitation visits, 2 focus groups with facilitators, and 7 individual facilitator interviews.

Findings Facilitation in this intervention was enacted mainly in the form of four different facilitator roles. These roles were combined in various ways during the visits with some roles being more pronounced than others. Although there were variations between the facilitators and the facilitation visits, we primarily focused on the similarities. A fifth role of coach was also described; however this role was rather absent in the observed visits.

The role of “teacher” consisted of knowledge dissemination and motivation regarding central elements of the CDMP (such as ICPC diagnosis coding, annual chronic disease check-ups, stratification etc.), the DCM, and relevant websites on professional guidelines and municipal chronic care services. The facilitators used more or less structured PowerPoint presentations (on PC or projector), talks, demo versions of the DCM, or demonstrations of relevant websites.

The role of “super user” also entailed factual knowledge, but used hands-on guidance within the practice’s own computer system, mostly with a GP from the practice in front of a PC. The facilitators demonstrated unfamiliar features in the electronic patient record system, showed how to use the DCM, discovered errors in the set-up of the DCM, and talked about data (e.g. coding percentages, missing annual chronic disease check-ups, and improvement in registration at subsequent visits). The facilitators could provide some technical insights, but they did not perceive their role as being technical experts and lacked knowledge on the different patient records systems. Factual knowledge dissemination had a rather educative character, but it also actively involved the participants by focusing on their wishes, questions, and comments. The active involvement was more pronounced during hands-on dissemination than the more presentation based.

The role of “peer” underlined the facilitators’ professional status as colleagues from general practice. They used the word “we” in the conversations, indicating their common professional identities and working conditions. They emphasised their perceptions of the benefits of increased systematisation in their own practices. They often referred to their own experiences when describing ways of organising the practice. In some cases the facilitators briefly mentioned such experiences during general topic presentations, in other cases they provided comprehensive

34

description of their own practice organisation in terms of structure, work division, annual chronic disease check-ups, and use of the DCM. By referencing their own lack of perfection, their process experiences, the changes they achieved quickly, and how they had overcome obstacles, the facilitators aspired to prevent the practices from seeing the change process as overwhelming.

The role of “process manager” was identified in agenda setting, task definition, and meeting management activities. Agenda setting seemed influenced by a combination of the practices’ considerations before the visits, the practices’ ideas during the visits, the responses to the initial questionnaires, and facilitator inspired and driven topics. The facilitators promoted agreement on tasks, though in varied levels of systematisation, and kept up momentum in the process by status reporting at subsequent visits. As for meeting management, the visits were often held with a limited degree of structure, with several topics and subtopics covered in varied levels of detail, and with a varied degree of active participation (especially by the staff). The facilitators often sought to include all the different participants in the meetings. The observations point to occasional challenges with this role. The challenges mentioned by the facilitators were related to lack of time in larger practices, some practice members not attending, and delayed or rescheduled meetings.

The role of “coach” was explored in the analysis, because the facilitators were meant to help the practices articulate various problems and solutions and support internal reflections and discussions. However, the enactment of the role of coach appeared to be limited in this intervention. The facilitators did little to stimulate such reflective discussions during the visits. Discussions and clarifications in the practice were often emphasised as important, but it was suggested that these could be done outside of the facilitation visits. Further, subsequent visits often focused on status reporting, additional knowledge provision linked to experienced challenges concerning the DCM, and not on discussion e.g. on implementation of new procedures. Few facilitators described practising aspects of this role in visits other than the ones observed in the study. The facilitators primarily linked the limited use of the coaching role to the fact that the practices had very basic and concrete needs that required definite guidance, and they also said that time was too limited to take such an approach. However, some facilitators also indicated that it was related to their own comfort zones, competences, and personalities.

35

Additional findings of facilitation enactment not presented in paper 1 These additional findings on how facilitation was enacted in this intervention elaborate on two issues briefly mentioned in Paper 1. First, variations in the facilitation enactment across visits were briefly described in the paper, but the focus was mainly on the similarities. Second, we indicated that there were differences in the degree to which the facilitators guided the process, mainly in agenda setting. Thus, to further nuance the understanding of the facilitation enactment in this intervention, findings of variations being both practice and facilitator dependent, and facilitation processes being both practice and facilitator driven, are elaborated below.

Practice and facilitator dependent variations There were several variations in how facilitation was enacted at the visits regarding choice of topic, structure of the visits, and the most marked facilitator role adopted. Based on the interviews, focus groups, and observations these variations seemed influenced by both variations between practices, and thereby a direct result of the tailored approach, and variations between facilitators, due to differences in their interpretation of the concept of facilitation. Based on the data I could not determine in which situations variations were practice dependent and in which situations the variations were facilitator dependent.

Practice dependent variations In accordance with the principal idea of tailoring the facilitation visits to practice needs and wishes, the facilitators emphasised that there were variations between their visits due to differences between practices and even due to different situations within the same practice. According to the facilitators, differences between practices were due to factors such as differences in practice size, differences in prior knowledge about the topics, traditions within practices for internal discussions, and the ability of practices to implement new things. Regarding the size of the practice, some facilitators perceived certain topics as more relevant for the larger practices, e.g. organisation of meetings and division of tasks, and likewise that a more hands-on facilitation approach was often more relevant in the smaller practices (solo practices). Additionally, some facilitators perceived that the coaching based approach was more obvious and easier to apply in larger practices with a tradition of developmental discussions. In accordance with the continuum of facilitator roles emphasised in the intervention, the facilitators described how they shifted between approaches in different situations at the same visit, and generally experienced moving towards a more coaching based approach during subsequent visits to the practices.

Facilitator dependent variations Even though the facilitators in the interviews and focus groups focused on practice dependent variations, they also expressed inter-facilitator variations, and these types of variation were also apparent during the observed visits.

Facilitator dependent variations were noticeable in the structure of the visits, the facilitators’ use of the facilitator roles, the degree of detail and time spent on introducing topics, the use of IT equipment, the facilitators’ knowledge about the specific IT systems within the practice, the extent to which the facilitators guided the topics and the tasks, and the extent of the facilitators’ ambitions

36

on behalf of the practices. For example, when introducing the DCM, some facilitators used extensive PowerPoint presentations while others gave brief motivational speeches. Further, regarding the interventions’ facilitation continuum, the facilitators varied in which role they placed themselves overall. Thus, some facilitators indicated that they generally were somewhere between teaching and sparring and others between sparring and coaching

The facilitators perceived the educational programme and the network meetings as a source of inspiration to their individual interpretation of the facilitator role and generally not as requirements from the project managers. They differed in their understanding of which elements of the project were mandatory. Even when something was perceived to be mandatory, it was also perceived as relatively minor issues which had to be mentioned during the visits, but was not related to the visits’ overall structure or to the facilitation approaches. The facilitators described that their interpretation of being a facilitator was influenced by a mixture of their facilitation education, experiences of change processes and structures in their own practices, their individual areas of interest, and other job-functions they possessed (e.g. working as a practice coordinator regarding COPD or municipal services). Further, the facilitators described that their personalities and comfort zones influenced their enactment of facilitation.

In general, there was consensus among the facilitators that variations and flexibility were natural, and they expressed a respect for each other’s differences. They did not seek conformity and did not perceive that some types of facilitation enactments were more correct than others. They emphasised the importance of flexibility to take into account both the practices’ different needs and the facilitators’ different preferences and personalities. Consensus seeking was also evident in the focus groups where none expressed disapproving attitudes towards the variations. When differences in their approaches to facilitation were brought to light, the immediate response was acceptance. This acceptance was linked to the facilitators’ comfort zones and a belief that different means would lead to the same results. In the focus groups there were seldom indications of disagreement. However, when some facilitators indicated having ambitions on behalf of the practices about certain topics or levels of progress, others responded that the facilitator’s ambition was not the practice’s, suggesting an undertone that the ambition was perhaps inappropriate. Nevertheless, disapproval was not explicitly stated. In an individual interview, one facilitator expressed that even though she personally appreciated that she could influence the facilitation as she pleased, she found it problematic that there was no uniformity between the facilitators, and that they predominately drew upon their own practices and experiences. In her opinion, this meant that the offer provided to the practices was too dependent on which facilitator was allocated to the practice. She also expressed this view at the final evaluative network meeting, however the other facilitators did not appear to reflect or comment on her observation.

A facilitator driven vs. a practice driven process A practice driven approach was a core component in the intervention design. The facilitators expressed in the interviews that they considered it important that the process was practice driven, i.e. that the practices should choose the focus and topics they found most relevant to ensure motivation and ownership. In general, the facilitators asked the practices which topics they wanted

37

to focus on, the practices chose the overall topics, and during the visits questions from the practice to the facilitator often steered the direction of the visit.

However, there were also examples in both interviews and observations where the process appeared to be more facilitator driven. In the interviews, the facilitators described the first visit as an introduction and a way for the facilitator to “scout the ground” and encourage the practice to choose a focus area. This act of “scouting the ground” was also evident in the observations where, without a formal structure, the facilitators asked questions about the practices’ status and knowledge in various areas and suggested topics based on the answers. In the interviews, the facilitators described this process as trying to catch the practice’s interest by using “appetizers”, and underlined that they merely put forward suggestions which they did not argue for. However, this did not always seem to be the case at the observed visits where practices tended to agree on the topics suggested by the facilitators, and facilitators occasionally added minor topics without explicitly having clarified the practices’ interest in them beforehand. Additionally, several facilitators mentioned that there were topics they always introduced during the visits, because either they or the project managers believed them to be important. As mentioned some facilitators also expressed having ambitions on behalf of the practices on certain topics or on the level of progress.

Another example of a seemingly more facilitator driven approach was when the practices chose the overall topic, e.g. the DCM or chronic care procedures for either COPD or diabetes. Having made the choice, it was generally not clarified what they wanted to focus on within the given topic, what their prior knowledge was, or their preferences regarding the structure of the visits, and the balance between different facilitator roles. Hence, the facilitator tended to decide on both the specific content within the topics and the facilitation approach. In the interviews, the facilitators said that they perceived this type of initial clarification as unnecessary, since by “testing the waters” they sensed what was relevant. This approach tends towards a more facilitator driven process. The content was, however, influenced by the questions the practice asked during the visits, and this was especially apparent during the hands-on facilitation approach.

Additionally, part of the role of process manager could be seen to coincide with a more facilitator driven process. The facilitators sought to ensure task definition and completion, they took notes on agreed upon tasks and summed these up in the visit report. They occasionally contacted practices prior to a visit to remind them of the upcoming visit, the tasks, and the agenda; and some contacted practices that had not scheduled or had cancelled their subsequent visit. The observed practices on the other hand were prepared in a quite limited way prior to the first visit and did not take notes during the visits.

38

Paper 2. Influences of peer facilitation in general practice – a qualitative study (Appendix 4)

Purpose The purpose was to explore facilitation from the recipients’ (i.e. general practice) perspective to gain a more detailed understanding of how peer facilitation influenced practices and how they valued the facilitation.

Method The data were 20 practice interviews conducted after the first and the last facilitation visit and observations of 30 facilitation visits.

Findings The practices’ expectations prior to the visits were generally vague and their understanding of the intervention and of facilitation was limited. Further, they did not appear to experience a strong need or motivation for change. The study also showed that the facilitation concept was understood differently by different practices, and therefore they had diverse expectations and wishes for the facilitation approach. Hence, although most were pleased with the content of the visits, a few were disappointed, mainly because they had expected a more coaching based facilitation approach.

The changes after the visits were mainly related to increased use of diagnosis coding and installation, improved set-up, and registration of the DCM. None came as far as using the DCM for quality improvement. Two practices had written new chronic care procedures and a few expressed increased attention towards annual chronic disease check-ups and municipal chronic care activities. Some did not have any tangible changes, or explicitly stated limited or no impact from the visits.

The practices were influenced by the facilitation visits in a variety of ways. That the facilitators were peers generally induced a perception of credibility and a sense of relevance, trustworthiness, and transferability of the facilitators’ descriptions. The practices generally experienced that the visits influenced their knowledge, awareness and skills, and their perceptions of the usefulness and manageability of the area of change. In general, the practices described the content of the visits as relevant due to the self-chosen topics, visits being focused on their experienced challenges, and the hands-on approach to knowledge dissemination. Additionally, the visits also influenced the conditions for change in the practices by being an occasion to initiate change, providing protected time, ensuring task definition and delegation, and a sense of deadlines due to the recurrent visits. Although the practices were influenced in various ways, there were examples of redundant, inadequate or forgotten knowledge, and limitations in some practices’ motivation to change after the visits. Additionally, contextual conditions like IT challenges and limited time during the working day were experienced as limiting the change process.

39

Paper 3. The effectiveness of a semi-tailored facilitator-based intervention to optimise chronic care management in general practice: a stepped-wedge randomised controlled trial (Published in BMC Family Practice 2014;15:65) (Appendix 5)

Purpose The purpose of the study was to examine the effectiveness a semi-tailored, facilitator-based intervention developed by the Capital Region of Denmark to support the implementation in general practice of CDMP for COPD and type-2-diabetes.

Method The study was a stepped-wedge, randomised, controlled trial among general practices in the Capital Region of Denmark. Practices were allocated to either an intervention group or a delayed intervention group (control group). The data were retrieved from registries and questionnaires.

Findings We randomised 189 general practices; 96 practices were allocated to the intervention group and 93 to the delayed intervention group. Six practices dropped out of the study due to retirement and were excluded from the analysis, and 12 practices did not respond to the questionnaires and were only represented in the registry data. For the primary outcome, we analysed respectively 94 and 89 practices in the two groups.

Of the 96 practices in the intervention group, 24% received all three hours of facilitation outlined in the design of the intervention, 39% received two hours, 29% received one hour, and 9% had no facilitation visit. With a few exceptions, all outcome measures improved from baseline to follow-up in both allocation groups. The use of annual chronic disease check-ups (the primary outcome) increased in both allocation groups (72% and 55% respectively), but there was no significant difference between the groups at follow-up (p=0.2788) and no significant difference in the groups’ level of change from baseline to follow-up (p=0.1639). As for the secondary outcomes, there were no significant differences regarding the self-reported use of annual chronic disease check-ups for diabetes and COPD (p=0.2345; p=0.0787 respectively), the reduction in the number of practices with few annual chronic disease check-ups (p= 0.4403), or in the use of spirometry (p=0.0835). The self-reported use of ICPC diagnosis coding was significantly in favour of the intervention group for both diabetes (p=0.0050) and COPD (p=0.0243). The difference between the groups in the proportion of practices that reported always using ICPC diagnosis coding at follow-up was 20 percentage points for diabetes and 16 percentage points for COPD. The self-reported use of stratification was significantly in favour of the intervention group for COPD (p=0.0185) and there was a non-significant trend for diabetes (p=0.0598). The differences between the groups in the proportion of practices that reported always stratifying patients were 13 percentage points for COPD and 14 percentage points for diabetes. During the study period, the intervention group had a faster sign-up rate to the DCM; however, this did not result in a significant difference at the end of the study.

40

Discussion As illustrated in the thesis introduction the field of facilitation is complex and diverse. Conceptual clarity seems to be lacking due to no clear and consistent operational definition of facilitation and considerable variation in the content of facilitation interventions.

The purpose of the thesis was to explore facilitation visits as a method for implementation and practice development in general practice. By exploring in detail facilitation enactment, how recipients were influenced by the facilitation visits, and the effectiveness in a single intervention, I have worked to provide a more nuanced picture of what may lie beneath the label of facilitation.

The thesis provides a comprehensive insight into a specific intervention within the complex field of facilitation. It does not portray a fully representative picture, but offers an understanding of what facilitation can encompass and gives rise to some conceptual and methodological considerations. Hence, in the following discussion I will present:

- a summary and discussion of my findings in relation to current facilitation research - strengths and limitations of my studies - arguments for why the concept of facilitation needs to be discussed and clarified - a methodological discussion of the evaluation of facilitation interventions.

Summary of findings in relation to current facilitation research

Facilitation enactment In this study, facilitation was mainly enacted in four facilitator roles: the teacher, the super user, the peer, and the process manager. This meant that the facilitators provided factual knowledge either through presentations or by a hands-on approach. Further, they passed on experience-based knowledge for inspiration in the form of their own experienced benefits and their own practice organisation. Finally, they sought to ensure task definition and delegation at the end of the visits, and commitment and progress through status reports on the agreed upon tasks. The facilitators generally did not enact a fifth role – that of a coach, enabling collective reflections and internal discussions on the practices’ organisation and implementation during the visits. The facilitators primarily linked the limited use of this role to practices´ profound concrete needs and the amount of time available in a visit, but some also indicated that it was related to their own comfort zones, competences, and personalities. There were also indications of occasional problems with some of the other facilitator roles, e.g. limited knowledge of the specific patient record systems hindered the super user role, limited structure during the visits, and tasks not being clearly defined limited the process manager role. Other recent facilitation studies have found enacted activities and facilitator roles similar to those in this intervention and with a similar breadth in the types of activities [4, 9, 19, 20, 26, 31, 71]. However, this study elaborates on the content and balance of the specific facilitator roles, and on the challenges involved in managing diverse roles. The study also elaborates

41

on the enactment of facilitator roles when the facilitator is a peer, i.e. a GP facilitating in general practice.

I found that besides the previously identified variations between studies under the label of facilitation, there was also great flexibility and variation among facilitation visits in this single intervention, due to both practice and inter–facilitator variation. Hence, some variations in this study seemed linked to the tailored approach where the enacted facilitation is closely aligned to different practice needs. While other variations were associated with differences in how the facilitators interpreted their role, influenced by their facilitation education, their attitudes, comfort zones, and their own practice organisation. Other facilitation studies have likewise identified inter-facilitator variation besides those attributable to the intervention being tailored to the practices [4, 28]. However, the facilitators’ acceptance of this flexibility and inter-facilitator variability has not to my knowledge previously been identified.

The influence of facilitation visits With the enactment of facilitation and the contextual conditions of this intervention, practices were influenced by facilitation visits in several ways. The majority of the practices were pleased with the visits and the facilitation approach. However, it should be noted that most did not have clear expectations prior to the visits, and also they were compensated for lost income which might have influenced their level of satisfaction. However, a few practices were disappointed because they had expected more of a coaching based facilitation approach or because the visits lacked structure. The changes captured after the facilitation visits seemed more like initial change components rather than actual changed chronic care management. For example, sign-up for the DCM, improved data registration, and new procedures. However, the DCM was not used for quality improvement in any of the practices and there were only a few examples of actual use of the new procedures.

The practices generally experienced that the facilitation visits gave them increased knowledge, skills, and awareness. It also influenced their motivation and confidence to change due to changes in their perceptions of the usefulness of the selected areas of change and the manageability of the change process. Thus, it influenced their motivation to change. Besides being influenced by the content of facilitation visits, the study also revealed that an equally important influence from the visits was the provision of protected time, definition and delegation of tasks, and a sense of deadline due to recurrent visits. However, there were limitations to the knowledge, skills, and motivation to change after the visits in some practices. Other facilitation studies have likewise found that practices generally appreciated the visits and experienced that the facilitators ensured motivation and goal setting, gave advice and guidance in relation to specific problems, kept them focused with recurrent visits, and perceived visits as protected time from the demands of daily work life [9, 16, 17, 19, 25]. By offering a comprehensive exploration of how practices valued and were influenced by the visits and the facilitation, this study contributes to a more nuanced picture of how facilitation might be working, variations in how it is valued, and potential limitations.

42

The peer facilitator As described facilitators in primary care are often not GPs, but nurses or practice assistant. This thesis indicates the influence of the facilitators being peer GPs on facilitation enactment and how practices were influenced by and valued the facilitation visits. In the enactment, the role of the peer was evident in references to their common work life and to the facilitators’ practices and experienced benefits. Other studies have likewise found that health professionals bring their professional identity with them into new organisational roles within quality improvement programmes [104, 105]. The professional background of the facilitators in combination with the content and length of the facilitator education might also have influenced their basic comfort zones and competences. This seems to some extent to have resulted in facilitators being less experienced with a coaching based approach to organisational development and with the role of a process manager. Further, most practices appreciated that the facilitators were peers because this gave a sense of credibility and eased the process by ensuring a sense of usefulness and manageability. Another facilitation study also indicated the importance of the facilitator being a fellow GP [16]. However, studies with non-peer facilitators have likewise found that practices were satisfied with the visits [15, 24, 25]. Two studies that compared visits by a peer GP with visits by a non-peer within the same intervention indicated that facilitation by peer GPs was more effective. However, in both studies the two groups lacked comparability due to differences between the facilitators’ education in facilitation and their prior facilitation experience [4, 106]. Thus, some might perceive peer-based facilitation as being more credible and motivational, but since it is not compared with other professions in the same intervention, with a common purpose and context, this perception cannot be either confirmed or rejected.

The effectiveness of the intervention Several RCTs and the meta-analysis of practice facilitation have found facilitation to be effective. Contrary to this, the RCT of this intervention, like some other studies, found mixed results. There was no significant difference between allocation groups regarding the primary outcome, but significant differences regarding some of the secondary outcomes. Further, there were general improvements in both allocation groups. This would appear to indicate that this particular facilitation intervention enacted in a context with an existing focus on chronic care management, simultaneous interventions, and mandatory requirements did not add substantially to a change in chronic care management.

Additionally, the qualitative studies can to some degree illuminate these findings. First, the qualitative findings indicate that the scope of the outcome measures might have been too limited. Since both allocation groups improved in the number of annual chronic disease check-ups and in sign-ups to the DCM, and the qualitative studies identified challenges with the DCM, a need for assistance with setting it up correctly, and initial limited motivation, it could mean the practices that received facilitation visits have a better functioning DCM and use it more often for quality improvement because the facilitators ensured the set-up and demonstrated its usefulness. This could indicate that instead of only assessing the amount, it may have been appropriate also to assess the content and quality of the annual chronic disease check-ups and whether the DCM was correctly set-up and regularly used. Likewise, those who focused on new procedures for diabetes or COPD

43

may not have more annual chronic disease check-ups, but better ones. However, these content and quality measures might have been difficult to assess. For example, a difference in the use of the DCM would not have been measurable because it was hardly present at baseline.

Second, the qualitative studies also indicate issues in the facilitation process that might have influenced the effects measured in the RCT. Since both the DCM and chronic care procedures improve the overview of patients who need annual chronic disease check-ups, then likely limitations are the identified IT challenges, occasional long periods from the first to the last visit, and that practices did not get to use the DCM for quality improvement and only changed their procedures in a limited way within the three facilitation visits. Further, practices were not profoundly motivated for change; some did not have any tangible changes within the three visits, and some still experienced challenges after the last visit. However, the identified issues can either be an indication of the intervention being ineffective, or that the process of change might be longer than the follow-up time allowed in the RCT. Additionally, since we identified a sense of deadline linked to recurrent facilitation visits, this could indicate that more visits might have ensured additional change. Regarding the identified effect on increased diagnosis coding, the qualitative data showed that several practices increased their use of diagnosis coding within the three facilitation visits and this was often linked to visits where the topic was the DCM. Hence, this result in the RCT is likely to be a combination of increased diagnosis coding being attainable quite fast and that the DCM was the most profound topic covered during the facilitation.

Therefore, the limited effect measured in the RCT is likely due to a combination of contextual conditions (challenges and existing trends), the outcome measures, the facilitation processes, a small number of intended visits, and a tight follow-up time in the study (entailing only a fourth of the practices had all their visits and in opposition to a potential longer change process). However, though mixed methods were applied in the exploration of this intervention, a clear conclusion about what influenced the level of effect identified in the RCT is not possible.

The intervention context Based on all three studies in this thesis, contextual conditions are identified as something that need to be taken into account within the discipline of facilitation. First, contextual conditions appeared to influence the facilitation enactment in a more knowledge based and technical direction due to the DCM becoming mandatory and bringing with it the related IT challenges. Second, the contextual conditions seem to have affected how the practices were influenced by the visits and how they valued them. For instance, in this study the practices generally did not experience having either the time or the tradition for formal, collective quality improvement meetings. In another setting, such meetings might be more common, and the need for the facilitation visits to provide protected time and deadlines to ensure progress would in all likelihood be less pronounced. Similarly, in this study the practices in general had limited prior motivation to get behind the mandatory DCM which was the predominant topic during the facilitation visits. Here inspiration and knowledge became important. This would not be needed so much in a context where there is already some motivation to adopt new procedures, and only a need for support to develop them. Third, some contextual conditions like challenges with the DCM and limited time during the working day were

44

counterproductive in the facilitation process. Other facilitation studies have identified similar contextual conditions influencing the process, e.g. competing priorities and heavy clinical workload; areas with national attention, e.g. national targets with financially incentivised indicators; IT challenges; and a limited tradition for formal quality improvement meetings [9, 16, 23, 29, 67]. Hence, though specific conditions and their influence will vary across studies these examples of similarities also indicate that some types of influential contextual conditions are likely to be more common.

The RCT does not portray how context is influential, but it demonstrates that contextual conditions are something that needs to be addressed when assessing the effect of a facilitation intervention. That there were improvements in both allocation groups demonstrates that most changes identified in the intervention group were likely not attributable to the intervention. Chronic care management in general, and several of the components of the intervention in particular, were given profound attention in the health care sector and in general practice in the region during the same time period and this is likely to have influenced both the intervention group and the control group. For instance, there were courses on the DCM, courses on ICPC diagnosis coding, regional data consultants and the DCM became mandatory during the intervention period. On that basis, it is possible that the intervention simply only has limited effect in regards to the chosen outcome measures and in the given follow-up time. On the other hand, strong secular trends and simultaneous initiatives might have overruled a potential effect of the intervention or it is possible that practices in the control group made use of some of the simultaneous offers to assist them with the changes. This would imply that the effectiveness might have been different had the DCM not become mandatory, and had there not been simultaneous initiatives focusing on the handling of chronic care.

45

Strengths and limitations of my studies

Strengths This study explored a single intervention from different perspectives, and entailed methodological triangulation in the qualitative studies (focus groups, practice and facilitator interviews, and observations) and an RCT. This particular mixture of methods has not to my knowledge been combined in other facilitation studies.

The combination of different qualitative methods ensured a more detailed exploration of what lies beneath the label of facilitation in this intervention. The use of focus groups was beneficial because they enabled reflection among the facilitators of their different enactments and attitudes. Hence, this revealed differences and similarities among the facilitators. It also gave an insight into the group atmosphere, dialogue, and how inter-facilitator differences were handled. The individual facilitator interviews, on the other hand, provided details of individual facilitators’ rationales, experiences, and attitudes because each facilitator had more time and potentially expressed views that were not revealed in the focus groups.

It is a strength that the practice and facilitator interviews were conducted prospectively while the intervention was carried out, because it ensured better recollection by the participants. The observations, which are rarely used in facilitation studies, were used both to qualify the interview guides and as data in themselves, providing insight into behaviours and interactions.

In the qualitative studies, the selection of practices was made strategically to secure variation in relation to geography, practice type, and practice answers to the initial questionnaire. Hence, the practices included in this study should not be substantially different from the remaining practices that received the intervention. Further, because the practices were selected from both years of the intervention there was room to explore if the facilitation enactment was influenced by increased experience among the facilitators, changes in the context, or changes in the practices’ level of implementation of, for instance, the DCM.

As for the RCT, the fact that nearly all outcome measures improved from baseline to follow-up in both allocation groups demonstrates the value of the design, because it displays changes that are attributable to the intervention. If only before-after measurements had been applied, the effect would have been profoundly overestimated, because such measures would have indicated a difference that mostly would be due to existing trends and other influences. Further, in the RCT it is a strength that a large number of practices were included and there was a relatively low dropout rate. We also had the advantage of using administrative registers for the primary outcome measure.

Though mixed methods are beneficial, there is a risk that observations, questionnaires, and interviews in practices participating in an RCT might influence the behaviour and perceptions of the practices and hence, influence the results of the RCT. In this study, I minimised this risk by conducting the qualitative study among practices not participating in the RCT.

46

It is a strength for both the qualitative studies and the RCT that the intervention took place in a real-life setting, with real-life variations and challenges, because it illustrates how facilitation may be enacted and how it may influence the practices when the concept of facilitation is developed and managed by a health planning authority, and not controlled by researchers.

Limitations One limitation of the studies is that GPs and staff were interviewed together and this might have limited the staff’s ability to express conflicting opinions and the GPs’ ability to express insecurities and issues concerning the staff. Nevertheless, I deemed it important to give room for a dialogue about a common experience and to increase the number and variation of interviewees. Additionally, it is a limitation that not all practices were interviewed twice and that not all people that had participated in the facilitation visits within the individual practices also participated in the interviews.

I might not have reached data saturation in the observations, and this could be a limitation. I did not see the coaching role enacted in the observed visits, although a few facilitators did describe activities related to this role in other, unobserved, visits.

It might have been beneficial to have observed the facilitators’ education and interviewed the project managers before the intervention was initialised. This thesis focuses on facilitation enactment and inter-facilitator variation, but the available intervention documents describing the project managers’ intentions and the content of facilitation education were limited. Hence, such earlier interviews and observations of the facilitators’ education could have given better insight into the planned intervention, the project managers’ understanding of facilitation and the different facilitator roles, and likewise, how the facilitators were prepared to handle their roles. This might have enabled a better comparison between intentions and facilitation enactment and an elaboration of what influenced the enactment of facilitation.

The qualitative data collected in the practices might have been influenced by my presence during the visits and the first interview. If the facilitators and practices acted differently because they were observed, this could have influenced the enacted facilitation. Likewise, the practices’ perceptions of the visits might have been influenced by the interview after the first facilitation visit and could have led to reflections they otherwise would not have had. These reflections might also have influenced their behaviour during subsequent visits. However, I believed the benefits of prospectively following the practice with both interviews and observations outweighed this risk.

The studies provided an insight into the range of facilitator roles and common challenges plus a range of ways in which the practices were influenced by and valued the visits. However, because of multiple variations between practices, it was not possible to identify specific patterns connecting facilitation approaches, practice characteristics, topics at the visits, contextual conditions, how practices were influenced and valued the visits, and their subsequent changes. This would have provided a better understanding of facilitation enactment, how facilitation was working, and how facilitation could be matched to the level of change in the practices.

47

The practices in the RCT differ from the remaining practices in the intervention in one aspect, but otherwise seem comparable. The practices included in the RCT were the first to sign up to participate in the facilitation intervention, therefore they most likely represent practices that are more interested in making changes. This might have been an influence on the fact that both the intervention group and the control group improved during the intervention period. Further, it is possible that the limited effect of the intervention was due in part to some practices in the intervention group not receiving the intended number of facilitation visits, and some practices in the control group receiving facilitation visits during the intervention period. However, the per-protocol analysis only revealed minor changes in the results regarding some of the secondary outcome measures. As mentioned, due to the improvements in both allocation groups, additional outcome measures focusing on content and quality in the improved areas are likely to have been beneficial. However, some of these measures were not available at baseline. Likewise, it would have been beneficial to follow some of the practices in the control group to explore what made them change, but that could have risked interfering with their process and hence the results of the study. The timeframe of the data collection for the qualitative studies and the follow-up time for the RCT might have been too short to obtain concrete changes. The practice interviews in the qualitative study were conducted approximately a week after the last visit and this was perhaps too limited a time span. However, I chose this time frame to ensure better recollection from the respondents, and to increase the likelihood that the experienced changes were attributable to the visits, rather than to contextual conditions that occurred later.

Although I see the benefits of the study taking place in a real-life setting, a potential limitation is that not all practices received the intended number of visits, which influences the ability to be conclusive on the effect of the planned facilitation intervention. Further, studying an existing intervention meant that I did not have any influence on the intervention content, the timeframe, or the ability to ensure its implementation.

48

A need for discussion and clarification of the facilitation concept This thesis illuminates the need for additional discussion and clarification of the facilitation concept which seems required due to an absence of a unified conceptual definition, heterogeneity among interventions, multiplicity and limited content clarity concerning facilitator roles and activities, and a tailored approach being emphasised but not specified. In the following reflections, three areas in need of discussion and clarification are presented.

First, it seems that clarification is needed on what facilitation can and should encompass, and what unifies the facilitation concept. Facilitation often appears to be perceived as a unified concept, and it is questionable if that is meaningful given the diversity in the field. One option would be to agree that facilitation is not a single type of intervention, but is instead an overall label where the only unifying feature might be 1) practice visits, 2) support in identifying a need for change, and 3) assistance in the planning and implementation of change (which then can be done by various supporting activities). Otherwise, if an intervention should only be perceived as facilitation when it is conducted in specific ways, then exactly what facilitation should encompass needs to be discussed and clarified, as well as how it is differentiated from related concepts.

In this regard, one could question whether a single concept can meaningfully encompass substantially diverse facilitation approaches. Approaches like, for instance, provision of factual knowledge and collegial experiences, process management, and support of reflection and group discussion, as intended in this intervention, and the even greater range across interventions – from facilitators performing tasks and providing technical and practical assistance to ensuring critical reflection and empowerment. There needs to be clarity about whether facilitation is all of these diverse approaches, or whether some of them are preferred over others, whether all approaches should be present in each intervention, or if interventions labelled facilitation could differ on a sort of continuum.

Additionally, the absent enactment of the coaching role, and the dominant role of provision of knowledge and experience in the present study highlights a need to consider not merely which activities should be included within the facilitation concept (e.g. audits, consensus building, and goal setting), but also how these activities ought to be carried out for the intervention to be classed as facilitation. Further, given that contextual conditions are both influential and unavoidable in various ways when it comes to facilitation, the variations these conditions impose also need to be taken into account in future conceptual discussions about what unifies the facilitation concept.

Second, the idea of a facilitator being capable of handling diverse facilitation activities and roles needs to be considered. While it is intuitively appealing to have a continuum of facilitation purposes and roles, and a facilitator able to switch between these based on situational needs and tailored to the needs of the practice, my findings question the possibility of enacting this. Although the facilitators in my study linked the limited application of the coaching role to the very concrete needs of the practices, they also made a link with their own preferences and comfort zones. A study by Rhydderch et al. likewise identified challenges in performing certain aspects of the facilitator roles [27].

49

In support of the idea of a continuum of facilitator roles, one could argue in accordance with the revised PARIHS framework [2] that the facilitators in this study were novices and therefore perhaps not yet skilled in the whole continuum. Additionally, challenges with the continuum of roles could also be linked to the facilitators’ education. In the literature of facilitation in health care it is occasionally mentioned which facilitation skills are perceived as essential [13, 20, 27, 30, 61, 63]. Nevertheless, the content and length of the facilitators’ education is rarely described in current studies, and knowledge of what their education ought to contain also seems to be lacking [5, 59]. In this intervention, as in some others, the education of the facilitators was quite short [19, 24, 34], and although the coaching based approach was given attention it might not have been sufficient. Hence, it is possible that an expanded educational programme and supervision or mentoring of new facilitators could change and improve facilitators’ skills and range. It seems relevant at this point to consider whether facilitation is understood as a role that is quite easily obtained by a practicing health care professional through a fairly limited introduction to facilitation techniques and content knowledge, or if it is a role that needs persistent and specialised experience. However, while education and increasing experience in being a facilitator are likely to influence facilitators’ abilities to manage various and less familiar roles, even very experienced facilitators may not be able to move easily along a continuum of roles. This is because few people will have the knowledge (whether medical or technical), the experience, the process management skills, and the interactive facilitation skills required for fluid movement between these roles.

Third, the meaning and attainment of a tailored facilitation approach needs some elaboration. According to the present study, different practices likely understand the concept of facilitation differently; hence, they will have diverse expectations and wishes for the facilitation approach. Though most practices in this intervention were pleased with the visits, a few were disappointed mainly because they had expected another sort of facilitation than the one they received. Hence, the study highlights a need to explicitly discuss and match practices and facilitation approaches and have facilitators with diverse skills. A study by Watkins et al likewise discussed the need for an introductory session about purpose and rules of engagement [30]. Though tailoring is emphasised in the facilitation research literature, descriptions are lacking as to what tailoring more precisely entails, which aspects should be tailored, and how a tailored approach is achieved.

The findings in this study of both practice and facilitator driven elements in the process contribute with an additional perspective to this discussion: is tailored facilitation necessarily a practice-driven process, or is a more facilitator-driven process equally suitable? For instance, do the facilitators choose an approach in accordance with what they consider to be the practices’ needs; or do the practices choose both the topics they need support with and the facilitation approach, or should it be a sort of a mix? In this intervention, the facilitators perceived their approaches to be tailored to the needs of the practices. They expressed in the interviews that they judged the appropriate approach by “testing the waters”, and although the practices chose the topics, this was to various degrees influenced by the facilitators. This approach appears less practice-driven, but whether it is tailored to the practice could be debatable. On the other hand, it might also be questionable whether practices, if they were to decide on all parts of the process including the facilitation approach, would request what was most appropriate for them.

50

Challenges of complexity in the evaluation of facilitation interventions This thesis is a thorough exploration of a peer facilitation intervention in general practice where mixed methods were used to illuminate different aspects of the intervention. Besides providing insight into what may lie beneath the facilitation label, the studies also illuminated some methodological challenges that are relevant for evaluations of facilitation interventions in general. In the following, I will take a step back from the thesis’ contribution concerning the content and effect of facilitation and discuss these methodological challenges. Facilitation interventions can be characterised as complex interventions, and based on the guidelines from the UK Medical Research Council (MRC) [107] on how to evaluate complex interventions, I will discuss these challenges both more broadly and by using examples from my own studies.

Complex interventions are defined by the MRC as interventions characterised by the:

• Number of, and interactions between, components within the experimental and control interventions

• Number and difficulty of behaviours required by those delivering or receiving the intervention • Number of groups or organisational levels targeted by the intervention • Number and variability of outcomes • Degree of flexibility, or tailoring of the intervention, permitted

The complexity relates to a combination of the intervention components, its implementation, and its interaction with the context [108]. The MRC further describes that complex interventions are greater than the sum of their parts due to the components acting in synergy to produce change [108]. The sources of complexity in facilitation interventions lie in the various required facilitator skills, the numerous activities, the great flexibility and tailoring, the potential variations in enactment and outcomes, the influence of contextual conditions, and that numerous independent practices need to be targeted. Further, it adds to the complexity that facilitation involves continuous meetings between people. Hence, facilitation is not just about identifiable activities, but also depends upon interactions between people.

Randomised controlled trials The Medical Research Council emphasises the RCT as the best way to study the effect of complex interventions and recommends their use when possible. However, below I discuss some challenges in relation to the choice of outcome measures, study design, and the influence of contextual conditions.

Outcome measures The number and variability of potential outcomes in complex interventions poses challenges to the choice of outcome measures in RCTs. This issue is highly relevant for facilitation interventions, where the object of change varies from specific and concrete guidelines to more broad programmes of chronic care management or practice development. If the objective is assessing adherence to specific guidelines, this will often entail a large number of explicit guideline recommendations and also potentially some of varying importance. Additionally, if the intended change is as complex as

51

the CDMP in this thesis, there will be an even greater number of equally relevant outcome measures due to the broad focus.

To address the number and variability of outcomes the MRC guidelines suggest using more than one primary outcome and a few secondary outcomes. This is in contrast to traditional RCTs. MRC guidelines state:

A single primary outcome, and a small number of secondary outcomes, is the most straightforward from the point of view of statistical analysis. However, this may not represent the best use of the data, and may not provide an adequate assessment of the success or otherwise of an intervention which may have effects across a range of domains [107].

However, in the literature of complex interventions, there are disagreements about whether to hold onto the traditional focus on one primary outcome and few secondary outcomes, or to increase the number of outcomes [107, 109, 110].

Though it intuitively seems appealing to have several outcome measures, this is not without problems. For instance, how should the outcome measures be ranked and compared when concluding whether the intervention was effective if only some of them are significantly different after the intervention? How many outcome measures should be significantly different between the allocation groups, how big a difference should there be, and are some of them more important than others? In the RCT in this thesis, some of the self-reported secondary outcomes were significantly different between allocation groups, but the primary outcome measure was not. How should this be interpreted, if the outcomes were not ranked and if there were perhaps even more outcome measures?

Further, in an ordinary RCT, the power calculations are performed based on the primary outcome and hence, the power of the study might not be accurate for additional outcome measures. This is not a problem if the outcomes indicate an effect of the intervention, but if they do not, it might be because of a lack of power, and not because the intervention is ineffective. Hence, more outcome measures on an equal level of importance entail the need for a power calculation to include all outcomes, and this is likely to require studies that are much larger. Further, the power calculations become more complicated and a sort of weighting of the outcomes might be needed if they are not of equal importance.

Having multiple outcomes also increases the risk of type 1 errors with false positive significant differences between the allocation groups. The Bonferroni adjustment method is the multiplicity adjustment method most commonly used to minimise this risk in health care studies [111]. In this method the significance level of 0.05 is divided by the number of tested outcomes. However, due to the decreased significance level, this method increases the risk of type 2 errors (false negatives) hiding true differences between the groups. It also means that the interpretation of a single outcome measure varies depending on how many other tests are performed, because the significance level changes according to the number of outcomes [112]. Other, more sophisticated methods exist, e.g. the false discovery rate, where the level of significance is not lowered quite as much [111].

52

However, using a different adjustment method still does not solve the problem of how many outcomes should be different, how different they ought to be, and which ones are most important in order to declare that an intervention is successful.

With several outcome measures, there is also the risk of a trade-off effect between them, whereby some outcome measures decrease when others increase. This is especially problematic if some practices increase in one outcome and decrease in another, while others do the opposite. If this occurs then the mean effect will even out and hide improvements and harms caused by the intervention.

Additionally, the complexity caused by flexibility and tailoring in facilitation interventions also challenges the choice of outcomes, because it gives great room for variation in topics and activities across sites. Most likely, none of the potential outcomes will be relevant for all participating practices, because the individual practices probably only change within the area of their chosen topic. Therefore, assessing mean changes in an outcome across all participating practices might not reveal change if not enough practices have changed within each area.

To counter the issues of multiple outcomes, the use of composite outcomes in the assessment of complex interventions has been suggested [113, 114]. Some facilitation studies have included such composite outcome measures [43, 44, 53, 64] where a group of outcome measures are combined into a single measure, and this is tested to discover if participants in the intervention group improve in more areas than those in the control group. However, composite outcomes have been criticised due to the risk of them being misleading or misinterpreted. This is because the individual components might vary in their level of seriousness or be affected differently by the intervention. Further, that it is not clear whether some components are profoundly more influential than others, and that there is a risk of arbitrariness in the ways in which components are combined [64, 115, 116]. Some of these issues might be countered by giving a different weight to the different components of the composite outcome and by conducting sensitivity analyses.

In my study, the CDMP combined clinical and organisational guidelines where the latter are less well defined, and there is no clear guidance on what constitutes a successful implementation of the programmes. This intervention is complex both in intervention content and potential outcomes. Other facilitation studies are likely to face the same challenges, but more simple and unified outcomes might be relevant for facilitation interventions focused on more specific guidelines.

The stepped-wedge design The MRC suggests that a stepped-wedge design is a good option to overcome practical or ethical objections to experimentally evaluate a complex intervention. In this study, the stepped-wedge design was chosen because all practices in the region had to be offered the intervention. Hence, this design with the control group being a delayed intervention group was the only option for randomisation in this case.

However, the study illuminates some challenges in this design. First, the follow-up period in my study was quite short, because all practices needed to have the intervention within two years. This

53

might have reduced the level of outcome, because not all practices managed to fit all their facilitation visits in within the time period, and also some changes might not have been carried out until later on. Therefore, it seems important with this design that the chosen outcomes are either those that can be expected to occur within a quite short time span, or that the total time period of the whole intervention needs to be substantially longer. Second, the behaviour in the control group might be influenced by their delayed intervention. In my study for example, the control group might have postponed initiating planned changes, because they wanted to wait until the facilitator could assist them or, on the contrary, they might have implemented changes before receiving the intervention because the act of signing up for the facilitation intervention motivated them for change. Other researchers also describe the risk of control practices making anticipatory adjustments, and hence contaminating the control group [117]. Third, the stepped-wedge design where all groups eventually receive the intervention removes the possibility to study the sustainability of change or long-term effect in comparison to no intervention, because when all groups complete the intervention, there is no longer a control group in the study. On the plus side, besides the practical and ethical benefits of the stepped-wedge design, it is also likely that when the control group knows that they too will receive the intervention, it probably increases their willingness to set aside time for data collection.

Therefore, using a pure control group might increase the validity of the results and the types of effects that can be measured. Nevertheless, unless the study is purely based on routine registry data, there is always a risk that participation in a trial will influence the control practices, because if they are to provide access to data or fill out questionnaires, it might make them consciously or unconsciously more attentive towards this area. Hence, a stepped-wedge design entails both possibilities and challenges that need to be considered.

The context of the intervention RCTs are highlighted for their ability to assess changes attributable to an intervention by eliminating the influence of other conditions. However, RCTs are criticised by researchers within theory-driven evaluation (especially realistic evaluation) for seeking to eliminate the influence of contextual conditions [108, 118]. According to realistic evaluation it is not meaningful to talk about an intervention as either working or not, because it will not always work for everybody and in all contexts. From this perspective, outcomes are generated by interactions between the intervention and its context [119-121].

According to this criticism, the use of RCTs simplifies the understanding of effect and produces less relevant results [122]. The aggregated effect in an RCT is said to obscure rather than enlighten [123], and according to Seers “RCTs of complex interventions suggest a robustness of the results that may not exist in reality” [124]. This is because an RCT assesses an average effect of an intervention, and whether the intervention group on average improves more than a control group. It does not provide knowledge about how the intervention will work in another setting under other conditions [122].

54

The MRC guidelines reflect on this critique from the realistic evaluation perspective, and respond with: ”these arguments typically misrepresent the assumptions made by RCTs [....] Randomisation aims to ensure that there is no systematic difference between groups in terms of participant and contextual characteristics, reflecting acknowledgment that these factors influence intervention outcomes” [108]. Additionally, both the MRC and others criticise theory-driven evaluations for not being able to assess causal effects because by declining randomisation they do not have counterfactuals showing what would have happened without the intervention [107, 125]. The MRC guidelines state “In rejecting randomisation, and in common with other theory-driven methods, realistic evaluation is limited in its ability to disentangle events observed from what would have happened anyway” [107].

Nevertheless, the MRC guidelines recognise the need to consider the contextual influence and stress that the context in which the intervention takes place should be described because what works in one setting might be ineffective or harmful in other settings due to differences in the contextual conditions [107]. Blackwood et al. argue that even though the MRC advocates for the exploration of context the guidelines seem to focus on whether contextual factors promote or inhibit the intervention’s effectiveness, instead of seeing the effectiveness of an intervention interconnected to its context and the people involved. Blackwood et al. describe it this way:

While accepting that trials on their own may not be able to provide comprehensively definitive information, it maintains the traditional assumption that interventions stand or fall on the basis of their inherent quality, notwithstanding the context in which they are operationalised [121].

Contrary to this, researchers using realistic evaluation seek to identify “what works for whom in what circumstances” and to describe context-mechanism-outcome configurations. These are specific patterns of how contextual elements influence mechanisms17 that produce the identified outcome when the participants encounter the resources of the intervention [119, 120].

In this thesis the qualitative studies showed a contextual influence on the facilitation enactment, how the practices were influenced by the visits, and their changes. The intervention seemed to be working through its interaction with the context and the practices’ circumstances. Thus, these interactions provide important knowledge about the intervention and are not just something that needs to be adjusted for by randomisation and stratification. This also means that findings of average changes in the RCT have limited transferability to other settings with a different context. At the same time, the RCT also demonstrated the value of conducting a trial that exposes changes attributable to the intervention as opposed to secular trends and simultaneous interventions. Hence, it countered a risk of overestimating the influence of the facilitation visits.

The presence of secular trends in evaluations of complex interventions has been described as a rising tide phenomenon [126]. Chen et al. have described that when improvement happens in both allocation groups yielding a null result, it might be due to a general improvement because the issue

17 The concept of mechanisms is described on page 57 in relation to process evaluations.

55

has already gained widespread attention and there is pressure to handle it [126]. The authors therefore discussed the importance of assessing whether the improvement in both groups is due to contamination of the control group or a rising tide phenomenon, and they suggested means for doing so (e.g. similar improvement across the health care system, started to improve before trial, qualitative data in both allocation groups) [126]. A reason why it is important to consider the secular trends is that these might not be present in another context, and hence an intervention like the one I studied that did not add substantially to change might prove effective in another context where it is not competing with the same improvement initiatives coming from other sources. Further, the researchers also described that knowing the existence of such secular trends prior to data collection enables to possibility to make adjustments to the sample size [126].

Process evaluations The MRC guidelines recommend conducting process evaluations as a supplement to RCTs of complex interventions to assess implementation fidelity, mechanisms of impact, and influencing contextual conditions [107].

Fidelity studies Intervention fidelity describes whether an intervention is actually delivered in accordance with the planned intervention [127]. Fidelity studies are used to assess whether negative findings of an RCT are due to an ineffective intervention or an unsuccessful implementation (type 3 errors) [127]. Hence, knowledge about fidelity is important for both internal and external validity of the findings [128]. Fidelity assessment is described as an emerging science; however, there is conceptual ambiguity and a lack of assessment tools [128-130]. In my first qualitative study, the purpose was to explore facilitation enactment and elaborate on what may occur under the label of facilitation to contribute to the understanding of the concept. Hence, the purpose was not to assess implementation fidelity to the planned intervention. However, the study does raise issues concerning fidelity assessment.

In the intervention studied in this thesis, the project initiators and project managers had only loosely defined the intended facilitation activities, roles, and the content and structure of the visits. Because of this, and because facilitation is not unanimous or well defined in the research literature either, it is difficult to talk about fidelity both in relation to this specific intervention’s planned content and to a broader intervention concept. Nevertheless, even if future facilitation interventions more thoroughly describe these elements, other challenges remain that complicate fidelity assessments. One challenge is linked to the emphasis within facilitation on flexibility and tailored interventions, and another challenge is linked to the difficulty of assessing the fidelity of some elements due to the nature of facilitation.

Regarding the first challenge to fidelity assessment, complex interventions often need to make changes and adaptations during the implementation process [107], and this is contrary to the standardised interventions traditionally tested in RCTs. This is a challenge, because when an intervention is tailored to the individual participants and adjusted throughout its implementation, it is difficult to describe which specific intervention content is evaluated in the RCT and also to

56

distinguish between successful tailoring and poor fidelity. In contrast, if flexibility is important for the desired change to take place, then standardisation would inhibit change and make it less relevant and acceptable [117, 124]. Mackenzie et al concluded:

Currently there are no evaluation approaches that are fit for all purposes. Over-standardisation of complex interventions is in danger of delivering precise but invalid effect sizes, while approaches that aim to understand complexity can rarely give definitive answers about whether a complex intervention is effective at the population level [117].

Fidelity is often described as adherence to a manual [129]. However, the usefulness of a manual appears somewhat limited when taking into consideration the wide number of potential facilitation activities and roles that can be tailored in each practice. In a study by Mars et al [128] advocating fidelity assessment, the authors described how facilitators (in a self-management course for patients) deviating from the manual can be reinterpreted as something positive, if it is done in response to the participants. But they did not elaborate on how to determine if this was the case. For instance in my study, it was difficult to differentiate between variations due to differences between practices or due to inter-facilitator variation, where the latter could be perceived as limited fidelity. This was because the intended facilitation was not well defined and because patterns were difficult to identify. Further, it is difficult to say whether the lack of a coaching approach in general was linked to facilitation being tailored to the practice needs and preferences, or if it was a result of poor fidelity to the intended facilitator roles and the idea of a continuum.

Some researchers have suggested ways to handle the conflict between fidelity assessment and tailoring. One suggestion is to focus on an intervention’s essential elements, which are then to be given priority in fidelity assessment and kept fixed while other elements can be flexible [108, 127, 131]. However, there do not appear to be any guidance on how to differentiate between essential and non-essential elements and hence fixed and flexible elements. In a study by Byng et al., the authors described the fixed elements in a facilitator intervention as being the facilitators’ education, a payment to the practices, and a toolkit [67]. However, according to my study this seems far from adequate for a fidelity assessment because it would not assess the enacted facilitation. Hawe et al. [132] have suggested that instead of talking about standardisation in form (activities and content), one should assess fidelity according to function. They give an example that form would be giving the same information kit to all participants, whereas function would be that the information is tailored to local culture and learning styles. However, this distinction seems difficult to apply and the example of function seems at risk of getting too broad and not providing applicable guidance. One could argue that with tailoring being a core element in facilitation, what should be assessed in a fidelity assessment is whether the facilitators succeeded in tailoring the approach to the practices. However, as mentioned in the discussion of the facilitation concept, how to assess whether the approach is sufficiently tailored is a questionable topic.

Regarding the second challenge, fidelity primarily seems assessable for selected elements of facilitation interventions, for instance the number of participating practices, the number of visits to the practices, whether required information was provided, whether goals were chosen and tasks

57

were defined and delegated. However, while such issues are important, they are far from sufficient to assess fidelity. Based on my study it is equally important to explore the interactions between facilitator and practice, the facilitator roles and practices’ responses, and the way the different facilitation activities are conducted. However, these elements seem more difficult to assess. If fidelity assessments are attempted the data sources should also be considered. Self-reports from intervention implementers (in the case of facilitation it would be the facilitators) in the form of check lists are often used, but they are criticised for overestimating fidelity [129, 130, 133]. Other sources are participant questionnaires or audio or video recordings [133]. However, the use of questionnaires or checklists filled out by either facilitator or participants is of limited usefulness when it comes to identifying more complex intervention elements. Audio or video recordings are described as the gold standard for an objective verification of delivery [133], but while some components might be easily assessed in such recordings, e.g. provision of specific information and definition and delegation of tasks, other less uniform components, such as empathy and communication style, seem difficult to assess and to objectively verify. Further, if such recordings are to be used in fidelity assessments, a large number of practices would be required given the tailored intervention and variations between practices and facilitators. Together with the multiple visits in each practice, this would mean that the material gathered for a fidelity assessment likely is too large to thoroughly analyse. Schlosser has suggested that in studies of treatment fidelity 20-40% of sessions should be assessed [134]. If this algorithm was transferred to my study of 431 practices, it would mean a fidelity assessment on a minimum of 90 practices with 1-3 visits each.

Mechanisms of impact and contextual conditions For the assessment of mechanisms of impact and contextual conditions the MRC guidelines mention the use of theory-driven evaluation (theory-based evaluation and realistic evaluations) [108].

In theory-driven evaluation, mechanisms18 are described as the participants’ reasoning and reactions when encountering an intervention (their cognitive, affective, and social responses) that generate the outcomes. It is emphasised that mechanisms are something different than intervention activities or parameters identified in subgroup analyses. Mechanisms are understood as being sensitive to contextual variations and context is described as the conditions in which the intervention is introduced that influence the triggering of mechanisms and thus the outcomes [119, 120, 135, 136]. Context in realistic evaluation addresses issues of “for whom” and “in what circumstances” an intervention works, and it is emphasised that context is not just the locality of an intervention, but also systems of interpersonal and social relationships, biology, technology, and economic conditions [119, 135, 136].

18 Reviews of studies exploring mechanisms have identified variations in the understanding and use of the concept of mechanisms [135, 136]. The following description is based on a book and a paper by the founders of realistic evaluation Pawson and Tilley [119, 120], a concept paper about mechanisms [135] and a recent review of mechanisms from a realist approach [136].

58

Although the MRC guidelines recommend assessment of mechanisms of impact and contextual conditions there is no guidance on how to do this or how to use the findings in relation to an RCT. Some researchers have tried to combine the RCT with realistic evaluation [125, 137]. Bonell et al have described a model for “realistic randomised controlled trials” to capture both intervention effects and which intervention activities work, for whom, and under what circumstances [137]. However, whether and how it is meaningful to combine these approaches is debated in the research literature [118, 121, 137].

In the intervention studied in this thesis plausible mechanisms can be identified from the ways practices experienced being influenced by the facilitation visits. Hence, these were increased knowledge, awareness and skills, an increased sense of usefulness and manageability, and that the visits were an occasion to initiate change, they provided protected time, and increased the progress of change due to a sense of deadline. Additional plausible mechanisms were a sense of credibility and transferability due to the facilitator being a peer and a sense of relevance due to the attention given to specific challenges and a hands-on approach. I describe these as plausible mechanisms because they are not directly linked to the impact of the facilitation visits but are based on how the practices experienced being influenced by the visits. The study also identified a number of contextual conditions that are likely to have affected how the practices were influenced by the facilitation visits and also to have set limitations for the change process. This insight into how practices were influenced by the visits and the contextual conditions gives a more in-depth knowledge of how facilitation functions instead of merely assessing provided facilitation activities and outcomes. Thus, in accordance with the MRC guidelines, this study supports the viewpoint that exploring mechanisms and contextual conditions is important in facilitation studies, especially because contextual conditions and intervention design vary across facilitation interventions.

Though my study did not directly identify mechanisms, it reveals some likely challenges in identifying these within facilitation interventions. These challenges are difficulties identifying 1) level of change, 2) patterns between practices, and 3) potentially triggered, but inadequate mechanisms.

The first challenge concerns the assessment of the practices’ level of change, which is needed to identify patterns for mechanisms of impact. In my study, it proved difficult to qualitatively compare, and therefore categorise, the practices change levels. Four reasons for this are listed below and some of them are very likely to be relevant for other facilitation studies. First, the subject of implementation was the broadly focused CDMP, which contained both clinical and organisational components and no clear guidance on what constitutes a successful implementation. Second, due to tailoring in the choice of topic, the practices achieved changes in different areas, which makes it difficult to compare changes across practices. Third, the time of data collection can be a challenge in the assessment and comparison of the levels of change. At the time of the interviews in this study the changes appeared to be initial change components rather than fully changed chronic care practices. Paterson el al [138] describe that changes from complex interventions do not have a single endpoint. Hence, with change being a continuous process, these initial changes could be a first step towards future improvements and implementation of the CDMP, or it could be that the

59

practices may not improve any further. Fourth, because the RCT revealed changes in both allocation groups, indicating a secular trend, it is not possible to assess how much of the change in the observed practices is attributable to having had facilitation visits. Similar issues of assessing levels of change seem probable for other facilitation studies as well. However, assessment and ranking of practices’ levels of change is likely to be less complicated in interventions with more simple and concrete intended changes.

The second challenge concerns identifying clear patterns in the data. To identify mechanisms it is necessary to identify patterns of how practices differ, for instance, did practices that changed in specific ways have a mechanism in common; or did those that did not change have some particular contextual conditions in common. However, in this study, it was difficult to identify patterns between practice characteristics, enacted facilitation, choice of topics, facilitators, contextual conditions, and changes, or lack thereof, because of various combinations between and across these practices. Thus, it was not possible to identify why facilitation was enacted in specific manners in individual practices and to understand specific connections between enactment, mechanisms, contextual factors, and outcomes. This is likely also the case in other facilitation interventions, because the visits are offered to a wide spectrum of practices with diverse characteristics, and because of the emphasis on a tailored approach. Additionally, it is also likely that some mechanisms and differences between practices’ levels of change are more hidden and therefore difficult to identify. For instance, because facilitation is an interaction between people, it is likely that some of the variations are due to the chemistry between the practice and the facilitator. Another complicating element in identifying patterns is that some researchers have described issues around differentiating when something is a mechanism, an intervention activity, or a contextual condition [67, 69, 122, 136, 139, 140].

The third challenge in identifying mechanisms has to do with mechanisms often being described as either present or absent. This study indicates a possibility that mechanisms can be influenced without triggering change, which likely makes it more difficult to identify them. For instance, the practices experienced an increased sense of usefulness regarding the potential changes and increased manageability of the change process; hence, an increased motivation to change. However, there were examples of this not being sufficient to trigger and ensure substantial changes. Likewise, though knowledge, awareness, and skills were improved, it was at times deficient; they were not improved enough to bring on the changes. This is in line with Dalkin et al., who suggested that instead of talking about a mechanism having an on/off switch and either triggering or not triggering an outcome, the activation of mechanisms should be understood as a continuum [139]. They exemplify this by describing that individuals, to a varying degree, can feel confident, angry, or mistrustful which can lead to a graduation of outcomes [139]. In addition to this, it is also reasonable that some mechanisms are only activated if specific concurrent mechanisms are present, which likely also makes it more difficult to identify them. Other studies have likewise mentioned a potential interaction between mechanisms which challenges their identification [69, 136]. Further, it seems likely that even though a mechanism is activated, contextual conditions might counteract its effect. This appeared to be the case in this study, where increased motivation to use of the DCM was counteracted by frustrations and misused time caused by IT challenges.

60

Hence, in the qualitative study in this thesis it was difficult to define a satisfactory outcome and categorise the practices accordingly, and this is most likely not straightforward in other facilitation studies either. Even if it is possible clearly to distinguish practices with high and low levels of change, there are likely, as in this study, numerous variations between practices that make it difficult to connect enactment, mechanisms, contextual conditions, and outcomes, and issues with partly triggered mechanisms.

Given these challenges the following three issues need to be considered in future studies of mechanisms of impact and contextual conditions, which might improve the ability to identify patterns. First, according to both theory driven evaluation literature and the MRC guidelines, it is important for the design, implementation, and evaluation of complex interventions to have an initial programme theory that explicitly combines activities, mechanisms, contextual conditions, and expected outcomes [108, 141]. A programme theory of anticipated patterns between these might increase the likelihood of identifying clear patterns in the data. However, it is important to be aware of potential unexpected mechanisms and contextual conditions so they are not overlooked. The MRC guidelines also recommend having a pilot phase in the development of complex interventions. In this intervention the pilot phase was used as part of the facilitator education and was used to test the comprehensibility of the questionnaires. If an extended pilot phase had been performed, it might have given an impression of facilitation enactment, likely mechanisms, and outcome measures relevant for the RCT. Thus, a pilot phase might have ensured a more focused trial and qualitative study. Therefore, this study supports these recommendations in the MRC guidelines.

Second, one could question if another study design might increase the possibility of identifying patterns of mechanisms of impact and contextual influences. In this study, an option could have been to explore practices that had changed according to the effectiveness measurement in the RCT, and retrospectively assessed differences between them. However, this would entail a risk of recollection bias in the interviews, observations would not have been possible, and all changes might not be attributable to the intervention because practices in the control group also improved. Another facilitation study that sought to identify reasons for success and failure by studying practices with a high and a low level of change did identify enabling and limiting conditions, but clear patterns of differences between those who had succeeded and those who had not were not evident [23]. Hence, I still perceive the design chosen in this thesis as suitable. Nevertheless, based on the difficulties establishing patterns in my study, it might be necessary to include more practices than usual in qualitative studies of facilitation, given the complexity and variability in mechanisms, conditions, and outcomes. This is not to quantify the data, but to ensure similarities in conditions across some practices to enable the identification of patterns. It might also be beneficial to explore processes in a control group if there is an RCT, to understand why they change their behaviour and thereby better understand the contribution of the facilitation intervention.

Third, the data sources in these studies need to be considered. To be able to identify connections between facilitation enactment, mechanisms, contextual conditions, and outcomes, it appears important to have comprehensive descriptions of all these elements within the practices and of variations between practices. Therefore, because of the strength and limitations of the different data

61

sources, a combination of interviews with facilitators and practices, and observations of facilitation visits, appears to be a preferable design. Further, though the experiences from practices and facilitators are valuable and necessary to identify patterns, it is important not to rely solely on their descriptions of, for instance, which facilitation elements contributed to change or of barriers to the change process. This approach would carry a risk of being based on assumptions or speculation. Hence, there should still be a focus on the wide range of data provided by the combined data sources, in order to keep one’s attention on the importance of identifying patterns based on variations between practices’ experiences, facilitation approaches, processes, and levels of change. Concurrent quantitative data appear useful to ensure a sort of standardisation in the assessment of outcomes across practices and perhaps some of the contextual conditions. In a recent paper, Pawson [141] stressed the use of both qualitative and quantitative methods in realist evaluations and criticised realistic evaluation studies that have used qualitative methods to identify outcomes. He emphasised that in his understanding outcomes have to be assessed with quantitative methods and that there ought to be baseline measures as well. This is needed to categorise who changed and who did not and to avoid personalised outcome descriptions and bias [141]. Additionally, practice satisfaction assessments, which are used in some facilitation studies [15, 24], should be used with some caution because it is unclear what they represent. For instance, they risk representing something that only in a limited way is related to changes attributable to the facilitation visits. In my qualitative study, most practices were satisfied with the visits, independently of whether they had made concrete changes or not. In this study, these statements might have been influenced by limited expectations prior to the visits, cost-neutral visits, and that the practices did not want to state discontent due to the facilitator being a colleague. Other studies have likewise found issues that seem to problematise the use of satisfaction assessments. A study found that despite practices appreciating the dialogue during visits and being very satisfied with their facilitator, they had only a moderate wish to participate again if they were given the option [16]. Another study found that practices felt their participation had been a good experience regardless of the results [23], and a third study found that despite practices having trouble describing specific outcomes, they expressed having benefitted from the visits [19].

62

Implications for research and practice Future facilitation studies have been called for concerning the impact of the number of visits, diverse facilitation approaches, facilitators’ professional background, essential facilitation elements etc. [5, 12, 19, 32, 58, 59]. However, it seems rather unlikely that studies of facilitation can contribute with findings on how best to organise facilitation interventions regarding the most appropriate facilitator professions, length of the intervention, facilitator roles, and facilitation activities. This is because comparisons across interventions, both qualitatively and quantitatively, seem highly limited due to the differences between interventions in their purposes, practice characteristics, facilitation enactment, and contextual conditions. Therefore, what works best is likely to be linked to the specific purpose and context of each intervention. The difficulties described in this discussion in identifying patterns in a single intervention due to various crisscrossing conditions and also the emphasis on tailoring, are likely to impede conclusions describing under which conditions specific facilitation variants are best suited.

A further complication lies with the meaningfulness of systematic reviews, especially meta-analyses, of facilitation which seem limited given that they aggregate and compare data from interventions with heterogeneous content and diverse contextual conditions. This could introduce the risk of oversimplified or misleading conclusions. For instance, the meta-analysis on practice facilitation [6] concluded that there is a moderate effect from facilitation, and that it is more effective when there are a larger number of visits, a tailored approach, and a fewer number of practices per facilitator. Although the authors mention the diversity between the interventions, such conclusions incorrectly give an impression of facilitation as a unified concept working independently of the purpose and context. Hence, at best, the usefulness of a systematic review and particularly a meta-analysis of facilitation is limited, and at worst, it might lead to erroneous translations among those who develop new interventions.

In addition one could question if additional RCTs of facilitation are meaningful due to the heterogeneity, methodological challenges, and limitations in transferability. It seems doubtful that additional RCTs will provide new knowledge to the field of facilitation, and they might be best suited for more evaluative purposes. Such an evaluative purpose would be to assess if a given intervention was appropriate and if it ought to continue in the same setting (here it is important to consider if the contextual conditions have changed). If RCTs are to be used, there is a need to address the methodological challenges and to measure the sustainability of the intended changes.

With qualitative studies of facilitation researchers can attempt to explore mechanisms, processes, barriers, and contextual influences. However, in this thesis firm results seemed difficult to obtain because the numerous variations across practices complicated identification of explanatory patterns, and this is likely to be a general phenomenon. There are, as discussed, some ways to improve the possibility of identifying patterns, but then the question is transferability across purposes and contextual conditions. Therefore, the request in the facilitation literature for more studies of the “black box” and facilitation processes etc. still seems relevant, but the use of the knowledge gained needs to be considered.

63

It seems likely that the broader lines of the findings from this thesis and other qualitative facilitation studies are transferable. These include how facilitation is enacted in a variety of facilitator roles, and how facilitation influences practices in various ways. Further broad findings include that intervention design, practice conditions, the facilitators’ professional background and contextual conditions influence the facilitation enactment (including the balance between facilitator roles), how practices are influenced by the visits, and the effectiveness of the intervention. Additionally, that variation between visits occurs, which likely is a mixture of tailoring to the different needs and preferences of the practices and inter-facilitator variation. Contrary to these broad lines, the more detailed findings of this thesis and other facilitation studies in the form of specific facilitator roles, conditions, ways of influencing, and outcomes can primarily be used as inspiration for future facilitation studies in term of study design, the development of programme theories, and in the analysis of the data. To enable this, future studies need to be well-described e.g. as to what is understood by facilitation, the intended and implemented intervention, the facilitators’ professional background and their education in facilitation, and also influential contextual conditions.

Even if these detailed and specific findings are mainly factors for inspiration, they are valuable because they illuminate what may occur under the label of facilitation. Therefore, at this point it still seems beneficial with more qualitative or mixed method studies of different interventions regarding enactment, mechanisms, and barriers. It might also be beneficial to explore delimited questions, for instance, comparisons of peer and non-peer facilitators within the same intervention, the content and length of the facilitators’ education, different levels of facilitator expertise and inter-facilitator supervision, and the sustainability of changes in practices. Although findings are likely to be linked to the intervention purpose, intended facilitation, and context, these studies can inspire the design of future interventions. Additionally, even if transferability is difficult to attain, it is conceivable that some findings may be common across several studies, which could indicate that these elements may be more likely to be expected, but still not necessarily present in future interventions. If broad lines and inspirational knowledge are the main elements of transferable knowledge attainable from facilitation studies then the question could be when enough research knowledge is actually enough.

It should be noted that given the heterogeneity in interventions labelled as facilitation it is possible that in some areas of the facilitation literature with more simple purposes and set-ups, the obtained knowledge will be more transferable across studies. However, when a facilitation intervention can be defined as simple is difficult to determine.

These implications for research also influence a range of considerations needed in the design and implementation of future facilitation interventions. It is important for the health planning authorities who develop facilitation interventions to understand that facilitation is not a unified concept, and to understand the limitations behind how it is studied. They need to be cautious about how they interpret and use conclusions from systematic reviews and meta-analyses, and should not expect that studies can contribute with findings on how best to organise facilitation. Hence, in the design of new interventions one should be cautious about directly transferring designs and findings from other interventions. The detailed and specific findings of both this thesis and other facilitation studies will predominately contribute with elements that need to be addressed in the design and

64

implementation of future facilitation interventions, rather than with recommendations for an appropriate design. Therefore, conducting a detailed formative evaluation while the intervention is implemented will be very important in future facilitation interventions. This will enable adjustment and optimisation of the intervention as it proceeds, which should be expected given that insight into needs and circumstances at the outset is likely to be incomplete. The following points ought to be addressed in the design and implementation stages in order to improve the design, monitoring, evaluation, and likely the effectiveness of the intervention:

• A practice needs assessment ought to be conducted in the local setting to promote a match between the needs of the practices and the intended facilitation.

• A programme theory connecting intended activities and expected mechanisms, outcomes, and influential contextual conditions should be clarified and followed up in the formative evaluation.

• A clarification of how facilitation is understood in the particular intervention, not just as activities, but also by which kind approaches these should be conducted.

• A thorough consideration of which facilitator roles seem best-suited, how to ensure them, and which facilitator roles might come into play in the given context and with the chosen facilitator profession.

• The content and length of the facilitators’ education, and how it is best matched to practice needs and the programme theory should be carefully considered.

• Consideration of whether some parts of the facilitation enactment should be fixed across practices, and how to handle potential inter-facilitator variation.

• It is important to be aware of diverse practice characteristics, conditions, and preferred processes, and hence to consider the intervention’s understanding of a tailored approach and how it can be achieved.

65

Conclusion Facilitation is a complex phenomenon on both a conceptual level, in practice, and as a research object. It is a diverse and multifaceted concept with limited conceptual clarity and great diversity between interventions. With a comprehensive insight into one specific intervention, this thesis elaborates on what may lie beneath the facilitation label, and on facilitation visits as a method for implementation and practice development in general practice. This was achieved through these three research questions:

1. How is facilitation enacted during facilitation visits? 2. How does facilitation influence the visited practices and how do they value the facilitation? 3. What is the effectiveness of the facilitation intervention?

The thesis emphasises how facilitation is enacted in a variety of facilitator roles, demonstrates the content of these roles and how they might be enacted, and shows that when a continuum of roles is intended some are more pronounced than others. Further, it demonstrates that the enactment of facilitation is dependent on the design, content, and context of the intervention as well as on the professional background of the facilitators. In this intervention facilitation was mainly enacted as the provision of factual knowledge either by presentations or a hands-on approach; as the provision of experience-based knowledge, and as process management. The facilitators generally did not enact coaching-based facilitation enabling collective reflections and internal discussions during the visits. This was linked to practices´ concrete needs and requests for concrete guidance, the amount of time available, and the facilitators’ competences. The thesis also shows that in addition to variations between facilitation interventions there are likely also variations within interventions due to approaches tailored to different practice needs and preferences, and inter-facilitator variation in the interpretation of the facilitation concept. Further, the thesis illustrates that the facilitation process itself is a mix of practice driven and facilitator driven elements.

This thesis demonstrates that facilitation visits seem to influence practices in diverse ways, which is linked in part to the enacted facilitation and in part to the protected time and attention ensured by the visits. In this intervention, influence was in the form of increased knowledge, awareness, and skills, changed perceptions of usefulness and manageability, time being set aside, task definition and delegation, and a sense of deadline due to recurrent visits. The ways in which practices are influenced by and value the facilitation visits also seem shaped by the facilitators’ professional background and the intervention’s contextual conditions. The practices described in this thesis vary in numerous dimensions, e.g. practice characteristics, motivations, and preferred type of facilitation, all of which influence their processes and how they value the visits.

Furthermore, mixed results and improvement in both allocation groups in the thesis’ RCT show that this particular intervention under these contextual conditions had little added effect on chronic care management. Hence, it underlines that chronic care management is an area with influential secular trends that need to be taken into account in the design and evaluation of facilitation interventions. Further, because the findings from the qualitative studies indicated potential limitations in the outcome measures due to contextual conditions, the follow-up time, the intervention design, and in

66

the facilitation process, it underlines that a pilot study and studies of the implementation process would be beneficial.

The field of facilitation is characterised by an absence of a unified conceptual definition, heterogeneity among interventions, multiplicity and limited content clarity in facilitator roles and facilitation activities, and a tailored approach being emphasised but not specified. Therefore, a discussion and clarification process seems to be required of the facilitation concept. This should entail what unifies the concept of facilitation and what it should encompass, the diversity of roles in which a facilitator should be skilled, the facilitator’s professional background, the content and length of facilitation education, and what a tailored approach precisely entails and how it is achieved.

Further, the thesis stresses that facilitation is a complex area to study. Though both RCTs and qualitative studies of facilitation have their strengths, there are several methodological challenges linked to the choice of outcome measures, a connection between the intervention and the context, assessment of fidelity, and difficulties identifying explanatory patterns in the data in regards to facilitation enactment, mechanisms, outcomes and contextual conditions. These methodological challenges need to be addressed by both researchers and those developing facilitation interventions.

The transferability of the results of facilitation studies to other settings and the comparability of different facilitation studies seems impeded by difficulties in identifying patterns as well as variations between purposes, settings, facilitation enactment, and contextual conditions across interventions. Hence, it seems unlikely that studies of facilitation can contribute with findings on how best to organise facilitation interventions, or that meta-analyses of facilitation can contribute with meaningful conclusions. However, the specific findings from individual facilitation studies (facilitator roles, mechanisms, and contextual influences) can be used as inspiration and areas for consideration in future facilitation interventions and research studies.

67

References References in the thesis 1. Foy R, Eccles MP, Jamtvedt G, Young J, Grimshaw JM, Baker R. What do we know about how to do audit and feedback? Pitfalls in applying evidence from a systematic review. BMC Health Serv Res. 2005;5:50. 2. Harvey G, Kitson A. Implementing evidence-based practice in healthcare: a facilitation guide. 1 ed. London: Routledge; 2015. 3. Fullard E, Fowler G, Gray M. Facilitating prevention in primary care. Br Med J (Clin Res Ed). 1984;289(6458):1585-7. 4. Petrova M, Dale J, Munday D, Koistinen J, Agarwal S, Lall R. The role and impact of facilitators in primary care: findings from the implementation of the Gold Standards Framework for palliative care. Fam Pract. 2010;27(1):38-47. 5. Dogherty EJ, Harrison MB, Graham ID. Facilitation as a role and process in achieving evidence-based practice in nursing: a focused review of concept and meaning. Worldviews Evid Based Nurs. 2010;7(2):76-89. 6. Baskerville NB, Liddy C, Hogg W. Systematic review and meta-analysis of practice facilitation within primary care settings. Ann Fam Med. 2012;10(1):63-74. 7. Bidassie B, Williams LS, Woodward-Hagg H, Matthias MS, Damush TM. Key components of external facilitation in an acute stroke quality improvement collaborative in the Veterans Health Administration. Implement Sci. 2015;10:69. 8. Kauth MR, Sullivan G, Blevins D, Cully JA, Landes RD, Said Q, et al. Employing external facilitation to implement cognitive behavioral therapy in VA clinics: a pilot study. Implement Sci. 2010;5:75. 9. Kotecha J, Han H, Green M, Russell G, Martin MI, Birtwhistle R. The role of the practice facilitators in Ontario primary healthcare quality improvement. BMC Fam Pract. 2015;16:93. 10. Liddy C, Laferriere D, Baskerville B, Dahrouge S, Knox L, Hogg W. An overview of practice facilitation programs in Canada: current perspectives and future directions. Healthc Policy. 2013;8(3):58-67. 11. Nagykaldi Z, Mold JW, Aspy CB. Practice facilitators: a review of the literature. Fam Med. 2005;37(8):581-8. 12. Stetler CB, Legro MW, Rycroft-Malone J, Bowman C, Curran G, Guihan M, et al. Role of "external facilitation" in implementation of research findings: a qualitative evaluation of facilitation experiences in the Veterans Health Administration. Implement Sci. 2006;1:23. 13. Thompson GN, Estabrooks CA, Degner LF. Clarifying the concepts in knowledge transfer: a literature review. J Adv Nurs. 2006;53(6):691-701. 14. Atkins EM, Duffy MC, Bain DJ. The practice characterization model: the importance of organizational life cycles and targeted interventions in general medical practice. Int J Health Plann Manage. 2001;16(2):125-38. 15. Baskerville NB, Hogg W, Lemelin J. Process evaluation of a tailored multifaceted approach to changing family physician practice patterns improving preventive care. J Fam Pract. 2001;50(3):W242-9. 16. Buch MS, Edwards A, Eriksson T. Participants' evaluation of a group-based organisational assessment tool in Danish general practice: the Maturity Matrix. Qual Prim Care. 2009;17(5):311-22. 17. Bunniss S, Gray F, Kelly D. Collective learning, change and improvement in health care: trialling a facilitated learning initiative with general practice teams. J Eval Clin Pract. 2012;18(3):630-6. 18. Burtonwood AM, Hocking PJ, Elwyn G. Joining them up: the challenges of organisational change in the professional politic of general practice. J Interprof Care. 2001;15(4):383-93. 19. Dalsgaard T, Kallerup H, Rosendal M. Outreach visits to improve dementia care in general practice: a qualitative study. Int J Qual Health Care. 2007;19(5):267-73. 20. Dogherty EJ, Harrison MB, Baker C, Graham ID. Following a natural experiment of guideline adaptation and early implementation: a mixed-methods study of facilitation. Implement Sci. 2012;7:9. 21. Dogherty EJ, Harrison MB, Graham ID, Vandyk AD, Keeping-Burke L. Turning knowledge into action at the point-of-care: the collective experience of nurses facilitating the implementation of evidence-based practice. Worldviews Evid Based Nurs. 2013;10(3):129-39.

68

22. Elwyn G, Hocking P, Burtonwood A, Harry K, Turner A. Learning to plan? A critical fiction about the facilitation of professional and practice development plans in primary care. J Interprof Care. 2002;16(4):349-58. 23. Hogg W, Baskerville N, Nykiforuk C, Mallen D. Improved preventive care in family practices with outreach facilitation: understanding success and failure. J Health Serv Res Policy. 2002;7(4):195-201. 24. Huston P, Hogg W, Martin C, Soto E, Newbury A. A process evaluation of an intervention to improve respiratory infection control practices in family physician offices. Can J Public Health. 2006;97(6):475-9. 25. Kousgaard MB, Thorsen T. Positive experiences with a specialist as facilitator in general practice. Dan Med J. 2012;59(6):A4443. 26. Liddy CE, Blazhko V, Dingwall M, Singh J, Hogg WE. Primary care quality improvement from a practice facilitator's perspective. BMC Fam Pract. 2014;15:23. 27. Rhydderch M, Edwards A, Marshall M, Elwyn G, Grol R. Developing a facilitation model to promote organisational development in primary care practices. BMC Fam Pract. 2006;7:38. 28. Shaw E, Looney A, Chase S, Navalekar R, Stello B, Lontok O, et al. 'In the Moment': An Analysis of Facilitator Impact During a Quality Improvement Process. Group Facil. 2010;10:4-16. 29. Tierney S, Kislov R, Deaton C. A qualitative study of a primary-care based intervention to improve the management of patients with heart failure: the dynamic relationship between facilitation and context. BMC Fam Pract. 2014;15:153. 30. Watkins C, Timm A, Gooberman-Hill R, Harvey I, Haines A, Donovan J. Factors affecting feasibility and acceptability of a practice-based educational intervention to support evidence-based prescribing: a qualitative study. Fam Pract. 2004;21(6):661-9. 31. Aspy CB, Enright M, Halstead L, Mold JW, Oklahoma Physicians Resource/Research N. Improving mammography screening using best practices and practice enhancement assistants: an Oklahoma Physicians Resource/Research Network (OKPRN) study. J Am Board Fam Med. 2008;21(4):326-33. 32. Dickinson WP, Dickinson LM, Nutting PA, Emsermann CB, Tutt B, Crabtree BF, et al. Practice facilitation to improve diabetes care in primary care: a report from the EPIC randomized clinical trial. Ann Fam Med. 2014;12(1):8-16. 33. Dietrich AJ, O'Connor GT, Keller A, Carney PA, Levy D, Whaley FS. Cancer: improving early detection and prevention. A community practice randomised trial. BMJ. 1992;304(6828):687-91. 34. Engels Y, van den Hombergh P, Mokkink H, van den Hoogen H, van den Bosch W, Grol R. The effects of a team-based continuous quality improvement intervention on the management of primary care: a randomised controlled trial. Br J Gen Pract. 2006;56(531):781-7. 35. Eriksson T, Siersma VD, Logstrup L, Buch MS, Elwyn G, Edwards A. Documenting organisational development in general practice using a group-based assessment method: the Maturity Matrix. Qual Saf Health Care. 2010;19(5):e37. 36. Frijling B, Hulscher ME, van Leest LA, Braspenning JC, van den Hoogen H, Drenthen AJ, et al. Multifaceted support to improve preventive cardiovascular care: a nationwide, controlled trial in general practice. Br J Gen Pract. 2003;53(497):934-41. 37. Frijling BD, Lobo CM, Hulscher ME, Akkermans RP, van Drenth BB, Prins A, et al. Intensive support to improve clinical decision making in cardiovascular care: a randomised controlled trial in general practice. Qual Saf Health Care. 2003;12(3):181-7. 38. Goodwin MA, Zyzanski SJ, Zronek S, Ruhe M, Weyer SM, Konrad N, et al. A clinical trial of tailored office systems for preventive service delivery. The Study to Enhance Prevention by Understanding Practice (STEP-UP). American journal of preventive medicine. 2001;21(1):20-8. 39. Hogg W, Lemelin J, Graham ID, Grimshaw J, Martin C, Moore L, et al. Improving prevention in primary care: evaluating the effectiveness of outreach facilitation. Fam Pract. 2008;25(1):40-8. 40. Hogg W, Lemelin J, Moroz I, Soto E, Russell G. Improving prevention in primary care: Evaluating the sustainability of outreach facilitation. Can Fam Physician. 2008;54(5):712-20. 41. Jaen CR, Ferrer RL, Miller WL, Palmer RF, Wood R, Davila M, et al. Patient outcomes at 26 months in the patient-centered medical home National Demonstration Project. Ann Fam Med. 2010;8 Suppl 1:S57-67; S92. 42. Kinsinger LS, Harris R, Qaqish B, Strecher V, Kaluzny A. Using an office system intervention to increase breast cancer screening. J Gen Intern Med. 1998;13(8):507-14.

69

43. Lemelin J, Hogg W, Baskerville N. Evidence to action: a tailored multifaceted approach to changing family physician practice patterns and improving preventive care. CMAJ : Canadian Medical Association journal = journal de l'Association medicale canadienne. 2001;164(6):757-63. 44. Lobo CM, Frijling BD, Hulscher ME, Bernsen RM, Braspenning JC, Grol RP, et al. Improving quality of organizing cardiovascular preventive care in general practice by outreach visitors: a randomized controlled trial. Prev Med. 2002;35(5):422-9. 45. Lobo CM, Frijling BD, Hulscher ME, Bernsen RM, Grol RP, Prins A, et al. Effect of a comprehensive intervention program targeting general practice staff on quality of life in patients at high cardiovascular risk: a randomized controlled trial. Qual Life Res. 2004;13(1):73-80. 46. Margolis PA, Lannon CM, Stuart JM, Fried BJ, Keyes-Elstein L, Moore DE, Jr. Practice based education to improve delivery systems for prevention in primary care: randomised trial. BMJ. 2004;328(7436):388. 47. Meropol SB, Schiltz NK, Sattar A, Stange KC, Nevar AH, Davey C, et al. Practice-tailored facilitation to improve pediatric preventive care delivery: a randomized trial. Pediatrics. 2014;133(6):e1664-75. 48. Modell M, Wonke B, Anionwu E, Khan M, Tai SS, Lloyd M, et al. A multidisciplinary approach for improving services in primary care: randomised controlled trial of screening for haemoglobin disorders. BMJ. 1998;317(7161):788-91. 49. Mold JW, Aspy CA, Nagykaldi Z, Oklahoma Physicians Resource/Research N. Implementation of evidence-based preventive services delivery processes in primary care: an Oklahoma Physicians Resource/Research Network (OKPRN) study. J Am Board Fam Med. 2008;21(4):334-44. 50. Nutting PA, Crabtree BF, Stewart EE, Miller WL, Palmer RF, Stange KC, et al. Effect of facilitation on practice outcomes in the National Demonstration Project model of the patient-centered medical home. Ann Fam Med. 2010;8 Suppl 1:S33-44; S92. 51. Parchman ML, Noel PH, Culler SD, Lanham HJ, Leykum LK, Romero RL, et al. A randomized trial of practice facilitation to improve the delivery of chronic illness care in primary care: initial and sustained effects. Implement Sci. 2013;8:93. 52. Roetzheim RG, Christman LK, Jacobsen PB, Schroeder J, Abdulla R, Hunter S. Long-term results from a randomized controlled trial to increase cancer screening among attendees of community health centers. Ann Fam Med. 2005;3(2):109-14. 53. Solberg LI, Kottke TE, Brekke ML. Will primary care clinics organize themselves to improve the delivery of preventive services? A randomized controlled trial. Prev Med. 1998;27(4):623-31. 54. Stange KC, Goodwin MA, Zyzanski SJ, Dietrich AJ. Sustainability of a practice-individualized preventive service delivery intervention. Am J Prev Med. 2003;25(4):296-300. 55. van Bruggen R, Gorter KJ, Stolk RP, Verhoeven RP, Rutten GE. Implementation of locally adapted guidelines on type 2 diabetes. Fam Pract. 2008;25(6):430-7. 56. Zwarenstein M, Bheekie A, Lombard C, Swingler G, Ehrlich R, Eccles M, et al. Educational outreach to general practitioners reduces children's asthma symptoms: a cluster randomised controlled trial. Implement Sci. 2007;2:30. 57. Burrows DE. Facilitation: a concept analysis. J Adv Nurs. 1997;25(2):396-404. 58. Grumbach K, Bainbridge E, Bodenheimer T. Facilitating improvement in primary care: the promise of practice coaching. Issue Brief (Commonw Fund). 2012;15:1-14. 59. Harvey G, Loftus-Hills A, Rycroft-Malone J, Titchen A, Kitson A, McCormack B, et al. Getting evidence into practice: the role and function of facilitation. J Adv Nurs. 2002;37(6):577-88. 60. Kitson A, Harvey G, McCormack B. Enabling the implementation of evidence based practice: a conceptual framework. Qual Health Care. 1998;7(3):149-58. 61. Taylor EF, Machta RM, Meyers DS, Genevro J, Peikes DN. Enhancing the primary care team to provide redesigned care: the roles of practice facilitators and care managers. Ann Fam Med. 2013;11(1):80-3. 62. Van Hoof TJ, Miller NE. Consequences of a Lack of Standardization of Continuing Education Terminology: The Case of Practice Facilitation and Educational Outreach. J Contin Educ Health Prof. 2014;34(1):83-6. 63. Laferriere D, Liddy C, Nash K, Hogg W. Navigating change: how outreach facilitators can help clinicians improve patient outcomes. J Am Board Fam Med. 2012;25(2):232-7.

70

64. Liddy C, Hogg W, Singh J, Taljaard M, Russell G, Deri Armstrong C, et al. A real-world stepped wedge cluster randomized trial of practice facilitation to improve cardiovascular care. Implement Sci. 2015;10(1):150. 65. Lobo CM, Euser L, Kamp J, Frijling BD, Severens JL, Hulscher ME, et al. Process evaluation of a multifaceted intervention to improve cardiovascular disease prevention in general practice. The European journal of general practice. 2003;9(3):77-83. 66. Byng R, Jones R, Leese M, Hamilton B, McCrone P, Craig T. Exploratory cluster randomised controlled trial of shared care development for long-term mental illness. Br J Gen Pract. 2004;54(501):259-66. 67. Byng R, Norman I, Redfern S, Jones R. Exposing the key functions of a complex intervention for shared care in mental health: case study of a process evaluation. BMC Health Serv Res. 2008;8:274. 68. Waterman H, Boaden R, Burey L, Howells B, Harvey G, Humphreys J, et al. Facilitating large-scale implementation of evidence based health care: insider accounts from a co-operative inquiry. BMC Health Serv Res. 2015;15:60. 69. Byng R, Norman I, Redfern S. Using Realistic Evaluation to Evaluate a Practice-level Intervention to Improve Primary Healthcare for Patients with Long-term Mental Illness. Evaluation. 2005;11(1):69-93. 70. Seers K, Cox K, Crichton NJ, Edwards RT, Eldh AC, Estabrooks CA, et al. FIRE (Facilitating Implementation of Research Evidence): a study protocol. Implement Sci. 2012;7:25. 71. Nutting PA, Crabtree BF, Miller WL, Stewart EE, Stange KC, Jaen CR. Journey to the patient-centered medical home: a qualitative analysis of the experiences of practices in the National Demonstration Project. Ann Fam Med. 2010;8 Suppl 1:S45-56; S92. 72. Cook R. Primary care. Facilitators: looking forward. Health Visit. 1994;67(12):434-5. 73. Manley K, McCormack B. Practice development: purpose, methodology, facilitation and evaluation. Nurs Crit Care. 2003;8(1):22-9. 74. Rycroft-Malone J. The PARIHS framework--a framework for guiding the implementation of evidence-based practice. J Nurs Care Qual. 2004;19(4):297-304. 75. Cheater FM, Hearnshaw H, Baker R, Keane M. Can a facilitated programme promote effective multidisciplinary audit in secondary care teams? An exploratory trial. Int J Nurs Stud. 2005;42(7):779-91. 76. Berta W, Cranley L, Dearing JW, Dogherty EJ, Squires JE, Estabrooks CA. Why (we think) facilitation works: insights from organizational learning theory. Implement Sci. 2015;10(1):141. 77. Nazareth I, Freemantle N, Duggan C, Mason J, Haines A. Evaluation of a complex intervention for changing professional behaviour: the Evidence Based Out Reach (EBOR) Trial. J Health Serv Res Policy. 2002;7(4):230-8. 78. O'Brien MA, Rogers S, Jamtvedt G, Oxman AD, Odgaard-Jensen J, Kristoffersen DT, et al. Educational outreach visits: effects on professional practice and health care outcomes. Cochrane Database Syst Rev. 2007(4):CD000409. 79. Qureshi N, Allen J, Hapgood R. A systematic review of educational outreach visits for non-prescribing interventions in general practice. European Journal of General Practice. 2002;8(1):31-6. 80. Williams L, Burton C, Rycroft-Malone J. What works: a realist evaluation case study of intermediaries in infection control practice. J Adv Nurs. 2013;69(4):915-26. 81. Locock L, Dopson S, Chambers D, Gabbay J. Understanding the role of opinion leaders in improving clinical effectiveness. Soc Sci Med. 2001;53(6):745-57. 82. Pedersen KM, Andersen JS, Sondergaard J. General practice and primary health care in Denmark. J Am Board Fam Med. 2012;25 Suppl 1:S34-8. 83. Christiansen T. Organization and financing of the Danish health care system. Health Policy. 2002;59(2):107-18. 84. Praksistælling 2012 [Practice count 2012].2016. www.laeger.dk/portal/page/portal/LAEGERDK/Laegerdk/P_L_O/Om%20PLO/Tal%20og%20publikationer/Statistik%20om%20almen%20praksis. Accessed 12 March 2016 85. Kronisk sygdom. Patient, sundhedsvæsen og samfund [Chronic disease. Patient, health service and sociaty]. Copenhagen: 2005. 86. Bodenheimer T, Wagner EH, Grumbach K. Improving primary care for patients with chronic illness. JAMA. 2002;288(14):1775-9.

71

87. Bodenheimer T, Wagner EH, Grumbach K. Improving primary care for patients with chronic illness: the chronic care model, Part 2. JAMA. 2002;288(15):1909-14. 88. Forløbsprogrammer for kronisk sygdom - Generisk model og forløbsprogram for diabetes [Disease management programmes - generic model and Disease management programmes for Type 2 Diabetes]. Copenhagen: The Danish Health and Medicines Authority, 2008. 89. CHRONIC DISEASE MANGEMENT. A NATIONAL STRATEGY. Disease Management Programmes and self-Management Support. The National Board of Health, 2007. 90. Forløbsprogram for Type 2 Diabetes - Hospitaler, almen praksis og kommunerne i Region Hovedstaden [Disease management programme For Type 2 Diabetes - Hospitals, general practice and municipalities in the Capital Region of Denmark]. Hilleroed: 2009. 91. Forløbsprogram for KOL - Hospitaler, almen praksis og kommunerne i Region Hovedstaden [Disease management programme for COPD - Hospitals, general practice and municipalities in the Capital Region of Denmark]. Hilleroed: 2009. 92. Schroll H, Christensen RD, Thomsen JL, Andersen M, Friborg S, Sondergaard J. The danish model for improvement of diabetes care in general practice: impact of automated collection and feedback of patient data. Int J Family Med. 2012;2012:208123. 93. Model for Improvement. 2016. http://www.nichq.org/about/expertise/improvement-science/model_for_improvement. Accessed 12 March 2016 94. Arlbjørn JS. Process optimization with simple means: the power of visualization. Industrial and Commercial Training. 2011;43(3):151-9. 95. Patton M. Qualitative Research & Evaluation Methods. 3 ed. United States of America: Sage publication; 2002. 96. Morgan DL. Reconsidering the Role of Interaction in Analyzing and Reporting Focus Groups. Qual Health Res. 2010;20(5):718-22. 97. Warr DJ. "It was fun... but we don't usually talk about these things": Analyzing sociable interaction in focus groups. Qualitative Inquiry. 2005;11(2):200-25. 98. Crist JD, Tanner CA. Interpretation/analysis methods in hermeneutic interpretive phenomenology. Nurs Res. 2003;52(3):202-5. 99. Dahlager L, Fredslund H. Hermeneutic analysis – understanding and preunderstanding. In: Vallgårda S, Koch L, editors. Research methods in public health. 1 ed. Copenhagen: Gyldendal Akademisk; 2008. p 159-184 100. Geanellos R. Exploring Ricoeur's hermeneutic theory of interpretation as a method of analysing research texts. Nurs Inq. 2000;7(2):112-9. 101. Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3(2):77-101. 102. Brown CA, Lilford RJ. The stepped wedge trial design: a systematic review. BMC Med Res Methodol. 2006;6:54. 103. Andersen JS, Olivarius Nde F, Krasnik A. The Danish National Health Service Register. Scand J Public Health. 2011;39(7 Suppl):34-7. 104. Braithwaite J. Between-group behaviour in health care: gaps, edges, boundaries, disconnections, weak ties, spaces and holes. A systematic review. BMC Health Serv Res. 2010;10:330. 105. Kousgaard MB, Joensen AS, Thorsen T. The challenges of boundary spanners in supporting inter-organizational collaboration in primary care - a qualitative study of general practitioners in a new role. BMC Fam Pract. 2015;16:17. 106. van den Hombergh P, Grol R, van den Hoogen HJ, van den Bosch WJ. Practice visits as a tool in quality improvement: mutual visits and feedback by peers compared with visits and feedback by non-physician observers. Qual Health Care. 1999;8(3):161-6. 107. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M, et al. Developing and evaluating complex interventions: the new Medical Research Council guidance. BMJ. 2008;337:a1655. 108. Moore GF, Audrey S, Barker M, Bond L, Bonell C, Hardeman W, et al. Process evaluation of complex interventions: Medical Research Council guidance. BMJ. 2015;350:h1258.

72

109. Lancaster GA, Campbell MJ, Eldridge S, Farrin A, Marchant M, Muller S, et al. Trials in primary care: statistical issues in the design, conduct and evaluation of complex interventions. Stat Methods Med Res. 2010;19(4):349-77. 110. Wallin L. Knowledge translation and implementation research in nursing. Int J Nurs Stud. 2009;46(4):576-87. 111. Glickman ME, Rao SR, Schultz MR. False discovery rate control is a recommended alternative to Bonferroni-type adjustments in health studies. J Clin Epidemiol. 2014;67(8):850-7. 112. Perneger TV. What's wrong with Bonferroni adjustments. BMJ. 1998;316(7139):1236-8. 113. Mayo NE, Scott S. Evaluating a complex intervention with a single outcome may not be a good idea: an example from a randomised trial of stroke case management. Age Ageing. 2011;40(6):718-24. 114. Watt H, Harris M, Noyes J, Whitaker R, Hoare Z, Edwards RT, et al. Development of a composite outcome score for a complex intervention - measuring the impact of Community Health Workers. Trials. 2015;16:107. 115. Cordoba G, Schwartz L, Woloshin S, Bae H, Gotzsche PC. Definition, reporting, and interpretation of composite outcomes in clinical trials: systematic review. BMJ. 2010;341:c3920. 116. Freemantle N, Calvert M, Wood J, Eastaugh J, Griffin C. Composite outcomes in randomized trials: greater precision but with greater uncertainty? JAMA. 2003;289(19):2554-9. 117. Mackenzie M, O'Donnell C, Halliday E, Sridharan S, Platt S. Do health improvement programmes fit with MRC guidance on evaluating complex interventions? BMJ. 2010;340:c185. 118. Marchal B, Westhorp G, Wong G, Van Belle S, Greenhalgh T, Kegels G, et al. Realist RCTs of complex interventions - an oxymoron. Soc Sci Med. 2013;94:124-8. 119. Pawson R, Tilley N, editors. Realist evaluation. DPRN thematic meeting 2006 report on evaluation; 2009. 120. Pawson R, Tilley N. Realistic evaluation. 1 ed. London: SAGE Publications; 1997. 121. Blackwood B, O'Halloran P, Porter S. On the problems of mixing RCTs with qualitative research: the case of the MRC framework for the evaluation of complex healthcare interventions. J Res Nurs. 2010;15(6):511-21. 122. Salter KL, Kothari A. Using realist evaluation to open the black box of knowledge translation: a state-of-the-art review. Implement Sci. 2014;9:115. 123. Shepperd S, Lewin S, Straus S, Clarke M, Eccles MP, Fitzpatrick R, et al. Can We Systematically Review Studies That Evaluate Complex Interventions? PLoS Med. 2009;6(8):e1000086. 124. Seers K. Evaluating complex interventions. Worldviews Evid Based Nurs. 2007;4(2):67-8. 125. Cook TD. The false choice between theory-based evaluation and experimentation. New Directions for Evaluation. 2000;2000(87):27-34. 126. Chen YF, Hemming K, Stevens AJ, Lilford RJ. Secular trends and evaluation of complex interventions: the rising tide phenomenon. BMJ Qual Saf. 2015. 127. Carroll C, Patterson M, Wood S, Booth A, Rick J, Balain S. A conceptual framework for implementation fidelity. IS. 2007;2. 128. Mars T, Ellard D, Carnes D, Homer K, Underwood M, Taylor SJ. Fidelity in complex behaviour change interventions: a standardised approach to evaluate intervention integrity. BMJ Open. 2013;3(11):e003555. 129. Cross WF, West JC. Examining implementer fidelity: conceptualizing and measuring adherence and competence. J Child Serv. 2011;6. 130. Slaughter SE, Hill JN, Snelgrove-Clarke E. What is the extent and quality of documentation and reporting of fidelity to implementation strategies: a scoping review. Implementation Science. 2015;10(1):1-12. 131. French SD, Green SE, Francis JJ, Buchbinder R, O'Connor DA, Grimshaw JM, et al. Evaluation of the fidelity of an interactive face-to-face educational intervention to improve general practitioner management of back pain. Bmj Open. 2015;5(7). 132. Hawe P, Shiell A, Riley T. Complex interventions: how "out of control" can a randomised controlled trial be? BMJ. 2004;328(7455):1561-3. 133. Borrelli B. The assessment, monitoring, and enhancement of treatment fidelity in public health clinical trials. J Public Health Dent. 2011;71.

73

134. Schlosser R. On the Importance of Being Earnest about Treatment Integrity. Augmentative and Alternative Communication. 2002;18(1):36-44. 135. Astbury B, Leeuw FL. Unpacking Black Boxes: Mechanisms and Theory Building in Evaluation. American Journal of Evaluation. 2010;31(3):363-81. 136. Lacouture A, Breton E, Guichard A, Ridde V. The concept of mechanism from a realist approach: a scoping review to facilitate its operationalization in public health program evaluation. Implement Sci. 2015;10:153. 137. Bonell C, Fletcher A, Morton M, Lorenc T, Moore L. Realist randomised controlled trials: a new approach to evaluating complex public health interventions. Soc Sci Med. 2012;75(12):2299-306. 138. Paterson C, Baarts C, Launso L, Verhoef MJ. Evaluating complex health interventions: a critical analysis of the 'outcomes' concept. BMC Complement Altern Med. 2009;9. 139. Dalkin SM, Greenhalgh J, Jones D, Cunningham B, Lhussier M. What's in a mechanism? Development of a key concept in realist evaluation. Implementation Science. 2015;10. 140. Marchal B, van Belle S, van Olmen J, Hoeree T, Kegels G. Is realist evaluation keeping its promise? A review of published empirical studies in the field of health systems research. Evaluation. 2012;18(2):192-212. 141. Pawson R. The science of evaluation : a realist manifesto. London: SAGE; 2013. xix, 216 sider p.

Additional references in the three papers 142. Dixon A. Motivation and Confidence: what does it take to change behaviour? . London: The King’s Fund; 2008. 143. Due TD, Thorsen T, Kousgaard MB, Siersma VD, Waldorff FB. The effectiveness of a semi-tailored facilitator-based intervention to optimise chronic care management in general practice: a stepped-wedge randomised controlled trial. BMC Fam Pract. 2014;15:65. 144. Due TD, Thorsen T, Waldorff FB, Kousgaard MB. Role enactment of facilitation in primary care– a qualitative study (unpublised). 145. Helfrich CD, Damschroder LJ, Hagedorn HJ, Daggett GS, Sahay A, Ritchie M, et al. A critical synthesis of literature on the promoting action on research implementation in health services (PARIHS) framework. Implement Sci. 2010;5:82. 146. Kitson AL, Rycroft-Malone J, Harvey G, McCormack B, Seers K, Titchen A. Evaluating the successful implementation of evidence into practice using the PARiHS framework: theoretical and practical challenges. Implement Sci. 2008;3:1. 147. Crotty M, Whitehead C, Rowett D, Halbert J, Weller D, Finucane P, et al. An outreach intervention to implement evidence based practice in residential care: a randomized controlled trial [ISRCTN67855475]. BMC Health Serv Res. 2004;4(1):6. 148. McCowan C, Neville RG, Crombie IK, Clark RA, Warner FC. The facilitator effect: results from a four-year follow-up of children with asthma. Br J Gen Pract. 1997;47(416):156-60. 149. Stevens SA, Cockburn J, Hirst S, Jolley D. An evaluation of educational outreach to general practitioners as part of a statewide cervical screening program. Am J Public Health. 1997;87(7):1177-81. 150. Dalsgaard T, Rosendal M. [Motivation and barriers to the use of facilitator visits in general practice]. Ugeskr Laeger. 2008;170(9):731-5. 151. Smidth M, Christensen MB, Olesen F, Vedsted P. Developing an active implementation model for a chronic disease management program. International journal of integrated care. 2013;13:e020. 152. Mdege ND, Man MS, Taylor Nee Brown CA, Torgerson DJ. Systematic review of stepped wedge cluster randomized trials shows that design is particularly used to evaluate interventions during routine implementation. J Clin Epidemiol. 2011;64(9):936-48. 153. Janssens I, De Meyere M, Habraken H, Soenen K, van Driel M, Christiaens T, et al. Barriers to academic detailers: a qualitative study in general practice. The European journal of general practice. 2005;11(2):59-63. 154. Kronikerkompasset evalueringsrapport [The chronic care compass an evaluation report]. Aarhus: The Research Unit for General Practice in Aarhus, 2012. 155. Improving Chronic Illness Care. 2013. www.improvingchroniccare.org. Accessed 2013

74

156. Praksisbesøg i Region Syddanmark [practice visits in Region Southern Denmark] 2013. www.regionsyddanmark.dk/wm306156. Accessed 2016. Changed webpage. 2016. www.sundhed.dk/sundhedsfaglig/praksisinformation/almen-praksis/syddanmark/konsulenthjaelp-til-praksis/sydpol/praksisbesog-hvad-er-det/. Accessed 12 March 2016

75

Appendices Appendix 1: Selection of definitions of facilitation from the literature

Appendix 2: Description of the Data Capture Module

Appendix 3: Paper 1

Appendix 4: Paper 2

Appendix 5: Paper 3

76

Appendix 1: Selection of definitions of facilitation from the literature

“Facilitation is a technique by which one person makes things easier for others. The term describes the type of support required to help people change their attitudes, habits, skills, ways of thinking, and working [……] facilitators are seen as people who make things easier, help others towards achieving particular goals, encourage others, and promote action.” [60]

“The facilitator role was described at times as: an agent of change; co-ordinator; a cross-pollinator of good ideas; a resource-provider; an information-giver; a trainer; researcher; adviser and mentor” [72]

“Practice facilitation is a multifaceted approach that involves skilled individuals who enable others, through a range of intervention components and approaches, to address the challenges in implementing evidence-based care guidelines within the primary care setting” [6]

“Facilitation is the process of providing support to individuals or groups to achieve beneficial change. It has been described as ‘the provision of opportunity, resources, encouragement and support for the group to succeed in achieving its own objectives and to do this through enabling the group to take control and responsibility for the way they proceed’. A facilitator has been defined as a ‘catalyst for change’, as someone who ‘helps forward’ and ‘gives direction, by drawing upon their own experience’. The facilitators’ tools of trade consist primarily of knowledge, skills and techniques for structuring and driving a process of change and occasionally expertise in the clinical area addressed by the intervention” [4]

“PF [practice facilitation] occurs when a trained facilitator provides support services to a primary care practice for an improvement initiative. The PF approach enables teams to overcome challenges encountered when implementing changes in the office setting by building their internal capacity to engage in redesign or improvement efforts [4]. Facilitators assist teams within the practice as they identify and prioritize areas of change as well as help them develop tailored action plans for improvement” [51]

“An OF [outreach facilitator] is a health care professional with expertise in organizational change management who can lead and support health care providers down the path of change. Using tools such as chart audits and feedback, OFs help practice physicians recognize challenges and identify goals. Then, by working collaboratively, the OF and the physicians can develop plans to accomplish these goals” [63]

“Facilitation is a goal-oriented, context-dependent social process for implementing new knowledge into practice or organizational routines. It typically involves individuals learning together in the context of a recognized need for improvement and supportive relationships. Effective communication and interactive problem solving are key process components” [76]

“The process by which a group is helped to achieve its purpose by a facilitator who promotes the improvement of team dynamics and the active involvement of all group members” [2]

“A facilitator is an individual who is skilled in working with the concepts of change management and individual and organisational development. Facilitation involves the facilitator working with individuals, teams, and organisations to prepare, guide, and support them through the implementation process.” [70]

“Practice development is a continuous process of improvement towards increased effectiveness in patient centred care. This is brought about by helping healthcare teams to develop their knowledge and skills and to transform the culture and context of care. It is enabled and supported by facilitators committed to systematic, rigorous continuous processes of emancipatory change that reflect the perspectives of service users.” [73]

“Facilitation is a goal-oriented dynamic process in which participants work together in an atmosphere of genuine mutual respect in order to learn through critical reflection” [57]

“the practice development literature is characterised by a central focus on critical reflection, experiential learning and changing practice cultures […]. The facilitator’s role, therefore, is to enable the process of reflective learning, with a particular emphasis on developing insights that can enable individuals and teams to transform themselves and factors within their organisational environment that may act as barriers to the implementation of change and improvement.” [2]

77

Appendix 2: Description of the Data Capture Module

The DCM is meant as a quality improvement tool. Most of the data is automatically collected from the lab scheme in the electronic patient record system (including prescriptions, laboratory tests, and information measured and entered by the GPs, but during annual chronic disease check-ups the GP is prompted to add additional data in a so called “pop-up”. In the lab scheme the data needs to be structured and specific International Union of Pure and Applied Chemistry (IUPAC) codes need to be installed and correlated to each item for it to be gathered by the system. All data is collected in a central database. The data is then sent back to the GP in the form of an overview of all patients registered with the practice with the given disease, comparisons with GPs in the municipality, region, and country, and continuous data on individual patients to share with the patient in the consultation. In the patient overview the GP can easily identify patients who have not had annual chronic disease check-ups or who have values inconsistent with clinical guidelines [92].

“Pop-up” window in an annual check-up consultation

An overview of patients with the given disease

Comparison with other GPs

Overview to present to an individual patient

78

79

Appendix 3: Paper 1

Role enactment of facilitation in primary care – a qualitative study

Tina Drud Due, Thorkil Thorsen, Frans Boch Waldorff & Marius Brostrøm Kousgaard

80

Role enactment of facilitation in primary care – a qualitative study

Tina Drud Due1, Thorkil Thorsen1, Frans Boch Waldorff1, 2 & Marius Brostrøm Kousgaard1

Affiliation: 1The Research Unit for General Practice and Section of General Practice, Department of Public Health, University of Copenhagen, Denmark. 2Research Unit for General Practice, Department of Public Health, University of Southern Denmark, Denmark.

Corresponding author:

Tina Drud Due

E-mail: tina.due@sund.ku.dk

Co-authors:

Thorkil Thorsen: tthorsen@sund.ku.dk Frans Boch Waldorff: fransw@sund.ku.dk Marius Brostrøm Kousgaard: marbro@sund.ku.dk

81

Abstract Background Facilitation is widely used as an implementation method in quality improvement. Reviews reveal a variety of understandings of facilitation and facilitator roles. Research suggests that facilitation interventions should be flexible and tailored, meaning that facilitation approaches and tools are accommodated to the particular needs and circumstances of the receiving organisations. However, the multiplicity of potential facilitator roles may place heavy demands on the competences of facilitators, and so there is a need to investigate how facilitation is enacted in specific interventions. The purpose of this study was to explore the enactments of external peer facilitation in general practice. Methods The facilitation intervention under study was conducted in general practice in the Capital Region of Denmark in order to support an overall strategy for implementing chronic disease management programmes. We observed 30 facilitation visits in 13 practice settings and had interviews and focus groups with facilitators. We applied an explorative approach in data collection and analysis, and conducted an inductive thematic analysis. Results The facilitators mainly enacted four facilitator roles: teacher, super user, peer and process manager. Thus, apart from trying to keep the process structured and focused the facilitators were engaged in didactic presentations and hands-on learning as they tried to pass on factual information and experienced based knowledge as well as their own enthusiasm towards implementing practice changes. While occasional challenges were observed with enacting these roles, more importantly we found that a coaching based role which was also envisioned in the intervention design was only sparsely enacted meaning that the facilitators did not enable substantial internal group discussions during their visits.

Conclusion Facilitation is a complex phenomenon both conceptually and in practice. This study complements existing research by showing how facilitation can be enacted in various ways and by suggesting that some facilitator roles are more likely to be enacted than others, depending on the context and intervention design and the professional background of the facilitators. This complexity requires caution when comparing and evaluating facilitation studies and highlights a need for precision and clarity about goals, roles, and competences when designing, conducting, and reporting facilitation interventions.

[Keywords: facilitation, facilitators, outreach visits, primary care, qualitative study, general practice]

82

Introduction Facilitation has become a widely used method for implementing quality improvements in health care [1-9]. In primary care facilitation involves an external facilitator, often with a health care background, who visits the practice in order to support a process of change [1, 8]. A recent meta-analysis reported that practice facilitation has a moderately robust effect on the uptake of guidelines in primary care settings [1]. However, the literature on facilitation interventions is marked by substantial variations in design with regard to a) the object of implementation (from relatively simple guidelines to more complex guidelines for chronic care and/or organisational development), b) intensity (duration and number of visits), c) the professional background of the facilitators, and d) their pre-defined roles and tasks. Thus, reviews of the literature have identified a variety of understandings of facilitation and of potential facilitator roles [3, 4, 9]. A continuum has been proposed which conceptualises facilitation as ranging from a goal and task oriented approach to a more holistic approach focusing on organisational development in a broader sense [4, 10-12]. In the goal oriented approach, clinical units are assisted by a facilitator who supports goal setting, provides factual knowledge (e.g. about guidelines), diffuses ideas between settings, and provides project management and technical expertise. In the holistic approach, the facilitator supports a more transformative and empowering change process based on internal discussions, critical reflection, and interpersonal relations [4]. Most interventions are perceived to encompass aspects of both approaches which suggest that facilitators are not necessarily fixed at one point in the continuum, but should be able to move along it depending on the situation [4]. In line with this thinking, subsequent contributions underline the importance of facilitation interventions being flexible and tailored, meaning that facilitation approaches and tools are accommodated to the particular needs, skills and circumstances of the receiving organisations [13, 14].

Since each of the various facilitation activities outlined above can be associated with particular competences, the diversity of potential facilitator roles appears to place heavy demands on the competence span of facilitators. Combined with the lack of a clear and consistent operational definition of facilitation [2, 5, 7] this diversity of potential roles creates a need for investigating how facilitation is actually enacted in specific interventions where a broad understanding of facilitation is adopted. This line of enquiry also fits well with previous calls for more qualitative research aimed at improving our understanding of facilitation as an implementation approach [13, 15, 16].

Thus, the purpose of the present study was to explore the enactment of facilitation in specific roles carried out by external peer facilitators as they sought to support the uptake of chronic disease management programmes in general practice.

The setting of the intervention Danish health care is mainly tax financed with free-of-charge access to general practice and public hospital services. General practitioners (GP) are private entrepreneurs mostly financed through the tax financed health care reimbursement scheme. Services provided by general practice are regulated by collective agreements between the Danish Regions and the Organisation of General Practitioners [17, 18]. Danish general practice is divided by 40 % partnership practices (co-owned by 2-4 GPs) and 60% solo-practices (of whom some collaborate and share facilities or practice staff) [19]. Practice staff consists of secretaries and nurses, and nurses increasingly conduct selected chronic care consultations (only GPs can perform annual chronic disease check-ups).

83

The disease management programmes Chronic disease management programmes based on the Chronic Care Model [20, 21] have been developed in all five regions of Denmark [22]. As guidelines these programmes outline evidence based treatment and a systematic approach to chronic care including a division of tasks between GPs, hospitals and municipalities for a given chronic disease. The programmes describe the GP’s role as coordinator of care and a systematic proactive approach with population based patient registration, annual chronic disease check-ups and stratification of patients into three levels by risk of complications, complexity, and state of the disease [23, 24]. Several initiatives have been launched to promote the on-going implementation of chronic disease management programmes and to improve chronic care management e.g. IT solutions, lectures, and inter-sectorial collaborations and coordinators. The facilitation project studied in this paper was one of the supportive initiatives in the Capital Region of Denmark.

The facilitation intervention The facilitation intervention was carried out from 2011-2012 in general practice in the Capital Region of Denmark. The overall goal of the intervention was to support the implementation of chronic disease management programmes for type-2-diabetes and chronic obstructive pulmonary disease (COPD) in general practice. The facilitators were 14 GPs who were hired on a consultancy basis. It was assumed by the designers of the intervention that this was critical for increasing the legitimacy of the intervention and gaining access to general practice. The facilitators’ educational programme consisted of a one weekend seminar and 10 three-hour meetings over four months. During this period they were updated on the central elements of the disease management programmes, the Data Capture Module (DCM) (see below), and introduced to various implementation and facilitation tools such as the Plan-Do-Study-Act (PDSA) circle [25] and the brown paper method (a visual display of a process with big post-it notes on the wall, where a practice actively focuses current and future workflows and division of tasks)[26]. The facilitators also participated in workgroups where they developed additional tools. Throughout the rest of the project period, network meetings were held approximately every third month where the facilitators discussed their experiences, had further education, and adjusted some of their tools. Each participating practice was offered up to three visits of one hour each. Visits were free of charge, the practice was compensated for lost income and participation in the intervention was voluntary. The potential topics to cover in the facilitation visits, outlined in the information material provided to the practices, were: • Workflow and division of tasks for chronic disease check-ups • International Classification of Primary Care (ICPC) diagnosis-coding • Patient stratification • Leadership and organisation • Collaboration with municipalities and hospitals • The role of GPs as coordinators of care • The DCM: a software program for quality development that may provide GPs with an overview of

patients’ conditions and treatments. Patient data is automatically collected from the GPs’ electronic health record system. Soon after the facilitation intervention had been introduced it became mandatory for all practices to sign-up to the DCM within two years.

84

During the facilitation visits period the facilitators were intended to act as catalysts for change by:

1. Providing information to practices about the chronic disease management programmes. 2. Engaging in dialogue with the practices about goals for development. In the information material to

the clinics it was also stated that gathering the whole clinic for a facilitation visit was an obvious opportunity to talk about work flows, tasks and common goals.

3. Assisting the practices to define specific objectives for the visits and to choose suitable means for achieving them.

4. Providing relevant tools and suggesting courses and other regional initiatives if needed. 5. Ensuring the structure of visits by managing agenda setting and time frames, and after each visit

providing the practice with a standardised visit report containing the topics discussed, the goals agreed upon and the task to complete in-between-visits.

The change process was to be driven by the motivation of the practices and based upon their interests and choices of topic. The intended facilitation approach was elaborated in interviews with project initiators and project managers, and they emphasised multiple facilitator roles and a continuum hereof. In order to support a tailored approach the facilitators were supposed to be flexible and the idea of a continuum of facilitator roles was a central element during the education of the facilitators. This continuum ranged from an expert/teacher role at one end to a coaching role at the other end, with the role of a sparring partner in between. Although the facilitators were not expected to be technical or disease specific experts, they were expected to master most of the continuum and to switch between roles according to the situation. The training material stated that the coaching approach was supposed to generate a “helping and focused conversation between two (or more) persons, where one by using open and focused questions and neutral formulations, gives the other/others the possibility to formulate problems/challenges and create possible solutions”. The intervention was developed and implemented by the Capital Region of Denmark and the Regional Unit for Quality Development and Continuing Education in General Practice. As external researchers, we were not part of either the design or the implementation of the intervention. Our study on the effectiveness of the intervention has been published elsewhere [27].

Methods An explorative approach was applied in the data collection and analysis, but we were inspired by the various facilitator roles and activities described in the literature, and an idea of a continuum of roles in both the intervention documents and the literature. This study is based on observations, focus groups and individual interviews. We observed 30 facilitation visits in 13 practice settings with one to three visits in each setting; (4 of the 13 were joined visits with collaborating practices, hence a total of 18 practices). The practices were strategically sampled to ensure variation in geography, practice size, and facilitators [28]. An overview of practices and facilitators is presented in Table 1 and Table 2. Extensive notes were written during the observed visits and the visits were audio recorded. Apart from serving as primary data, the observations were also used to qualify the interview guides for individual interviews and focus groups. We carried out individual, semi-structured interviews with seven facilitators who took part in the observed visits and we ran two concurrent focus groups with approximately half of all the facilitators in each group). The purpose of the individual interviews was to get an in-depth understanding of the facilitators’ behaviours and perceptions. The focus groups explored similarities and differences in the participants’ views, experiences and behaviours which are potentially more clearly illuminated when the participants have the opportunity to reflect on each other's statements [28, 29]. We also interviewed the two leaders who initiated the intervention and the two project managers to elaborate on the intended intervention. The themes in the interview guides are presented in Table 3.

85

We also interviewed the observed practices and our findings will be reported in a separate paper. All interviews and focus groups were audio recorded and transcribed verbatim after a transcription protocol. We used thematic analysis based on the approach of Braun and Clarke [30] for the interviews and focus groups which were inductively coded sentence by sentence. The codes were then grouped into themes and sub-themes, and the themes were related to each other and the whole data material and thus refined and connected. We used the NVivo software program in the coding and theme construction process. We analysed the notes and audio recordings from the observations in relation to the interview based themes.

Findings According to our observations, facilitation was almost exclusively enacted into four roles during the visits: the teacher, the super user, the peer and the process manager. These roles were combined in various ways during the visits with some roles being more pronounced than others. Although there were variations between the facilitators and the facilitation visits (e.g. in the structure and content of the visits and in the balance between roles during the visits) we primarily focused our attention on the similarities in the enactment of facilitation in terms of these roles. Below we present the various roles based on our observations and interviews. Table 4 presents further illustrations of the various roles. At this point we should mention that we found the role of coach described in the intervention design as absent in the observed visits. Therefore, we also present the reflections of the facilitators on the absence of this role.

The teacher In the observed visits, the facilitators communicated factual knowledge to the practice about central elements of the chronic disease management programmes (such as ICPC coding, annual chronic disease check-ups, stratification etc.), the DCM, and relevant websites on professional guidelines and municipal health services. The facilitators used more or less structured PowerPoint presentations (shown on PC or projector), speeches, demo versions of the DCM, or demonstrations of relevant websites. In this role the facilitator did most of the talking, but the participants asked questions and commented on the presentations. The presentations on the DCM mainly focused on its potential benefits as the facilitators tried to motivate the practices by providing a rationale for adoption. Practical issues and requirements were often quite randomly provided. Written instructions were generally not provided and notes were not taken by the practice during the visits.

In the interviews, the facilitators emphasised that disseminating information was less important than having the practice articulate their own ideas and questions. Nevertheless, they also noted that they spent more time teaching than expected:

To a large extent we do become teachers. You start out with the intention of doing some coaching… but then, when they sit down at the table after a busy day in the clinic, they mostly want some help to get started. And then you often end up teaching. I mean, you have to change between the roles but there is a lot of teaching, I think. [Facilitator C, focus group]

Several practices had difficulty setting up the DCM correctly in between the visits and experienced challenges with their IT-system providers. Hence, subsequent visits often focused on these problems concerning the DCM.

The super user As a supplement or alternative to didactic presentations, the facilitators often sat down with the GP in front of the clinic computer to provide more practical, hands-on assistance and guidance. For instance, the facilitator would demonstrate unfamiliar features in the patient record system, show how to use the DCM, discover errors in the set-up of the DCM, and talk about DCM data (e.g. coding percentages, missing annual chronic disease check-ups, and improvement thereof at subsequent visits). The practices were asked to contact their IT-system provider and to set-up the DCM between visits (a facilitator did this during one visit,

86

because the practice had failed to do so). Although the facilitators could provide some technical assistance, they did not see themselves as technical experts, and according to both interviews and observations they lacked knowledge on the patient record systems that differed from those used in their own practices. Compared to the teaching role, the super user role was more focused on specific practical problems, and the practice participants were more active in terms of asking questions and commenting on the issues at hand. According to the facilitators, the high prevalence of this role in the intervention was due to the concrete and basic needs of the practices, and although used in most practice types the facilitators deemed the super user role highly relevant in smaller practices. The facilitators also perceived that the hands-on approach created a closer relation to the daily tasks of the practice, which was important for the motivation of the participants:

So in this way things become very hands on like … this makes it more clear to them… when we look up one of their patients in the system and talk about this patient [using the system data] it makes more sense to most of them. [Facilitator A, individual interview]

The peer Most facilitators repeatedly emphasised their professional status as colleagues from general practice. They tended to use the expression “we” in conversation, indicating their common professional identities and working conditions. They emphasised the benefits of increased systematisation in their own clinic (in terms of reduced workload, improved patient care, increased job satisfaction, and better finances). In the interviews, the facilitators often stressed that meaningfulness and ownership were important to ensure change and that they wanted to inspire the practices to change by passing on their own enthusiasm regarding practice development. They often referred to their own experiences from general practice when describing ways of organising the clinic. In some cases such experiences were briefly mentioned during general topic presentations, in other cases the facilitators provided a more comprehensive description of their practice organisation in terms of structure, work division, annual chronic disease check-ups, and use of the DCM. In the interviews the facilitators said that this was meant as inspiration ensuring the practices did not have to “reinvent the wheel” during the change process. By referencing their own lack of perfection, their process experiences, their quickly obtained changes, and how they had overcome obstacles, the facilitators aspired to prevent the practices from seeing the change process as overwhelming. It was important for the facilitators that they were not perceived as representatives of the regional health authorities, but as colleagues who knew the business, since this would create a sense of trust and acceptance. Some believed that they could only help the practices because they had been through practice change processes themselves:

So of course I can use it [own experience]… When I sit there [at the visit] I can say ‘look, we didn’t do any diagnosis coding but then we actually changed and went to almost a 100% in a very short time, so it’s not as difficult and time consuming as it appears‘ [Facilitator A, individual interview]

A few facilitators, however, had reservations about using their own practice as an example, being concerned that the process would then become too dependent on the specific facilitator. Additionally, some facilitators emphasised that their position gave them the opportunity to share knowledge and ideas between the visited practices. However, this kind of knowledge sharing was rare in the observed visits.

The process manager It was the responsibility of the facilitators to help clinics with structuring the change process. The facilitators took on this role as process managers in relation to the following aspects:

a) Agenda setting: Before the first visit, the clinics were asked to fill out an online questionnaire on their current knowledge and activities in relation to the disease management programmes, and to make suggestions for topics to be addressed at the visit. Nearly all practices filled out the questionnaire, but less

87

than half suggested a topic. At the first visit, the facilitators asked the practices to suggest and choose the overall topics. The extent to which the facilitators guided the choice of topic varied, but they often decided on the specific content influenced by comments and questions from the practice. However, the practices existing level of knowledge and improvement needs within the overall topic were not always clarified and only the choice of topic was considered; the structure of the visit and the practices’ preferred style of facilitation were not discussed.

b) Structure: At the observed visits there was generally a low degree of structure, several topics and subtopics were covered in varying degrees of detail, either initiated by facilitators or by practices asking questions or telling stories about specific patients. There were variations in the degree of structure, the length of time focused on one topic and whether and how slide presentations were used. The facilitators said that they often had to secure the participation of practice staff. Several times at the observed visits they asked the staff direct questions or suggested they took charge of particular tasks. However, there were also visits where the dialogue was primarily between the facilitator and the GPs.

c) Promoting agreement on tasks: At the first visits, the facilitators did not attempt to get the practice to set an overall goal although this was a stated intention in the intervention design. Instead the facilitators looked to find agreement on more tangible in-between visit tasks. At some visits both specific tasks and the people responsible for them were agreed upon; at other visits only the tasks were identified, and sometimes a visit ended abruptly without clearly defining tasks. In most cases the choice of tasks was primarily influenced by the facilitator who suggested the logical next steps. Generally, the facilitator ensured the scheduling of the next visit, but the content was often not explicit. Few facilitators had contact with the practices in between visits.

d) Follow-up: In the subsequent visits the facilitators had the practices do a status report on the previously agreed tasks (e.g. whether procedures had been made, diagnosis coding had improved, or the DCM was used). Hence, the subsequent visits became a deadline and a way of ensuring commitment throughout the process. As one of the facilitators put it:

It makes it easier when someone comes from the outside… and helps to define goals and tasks. It makes progress easier. Because it makes you think ‘Oh now they come back, now we better start’. So it keeps them at it [Facilitator G, individual interview]

However, while status reporting was a means to keep up momentum it rarely fostered further discussions e.g. on implementation of new procedures.

Overall the facilitators tried to manage the facilitation process through agenda setting, task agreement, and follow up. While our observations pointed to occasional problems with these activities the facilitators did not articulate such problems in managing the process. Rather, the challenges they mentioned were not related to their own actions, but to influential contextual conditions (e.g. that larger practices could be more difficult to handle due to lack of time, or that some practices had members that did not attend meetings and delayed or rescheduled meetings).

The (absent) role of the coach According to intervention design, the facilitators were intended to engage in a coaching approach to help practices to articulate various problems and solutions related to the overall goals of the intervention. The project managers of the intervention also described the coaching approach as one where the facilitator helped to generate internal reflections and discussions between participants at the visits by asking reflective questions and encouraging dialogue about current and future practice.

88

However, at the observed visits, the facilitators did little to stimulate such reflective discussions; rather, they suggested having such discussions in-between the facilitation sessions. Generally, the facilitators tended to do most of the talking at the first visits. At subsequent visits the practices were more actively status reporting, talking about their challenges regarding the DCM, and asking follow up questions. But the internal dialogue between practice members was still very limited. Thus, current organisation was only superficially explored (although GPs and staff answered some questions from the facilitator on current practice), specific change techniques (such as PDSA-circles or brown paper methods) were not used, and there were few discussions of implementation plans and ways of using work division or the DCM for continuous improvement. In only one of the observed visits (practice 1) did a facilitator attempt to engage in a more coaching based approach with reflections about current practice (Table 4). However, it did not seem fully accomplished, because the level of internal discussion was minimal, and the process was not followed up upon at the subsequent visit.

In the interviews, the facilitators were asked to reflect on the limited use of the coaching approach. They mainly connected it to the practices having more concrete needs and requesting inspiration from the facilitators’ ways of organising. However, the facilitators did not explicitly clarify the practices’ expected or preferred facilitation approach. Further, some facilitators found that the practices were not used to considering “the bigger lines” on a more reflective level, but were mostly orientated towards immediate solutions to current problems. A few facilitators commented that they had come into the sessions with the intention of coaching, and momentarily did get into in a reflective mode, but the conversation quickly became more focused on practical problems due to the practices’ needs and expectations:

The way that GPs think is very much about handling problems. It’s what we do with patients and this is also how GPs think when they work with [practice] development. That’s why I think it could be interesting to create a more reflective space, to get the thoughts going ‘how are we really doing at the clinic? Is this the clinic we want to be? Are there other areas we should work with to make things more interesting, easier, or better?’ So more general talks and reflections, that is exciting, but this is not the way they are used to think because they work under time pressure and very practically with the patients [Facilitator C, individual interview]

Some of the facilitators also indicated that the limited use of the coaching role was because it was less familiar to them, somewhat outside their comfort zone and competences, and linked the use of this role to their personalities. The facilitators also felt that the intervention design had inhibited a more reflective approach, e.g. the visits took place during the work day which made it difficult for the participant to get into a more reflective mode, and the number of visits was too limited to leave time for more general discussions.

Although we did not observe the coaching role enacted during the visits, a few of the facilitators said that they did practice aspects of this role in other visits (e.g. by using reflective exercises and reflective questioning). Further, the facilitators also seemed to vary in their understanding of when a coaching role was enacted. For some it was described as enabling an internal discussion of practice procedures, while others seemed to link coaching to the practice deciding upon the topic, and the facilitator initiating a change process, asking open questions, passing on ideas, and having the practice consider their organisation in between the visits.

Discussion The purpose of this study was to explore the enactment of external peer facilitation in a complex intervention in general practice. We found that facilitation mainly took the form of four different facilitator roles during the sessions: the teacher, the super user, the peer, and the process manager. Thus, facilitation largely took the form of a) didactic presentations and hands-on learning where the facilitators used factual information and experienced based knowledge as well as their own enthusiasm for change and b) process

89

management activities around agreement on tasks and deadlines. While other studies of facilitation have also found such activities to be central to facilitators [7, 13, 15, 31, 32], this study elaborates on the content and balance of the specific roles adopted by the facilitators during interaction, and on the challenges involved in managing several roles. Thus the facilitators sometimes lacked technical knowledge, had problems with structuring the visits, and did not always manage to ensure a systematic definition of tasks and responsibilities. Perhaps more interestingly we found that the role of the coach envisioned in the intervention design was generally not enacted (in terms of enabling collective reflections and internal discussions at the visits). This resonates with the study by Rhydderch et al [16], who found that facilitators had challenges with generating team learning and constructive discussions on practice improvements. Several factors influenced the particular enactments of facilitation in this case. First, the professional background of the facilitators shaped their behaviour as they used their professional identity and experiences as GPs to provide knowledge and motivation throughout the process. This is in accordance with studies showing that health professionals bring their professional identity with them into new organisational roles created by various improvement programmes [33, 34]. Further, the facilitators did not have much formal training in the coaching based approach to organisational development which differs significantly from the more familiar activities of intra-professional knowledge exchange familiar to GPs. Second, the context of the intervention probably influenced the enactment of facilitation in a more knowledge based and technical direction (teacher and super user roles) as many visits came to focus on how to install and use the DCM, which became mandatory for all practices during the intervention period. This entailed profound IT challenges and a need for technical support. If such challenges had not been present, more resources could have been devoted to discussions about how to use patient data in quality development. Third, the design of the intervention (three visits) gave it a relatively low intensity compared to other studies of practice facilitation [1], leaving less room for enacting a more coaching based approach which often requires more time. Although the aim of this study was not to assess implementation fidelity as such [35] the limited enactment of the coaching role could be interpreted as a token of limited fidelity. However, it could also be argued that since the enactments of facilitation were often related to both contextual conditions and the stated needs of the participating practices, the enactments were loyal to the intervention’s emphasis on flexibility and tailoring. This suggests that it may be difficult to establish clear fidelity criteria in facilitation interventions with a strong focus on tailoring and where the nature of facilitation likely renders essential elements difficult to assess. In light of the variety of facilitation definitions presented in the literature, the idea of facilitation as a continuum of potential roles to be enacted in a tailored intervention, as well as the findings from this study, it seems appropriate to ask whether the single concept of ‘facilitation’ can meaningfully encompass such a wide range of roles and activities. Thus, although the idea of a continuum of facilitator roles which the facilitator is able to switch between according to the situational needs of the practice is intuitively appealing, it seemed difficult to realise in this intervention. It is important to note that researchers working with a facilitation continuum recently have developed their understanding of facilitation, suggesting that facilitators are categorised according to their experience: novice, experienced, and expert facilitators. In this scheme the novice is not capable of performing all facilitation approaches or roles and needs supervision by more experienced facilitators [5]. This differentiated understanding of facilitation is supported by the findings from this study where most of the facilitators could be perceived as novices, and therefore not yet skilled in mastering a wider range of roles. Even though the facilitators felt well-prepared, the educational programme for the facilitators in this project was quite brief which is not unusual in such interventions [15, 36, 37]. Therefore, it is possible that an expanded educational programme and on-going

90

guidance from experienced facilitators could improve facilitators’ skills and range. However, we would add that even very experienced facilitators may not be able to move easily along the continuum since few people will have the knowledge (whether medical or technical), the experience, the process management skills, and the interactive facilitation skills required to do this. Future facilitation interventions should explicate exactly how facilitation is understood in the particular intervention, not just in terms of activities, but also in terms of how facilitation ought to be conducted. The best match between facilitator roles to be enacted and the purpose and the context of the intervention should be considered, including how the facilitators’ profession and contextual conditions might influence enactment. Likewise, the content and length of the facilitators’ education and how it is best matched to the practices’ needs, the programme theory and the facilitators’ professional background should be carefully considered. Furthermore, future facilitation studies could explore how different enactments of facilitation and facilitator professions are related to goal achievement. Strengths and limitations of the study The facilitation intervention explored in this study framed a variety of facilitator roles and employed a relatively large number of facilitators. Therefore, as a case, the intervention was well suited to shed light on the enactment of different facilitator roles and the potential challenges related to broad framings of facilitation. It may be argued that the design of the intervention was somewhat naïve in assuming that the facilitators would manage to easily move along the continuum of facilitator roles based on their brief education. However, the optimal combination of knowledge, time and resources is rarely present when complex ideas are translated into practice in real life settings and as mentioned above the education of the facilitators in these kinds of interventions is often quite brief. It is a strength of the study that the intervention was explored by methodological triangulation using focus groups, individual interviews, and direct observations. Although direct observations produce detailed insights into the “black box” of facilitation and serve to counteract the bias generated when relying solely on post-hoc interviews with participants [28], only a few other studies of facilitation have made use of direct observations or audio recordings of visits [15, 16, 38]. It is a limitation of the study that we may not have reached data saturation concerning the observations. Thus, at the observed visits we did not see the coaching role enacted although a few facilitators told of activities related to this role in other visits. However, our findings on the limited enactment of this role were generally supported by the interviews and focus groups.

Conclusion In this study of facilitation in the context of implementing chronic disease management programmes in general practice, facilitation was enacted through different facilitator roles. The facilitators engaged in various forms of factual and experienced based knowledge transmission using their peer status as a source of inspiration and credibility, and supported the process by ensuring task and subsequent follow-up. They generally did not enact the coaching role defined by the intervention in terms of generating collective reflections on problems and improvements at the visits. There were also indications of occasional challenges regarding some of the other roles (e.g. limited technical knowledge, limited structure during the visits, tasks not defined).

Previous reviews have established that facilitation is a complex phenomenon in both theory and practice. Our results complements the existing literature by suggesting that facilitation is enacted in various ways and that some facilitator roles are more likely to be enacted than others depending on the design, content, and context of the intervention as well as on the professional background of the facilitators. This complexity calls for caution when comparing results from facilitation studies and points to a critical need for precision and clarity about goals, roles, and competences when designing, conducting, and reporting facilitation interventions.

91

Abbreviations DCM: Data Capture Module GP: General practitioner ICPC diagnosis coding: International classification of primary care diagnosis coding COPD: Chronic obstructive pulmonary disease Competing interests The authors declare that they have no competing interests. Authors’ contributions TDD designed the study, collected the data, analysed and interpreted the data, and drafted the manuscript. FBW supervised the study and participated in the design of the study, the interpretation of the data, and the drafting of the manuscript. TT supervised the study and participated in the design of the study, the collection of the data, the interpretation of the data, and the drafting of the manuscript. MBK supervised the study and participated in the design of the study, the interpretation of the data, and the drafting of the manuscript. All authors read and approved the final manuscript. Acknowledgments We thank the Danish Research Foundation for General Practice, The Health Foundation, and The Research Foundation for Primary Care in the Capital Region of Denmark for funding the study. Our thanks also go to the practices, facilitators and the project initiators and managers for participating in the study. References 1. Baskerville NB, Liddy C, Hogg W. Systematic review and meta-analysis of practice facilitation within primary care settings. Ann Fam Med. 2012;10(1):63-74. 2. Bidassie B, Williams LS, Woodward-Hagg H, Matthias MS, Damush TM. Key components of external facilitation in an acute stroke quality improvement collaborative in the Veterans Health Administration. Implement Sci. 2015;10:69. 3. Dogherty EJ, Harrison MB, Graham ID. Facilitation as a role and process in achieving evidence-based practice in nursing: a focused review of concept and meaning. Worldviews Evid Based Nurs. 2010;7(2):76-89. 4. Harvey G, Loftus-Hills A, Rycroft-Malone J, Titchen A, Kitson A, McCormack B, et al. Getting evidence into practice: the role and function of facilitation. J Adv Nurs. 2002;37(6):577-88. 5. Harvey G KA. Implementing evidence-based practice in healthcare: a facilitation guide. 1 ed. London: Routledge; 2015. 6. Kauth MR, Sullivan G, Blevins D, Cully JA, Landes RD, Said Q, et al. Employing external facilitation to implement cognitive behavioral therapy in VA clinics: a pilot study. Implement Sci. 2010;5:75. 7. Kotecha J, Han H, Green M, Russell G, Martin MI, Birtwhistle R. The role of the practice facilitators in Ontario primary healthcare quality improvement. BMC Fam Pract. 2015;16:93. 8. Liddy C, Laferriere D, Baskerville B, Dahrouge S, Knox L, Hogg W. An overview of practice facilitation programs in Canada: current perspectives and future directions. Healthc Policy. 2013;8(3):58-67. 9. Nagykaldi Z, Mold JW, Aspy CB. Practice facilitators: a review of the literature. Fam Med. 2005;37(8):581-8. 10. Helfrich CD, Damschroder LJ, Hagedorn HJ, Daggett GS, Sahay A, Ritchie M, et al. A critical synthesis of literature on the promoting action on research implementation in health services (PARIHS) framework. Implement Sci. 2010;5:82.

92

11. Kitson AL, Rycroft-Malone J, Harvey G, McCormack B, Seers K, Titchen A. Evaluating the successful implementation of evidence into practice using the PARiHS framework: theoretical and practical challenges. Implement Sci. 2008;3:1. 12. Rycroft-Malone J. The PARIHS framework--a framework for guiding the implementation of evidence-based practice. J Nurs Care Qual. 2004;19(4):297-304. 13. Liddy CE, Blazhko V, Dingwall M, Singh J, Hogg WE. Primary care quality improvement from a practice facilitator's perspective. BMC Fam Pract. 2014;15:23. 14. Stetler CB, Legro MW, Rycroft-Malone J, Bowman C, Curran G, Guihan M, et al. Role of "external facilitation" in implementation of research findings: a qualitative evaluation of facilitation experiences in the Veterans Health Administration. Implement Sci. 2006;1:23. 15. Dalsgaard T, Kallerup H, Rosendal M. Outreach visits to improve dementia care in general practice: a qualitative study. Int J Qual Health Care. 2007;19(5):267-73. 16. Rhydderch M, Edwards A, Marshall M, Elwyn G, Grol R. Developing a facilitation model to promote organisational development in primary care practices. BMC Fam Pract. 2006;7:38. 17. Pedersen KM, Andersen JS, Sondergaard J. General practice and primary health care in Denmark. J Am Board Fam Med. 2012;25 Suppl 1:S34-8. 18. Christiansen T. Organization and financing of the Danish health care system. Health Policy. 2002;59(2):107-18. 19. Praksistælling 2012 [Practice count 2012].2016. www.laeger.dk/portal/page/portal/LAEGERDK/Laegerdk/P_L_O/Om%20PLO/Tal%20og%20publikationer/Statistik%20om%20almen%20praksis. Accessed 12 March 2016 20. Bodenheimer T, Wagner EH, Grumbach K. Improving primary care for patients with chronic illness. JAMA. 2002;288(14):1775-9. 21. Bodenheimer T, Wagner EH, Grumbach K. Improving primary care for patients with chronic illness: the chronic care model, Part 2. JAMA. 2002;288(15):1909-14. 22. Forløbsprogrammer for kronisk sygdom - Generisk model og forløbsprogram for diabetes [Disease management programmes - generic model and Disease management programmes for Type 2 Diabetes]. Copenhagen: The Danish Health and Medicines Authority, 2008. 23. Forløbsprogram for Type 2 Diabetes - Hospitaler, almen praksis og kommunerne i Region Hovedstaden [Disease management programme For Type 2 Diabetes - Hospitals, general practice and municipalities in the Capital Region of Denmark]. Hilleroed: 2009. 24. Forløbsprogram for KOL - Hospitaler, almen praksis og kommunerne i Region Hovedstaden [Disease management programme for COPD - Hospitals, general practice and municipalities in the Capital Region of Denmark]. Hilleroed: 2009. 25. Model for Improvement. 2016. http://www.nichq.org/about/expertise/improvement-science/model_for_improvement. Accessed 12 March 2016 26. Arlbjørn JS. Process optimization with simple means: the power of visualization. Industrial and Commercial Training. 2011;43(3):151-9. 27. Due TD, Thorsen T, Kousgaard MB, Siersma VD, Waldorff FB. The effectiveness of a semi-tailored facilitator-based intervention to optimise chronic care management in general practice: a stepped-wedge randomised controlled trial. BMC Fam Pract. 2014;15:65. 28. Patton M. Qualitative Research & Evaluation Methods. 3 ed. United States of America: Sage publication; 2002. 29. Warr DJ. "It was fun... but we don't usually talk about these things": Analyzing sociable interaction in focus groups. Qualitative Inquiry. 2005;11(2):200-25. 30. Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3(2):77-101. 31. Dogherty EJ, Harrison MB, Baker C, Graham ID. Following a natural experiment of guideline adaptation and early implementation: a mixed-methods study of facilitation. Implement Sci. 2012;7:9. 32. Petrova M, Dale J, Munday D, Koistinen J, Agarwal S, Lall R. The role and impact of facilitators in primary care: findings from the implementation of the Gold Standards Framework for palliative care. Fam Pract. 2010;27(1):38-47. 33. Braithwaite J. Between-group behaviour in health care: gaps, edges, boundaries, disconnections, weak ties, spaces and holes. A systematic review. BMC Health Serv Res. 2010;10:330.

93

34. Kousgaard MB, Joensen AS, Thorsen T. The challenges of boundary spanners in supporting inter-organizational collaboration in primary care - a qualitative study of general practitioners in a new role. BMC Fam Pract. 2015;16:17. 35. Carroll C, Patterson M, Wood S, Booth A, Rick J, Balain S. A conceptual framework for implementation fidelity. IS. 2007;2. 36. Engels Y, van den Hombergh P, Mokkink H, van den Hoogen H, van den Bosch W, Grol R. The effects of a team-based continuous quality improvement intervention on the management of primary care: a randomised controlled trial. Br J Gen Pract. 2006;56(531):781-7. 37. Huston P, Hogg W, Martin C, Soto E, Newbury A. A process evaluation of an intervention to improve respiratory infection control practices in family physician offices. Can J Public Health. 2006;97(6):475-9. 38. Shaw E, Looney A, Chase S, Navalekar R, Stello B, Lontok O, et al. 'In the Moment': An Analysis of Facilitator Impact During a Quality Improvement Process. Group Facil. 2010;10:4-16.

94

95

Table 2. Facilitator characteristics and data Facilitator Facilitator gender Facilitator age Individual interview Focus group Observed

A Female 45-49 + + + B Female 55-59 + + + C Male 40-44 + + + D Male 60-64 + + + E Female 60-64 - + + F Female 40-44 + - + G Female 45-49 + + + H Female 45-49 + + + I Female 45-49 - + + J Female 55-59 - + + K Male 60-64 - + - L Male 50-55 - + - M Female 50-55 - + - N Male 60-64 - - -

+ interviewed, in focus group or observed; - not interviewed, not in focus group or not observed

Table 3. Interview guides Individual interviews About the observed practice The preparation of the facilitator and the practice Their descriptions and experience of the meetings (structure, dialogue, what worked well and what did not) Their own role at the visits and in the change process The practice impact from the visits What they perceived as successful/less successful in the specific practice How visits differed from other practices and what is similar between practices About their general perception and practice Their understanding of the facilitator intervention Their contribution as facilitators How they were prepared for the role as facilitator What had influenced their understanding and enactment facilitation Variations between facilitators and practices The implication of being GPs Thoughts about project design (number of visits, meetings in the practice etc.) Focus groups Focus group 1 Their understanding and description of the facilitator intervention Their preparation The visit content The good visit Their tools Their competences Facilitator variations and unity The impact for the practices Focus group 2 Their understanding of facilitation and the roles of the facilitator If and why some roles were more often applied What influenced their roles Whether they had collaborated on a common understanding Improvements of their performance over time Their competences The significance of them being GPs Project initiators and project managers Background for the intervention The intervention design Their understandings of facilitation The education of the facilitators Expected changes Intervention flexibility

96

Table 4. Illustrations of enacted roles Role

Excerpts from observation notes

Teacher In practice 3, the facilitator gave a very structured PowerPoint presentation of the DCM. Before beginning, the facilitator said: ‘just interrupt, if anything is unclear’. He then described the system, how to sign-up, how to record and access the patient data, and how to use the system for quality improvement. The facilitator did most of the talking, sometimes answering questions from the practice. The presentation lasted for about one hour with the facilitator loosely skipping over some slides or just reading them aloud.

Super user In practice 10 with only one GP present, the facilitator emphasised that the GP should be the one sitting in front of the computer. The facilitator sat next to him, guiding him. The GP had installed the DCM some time ago but had not used it. They looked at his ICPC-diagnosis coding percentage and the facilitator showed how him to use the DCM. The facilitator found that the system set-up was not correct and that the GP was not typing all values in the right boxes. The facilitator suggested that the GP contacted his system provider […]. At the next visit, the facilitator asked the GP if he had increased his coding percentage and once again found problems in the system set-up. The facilitator contacted the IT-system provider who explained how to set up the system and the GP learned this as well.

Peer In practice 7 the facilitator explained that as an inspiration she would now describe how she had organized the COPD treatment in her own practice. She did so in detail using a PowerPoint presentation. There were a few comments along the way, but mostly the facilitator talked, while the practice was listening. The facilitator underlined that this was her way of organizing the clinic, and that the practice should find out how they wanted to do it. Prior to the visit, practice 3 had chosen to focus on the DCM. Before giving a detailed introduction to the DCM the facilitator stated ‘there are three main gains from using the DCM and I am not saying it as a representative of the Region, but because I am working with it myself in my practice’. During the visit several references to the facilitator’s own practice were made, both on the initiative of the facilitator and of the practice.

Process manager

Practice 7 and 12 had chosen to make new procedures for their COPD care. At the end of the first visit the facilitator ensured that 2-3 tasks were specified and that people in charge of each were chosen. At the second visit, the facilitator began going through the list of tasks asking about the status. In both practices, the appointed people answered that the procedures had been formulated. In practice 7 they were already using the new procedures, and the facilitator asked if they were functioning well, and they agreed that they were. In practice 12 one team member had made a draft and an internal meeting had been scheduled. In neither of the two practices was the content of the procedures discussed.

Coach As mentioned above Practice 1 was the only observed practice where the facilitator attempted to engage in a more coaching based approach, although this was not fully enacted. The facilitator tried to get the participants to reflect on their own practice through an exercise where each participant wrote down the things that worked well in their diabetes care as well as ideas for improvements and potential barriers. The facilitator then asked each participant about their thoughts. Several issues were brought up during the exercise, but not as a dialogue between the practice members. Rather they stated if they agreed or disagreed with each other’s statements addressing their comments to the facilitator. Also, they did not discuss how to proceed and instead the facilitator suggested that (before the next visit) the practice should arrange an internal meeting to discuss two patient cases and their ideas about how to improve the structure of diabetes care.

97

Appendix 4: Paper 2

Influences of peer facilitation in general practice – a qualitative study

Tina Drud Due, Marius Brostrøm Kousgaard, Frans Boch Waldorff & Thorkil Thorsen

98

Influences of peer facilitation in general practice – a qualitative study

Tina Drud Due1, Marius Brostrøm Kousgaard1, Frans Boch Waldorff1, 2 & Thorkil Thorsen1

Affiliation: 1The Research Unit for General Practice and Section of General Practice, Department of Public Health, University of Copenhagen, Denmark. 2Research Unit for General Practice, Department of Public Health, University of Southern Denmark, Denmark.

Corresponding author:

Tina Drud Due

E-mail: tina.due@sund.ku.dk

Co-authors:

Marius Brostrøm Kousgaard: marbro@sund.ku.dk Frans Boch Waldorff: fransw@sund.ku.dk Thorkil Thorsen: tthorsen@sund.ku.dk

99

Abstract Background Practice facilitation is increasingly used to support guideline implementation and practice development in primary care and there is a need to explore how this implementation approach works in real-life settings. We focus on a facilitation intervention from the perspective of the visited practices to gain a more detailed understanding of how peer facilitation influenced practices and how they valued the facilitation. Methods The facilitation intervention was conducted in general practice in the Capital Region of Denmark with the purpose of supporting the implementation of chronic disease management programmes. We observed 30 facilitation visits in 13 practice settings and interviewed the visited practices after their first and last visits. We then conducted a thematic analysis. Results Most of the respondents reported that facilitation visits had increased their knowledge and skills in relevant areas as well as their motivation and confidence to change. These positive influences were ascribed to a) the facilitation approach b) the credibility and know-how associated with the facilitators’ being peers c) the recurring visits providing protected time and invoking a sense of commitment. Despite these positive influences both the facilitation and the change process were also impeded by several challenges. Conclusion Practice facilitation is a multifaceted, interactive approach that may affect participants in several ways and it is important to attune the expectations of all the involved actors through elaborate discussions of needs, capabilities, wishes, and approaches. Future research should explore the complex links between the cognitive influences of practice facilitation and sustainable behavioural change.

[Keywords: facilitation, facilitators, outreach visits, primary care, qualitative study, general practice]

100

Background Various strategies are used to support guideline implementation and practice development in primary care e.g. regulations, financial incentives, and information dissemination. A more active and increasingly widespread strategy is practice facilitation [1-8]. This is a multifaceted intervention, where an external person (most often a health care professional) visits the practice and supports a process of change [1, 7]. A recent systematic review and meta-analysis concluded that practice facilitation has “a moderately robust effect on evidence-based guideline adoption within primary care” [1]. However, there is considerable heterogeneity between the included studies, and generally there is no clear and consistent operational definition of facilitation. Hence, specific facilitation interventions vary considerably in their form and content. In the literature, facilitators are portrayed as having multiple roles and performing multiple activities [3, 6, 8, 9]. Among these are audit and feedback, consensus building, plan-do-study-act circles, provision of advice and education, cross-pollination of good ideas and support of internal discussions, and critical reflection. Recent contributions have emphasised the importance of tailoring facilitation to the specific needs and circumstances of the targeted practices [1, 10, 11].

Given the increasing popularity of facilitation, the flexibility of the concept, and the heterogeneity among interventions labelled as facilitation, there is a need to explore how facilitation is actually performed in real-life settings, how it affects practices, and how it is experienced by participants. From January 2011-December 2012, the Capital Region of Denmark carried out a facilitation intervention to support the implementation of chronic disease management programmes for type-2-diabetes and chronic obstructive pulmonary disease (COPD) in general practice. The intervention relied on general practitioners (GPs) as facilitators. In two previous studies, we explored how facilitation was enacted in this intervention and the effectiveness of the intervention [12, 13]. First, based on observations and interviews with facilitators we found that facilitation was enacted through four major roles: the teacher (knowledge dissemination), the super user (hands-on knowledge dissemination on the practice’s computer system), the peer (facilitators conveying their experiences and information about their own practice organisation), and the process manager (selection of topics, tasks and status reporting at subsequent visits). We also found that the facilitators rarely enacted a more coaching based approach ensuring internal reflection and discussion during the visits [12]. Second, our randomised controlled trial on the intervention’s effectiveness showed mixed results with no difference between the allocation groups for the primary outcome, but some differences on secondary outcomes [13]. With the present study we supplement our previous results by focusing on facilitation from the recipients’ (i.e. general practice) perspective to gain a more detailed understanding of how peer facilitation influenced practices and how they valued the facilitation. We also identify several factors which inhibited the facilitation process.

Material and methods Setting and intervention The Danish health care system is primarily tax financed and offers free-of-charge access to general practice and public hospital services. The GP serves as the primary care provider and gatekeeper for patients’ referral to specialists and hospitals. They are private entrepreneurs, but mainly financed through the tax financed health care reimbursement scheme. The service provision of general practice is regulated via the collective agreement between the Danish Regions and the Organisation of General Practitioners [14, 15].

101

Chronic disease management programmes based on the Chronic Care Model [16, 17] have been developed in all five regions of Denmark [18]. The programmes outline evidence based treatment and a systematic approach to chronic care with division of tasks between GPs, hospitals and municipalities. They describe the GP’s role as coordinator of care and outline a systematic proactive approach with population based patient registration, annual chronic disease check-ups, and stratification of patients into three levels by risk of complications and complexity and state of the disease [19, 20]. Diverse initiatives have been initiated to support the implementation of the chronic disease management programmes and to improve chronic care management. The facilitation intervention in this study was one of these initiatives and it was developed and implemented by the Capital Region of Denmark. The overall aim of the intervention was to support the implementation of chronic disease management programmes for type-2-diabetes and COPD in general practice. Fourteen GPs were hired as facilitators. They went through an educational programme focused on an update of the content of the disease management programmes and related tools, and on how to be a facilitator. All practices in the region were offered up to three visits of one hour each. Visits were free of charge and the practices were compensated for lost income. The central principle of the intervention was that the change process should be driven by the practices’ own interests and choice of topics and that the facilitators should therefore tailor their activities to address the particular situation and needs of each practice. Thus, the intervention relied on the idea of a continuum of facilitator roles. The information material sent to the practices suggested relevant themes for the visits such as workflow procedures and division of tasks for chronic disease management, leadership and organisation, collaboration with municipalities and hospitals, the role of the GP as coordinator of care, and IT solutions for improved overview and systematisation, primarily the Data Capture Module (DCM). The DCM is a software program which automatically collects patient data from the GPs’ electronic health record system and provides individual and population based patient overview and data for quality improvement [21]. Shortly after the initiation of the facilitation intervention, sign-up to the DCM became mandatory and all practices were required to sign up no later than the 1st of April 2013. The intervention has been described in more detail elsewhere [12]. As researchers, our role was to evaluate the intervention and we were not involved in either the design or the implementation of the intervention. Methods We chose an explorative approach for both data collection and analysis. Practices were strategically sampled [22], to ensure variation in geography, size, current level of development in areas relevant to the disease management programmes (assessed by initial questionnaires), and the associated facilitator. We observed 30 facilitation visits in 13 practice settings. Extensive notes were written and the visits were audio recorded. Further, the first author conducted group interviews in the 13 practice settings after their first and their last visit (4 of the 13 facilitator visits were joint visits where collaborating practices where present; hence, a total of 18 practices were represented). The group interviews lasted approximately one hour, and we strived to include all GPs and staff who had been present at the facilitation visits. Table 1 presents an overview of the data material. As shown the data collection was not complete in all practice settings. We audio recorded the interviews, transcribed them verbatim, and analysed them using thematic analysis [23]. We used the software program NVivo in the coding and theme constructing process for the interviews. We grouped codes in themes and sub-themes and then related the themes to each other and to

102

the entire data material, thus refining and connecting them. The observations were primarily used to qualify the interview guides and the interviews but from the observation notes and audio recordings we also obtained information about task completion, the process between the visits, and potential challenges. According to Danish law a qualitative study like this one does not require ethical approval by the research ethics committee system.

Findings Prior to the visits most practices only had a vague notion of what to expect from facilitation, and their understanding of the intervention was generally limited. Also the practices did not appear to experience a strong need for change. The dominant reasons for participating in the intervention was to get help with the DCM (because it became mandatory), or because the visits were seen as an occasion to get started with developing more systematic procedures for chronic care check-ups. A few practices had merely signed-up because a colleague had mentioned the intervention. Most of the observed practices chose the DCM as their main topic while two practices focused mainly on developing new chronic care procedures for diabetes and COPD. The topics of the visits are described in Table 1.

At the first visits in the observed practices the practice decided on the topics of the visits. However, there was no introductory dialogue about the practices’ expectations or preferred facilitation approach and a limited clarification of their existing level of knowledge within the chosen topic. During the visits, the facilitators mainly engaged in various forms of knowledge dissemination, practical support and process management. Although the intervention design also comprised a more coaching based approach to support internal discussions and reflections (e.g. about existing and future procedures) this approach was not enacted during the observed visits. Still, the majority of the respondents were pleased with the visits and did not wish for this sort of facilitation approach. Several respondents appreciated the knowledge and inspiration offered by the facilitators, and some did not envisage that there was sufficient time at the visits for more elaborate discussions about their practice organisation. Nevertheless, two practices were quite dissatisfied with the visits because they had mainly expected the facilitators to engage the participants in an inspirational discussion about what changes were needed and how to implement them. Instead they experienced the facilitators taking an educative stance which did not involve asking the participants reflective questions and which lacked a focus on implementation:

it is not what a facilitator is supposed to do. When [the facilitator] is sitting on the side-line if you [the practice] are sitting and talking in the group, it is primarily making sure you do not lose focus, but also providing ideas in the process, saying… So that was what I had expected more of, more on the side-line, and then that we as a practice had tried to talk about how we would organise this. (GP, Practice 12)

One of these practices described that they rarely set time aside for discussions about practice development. Therefore, they had hoped that the visits would have focused more on supporting their internal discussions and development processes but they described that if a temporary doctor in training had not single-handedly taken upon her the task of making new procedures, they would not have accomplished much. In the other practice, the GPs were so disappointed with the facilitator that they declined more visits.

103

Across the observed practices generally, profound changes in direct patient care were not initialised after the facilitation visits, but there were several examples of practices having initiated components of change. Several practices increased their use of diagnosis coding and some installed and signed-up for the DCM, corrected the system set-up, and improved their data registration. However, none came as far as using the DCM data for quality improvement. Two practices wrote new chronic care procedures, and the practice interviewed after the last visit had begun to implement it. Additionally, a few practices expressed increased attention towards some of the addressed issues e.g. annual chronic disease check-ups and the webpage for municipal chronic care activities. However, some practices did not express any tangible changes and some reported limited or no impact from the visits.

Knowledge and skills The facilitators provided factual knowledge about International Classification of Primary Care (ICPC) diagnosis coding of individual consultations in the electronic patient records, the content of chronic disease check-ups, the DCM, and websites on professional guidelines and municipal chronic care services to which GPs can refer patients. This was either done by presentations, by showing demo versions of the DCM, by demonstrating relevant websites, or by hands-on guidance in the practices’ electronic patient record systems [9]. Prior to the visits, most practices had not used the DCM. Some had not yet installed it and some had not managed to set up the programme so as to generate accurate data. Further, they rarely diagnosis-coded individual consultations and they had little knowledge (and made little use of) the various websites introduced by the facilitators. On this background practices experienced that the facilitation visits increased their knowledge and awareness both of new tools and how to use them, and of errors in the set-up of the DCM. Some respondents stated that the knowledge provided by the facilitators ensured a faster implementation process due to knowledge being more easily accessible, and others perceived the knowledge, especially about the correct set-up of the DCM, as being essential for progress, because they would not have figured it out themselves:

We found out that we did not do it, that the computer was not set up properly... it turned out that the nurses’ computer was not set up to register the diagnosis-coding, which we had done through half a year. (Nurse, Practice 7)

Respondents generally described the content of the visits as relevant, because they had chosen the topics themselves, and because these topics were closely related to their daily practice and specific challenges (experienced prior to and in-between visits). The respondents also found that conducting the facilitation meetings in the practice constituted a beneficial frame for knowledge provision. Contrary to lectures in larger settings, the facilitation visits were focused on them, there were no disturbing questions from other practices, and they felt safe asking questions and revealing their weak points. Likewise, some appreciated that joint meetings in the practice increased the likelihood of the knowledge being applied, and relieved the GPs from spending time conveying it to the staff. However, other GPs preferred meetings without the staff so that the meetings were focused on the needs of the GPs.

Regarding patient related data for quality improvement, the practices generally did not review their own data prior to or in between the visits. However, the practices which looked at such data during the visits valued this experience. For them, the facilitation visits improved their appreciation of the relevance of patient data, helped them to identify problems, gave them an opportunity to consider data (which they

104

could not usually find time for), and reinforced them to improve the registrations even more. A few practices also improved their skills in using their information systems due to the hands-on approach. The practices were generally satisfied with the technical knowledge of the facilitator. Nevertheless, there were examples of facilitators lacking knowledge about the specific IT systems used by the practice (there are 11 IT systems in Danish general practice) and several times they asked practices to contact their IT-providers with questions and problems they could not handle themselves. Some practices would have preferred a facilitator that had experience with their specific IT system, while others did not perceive this as a barrier. Several practices experienced IT challenges and issues with IT providers between the visits. This seemed to slow down the implementation process as some practices did not complete tasks or did so at a slower pace. At the first visit, the facilitators did not clarify exactly what the visited practices wanted to focus on within a given topic or the level of their existing knowledge. Thus, although most practices reported that they obtained new knowledge from the facilitation visits, some of the knowledge provided was not new to everyone in the practices. While the GPs had generally gained little new knowledge from the presentations on medical and organisational aspects of chronic care, the practice staff often found this knowledge more relevant; not because it directly affected their own work, but because it gave them an improved understanding of the GPs’ work. There were also several examples of participants forgetting the knowledge provided during the visits and several participants still had questions about the correct use of the DCM after the visits. Some felt that too little time had been spent on some of the topics, and would have preferred the visits to be more structured and to have received written material on both the DCM and the facilitators’ organisation:

One might have been given a sort of a template. Because the problem is that you forget it a bit afterwards…what is it you need to remember to implement it… perhaps one might have needed that. So a small action card. How to do it… because we cannot remember it now right. (GP, Practice 13)

Motivation and confidence to change According to most respondents, their motivation and confidence to change increased as a result of the facilitation visits. They experienced the process of change as demystified and more manageable because the facilitators showed that the DCM was easier to use than they had assumed, and the facilitators’ descriptions of their own chronic care procedures gave them something to build upon:

It might seem a bit less unmanageable and hopefully a little less time consuming than I feared it would be. (GP, Practice 5)

And

It was really good to get it [description of facilitators’ chronic care procedures], so you did not have to reinvent the wheel. I think it was, was really good. (Nurse, Practice 7)

Further, the facilitators’ descriptions of the benefits they had gained from making the changes in their own practices as well as the content of their chronic care procedures inspired the visited practices by increasing their sense of usefulness in daily practice. Most GPs found that the facilitators being peers, with personal experience and knowledge of daily life in general practice, added to their credibility. This meant that they

105

generally perceived the facilitators’ statements as relevant, trustworthy, and transferable to their own practice:

I think it is true that a general practitioner will reach us more easily. We listen because there is a professional respect ... We listen more sharply and take it more seriously … than if it was a nurse… she would initially have to struggle against whether we would could use it for anything. (GP, Practice 2)

The GPs did not perceive the descriptions of the facilitators’ own practice organisation as something to be directly copied, but as a credible source of inspiration. The practices generally did not experience disadvantages from the facilitators being peers. Some could not see how the facilitators could have other professional backgrounds, while a few did not regard the peer component as crucial for the process. However, one of the previously mentioned dissatisfied practices felt provoked when the facilitator presented them with factual and experience-based knowledge because they did not perceive the facilitator as an expert or someone with an outstanding practice but just as a random GP. Also, while most GPs became more motivated by the visits some still expressed a feeling of obligation toward the DCM and doubted whether they would use the system beyond the required registrations:

Well it is the obligation that does it, because it is something that we have to do. If we had not had to, the question is whether we would have done it. That I don’t know. (GP, practice 3)

Additionally, the technical problems experienced in the process triggered increased frustration with the DCM:

Well it is just difficult to mobilise any energy among the doctors, who are to sit and code, if the shit does not work, excuse my directness. Then I bloody do not want to, and again I swear. Then I do not want to sit there and spend my time on something like that. Then it must be left to its own device until it is working. (GP, Practice 1)

Internal conditions for change Three aspects of the intervention, which did not relate to the specific content of the visits nor to the specific skills and actions of the facilitator, influenced the change process and how the practices assessed the intervention.

First, the visits offered an occasion to focus on and initiate changes and provided protected time which was much valued by the respondents, who reported on busy workdays where time was usually not set aside for practice development meetings with both GPs and staff attending. Thus the visits were described as a timeout for development which accentuated the focus on the chosen topics:

It also just helps quite a lot by creating a focus, because we devote an hour to it and sit here all of us together. Instead of in our busy workdays, where we just quickly went in and looked, and had set aside half an hour and then were fifteen minutes late and just got to look at something. Then this gives it much focus. (GP in training, Practice 2)

However, sometimes the observed visits got delayed and sometimes people were absent or left during the meeting. Thus, while most respondents appreciated the visits being held in the clinic for practical reasons, some mentioned that this also increased the risk of interruptions and delays since patients were waiting before, during, or after the visits.

106

Second, practices reported that the visits supported task definition and delegation and increased the sense of obligation, agreement, and mutual responsibility because the whole practice was gathered for the visits. However, from the observations it was clear that the clarity and systematisation of task definition and delegation varied and occasionally clear tasks were not explicitly defined.

Third, several practices described how the return of the facilitator at subsequent visits came to function as a reminder and deadline during the process. According to some respondents this speeded up the change process and ensured the completion of initiated projects that otherwise would probably not have been prioritised in a busy working day:

So you knew, that you had a meeting at this and that date and suddenly, you were a bit more motivated to go in and code and do things…. So the meetings have another function than just being a meeting, they also have the function of keeping you up to scratch. (GP, practice 3)

Thus several practices managed to perform their delegated tasks and/or to set a deadline for their implementation before the next visit. Still, most practices did not have a profound focus on the change process in-between visits. For instance they rarely discussed the tasks or changes in formal meetings, and some mentioned tasks only being done shortly before the visit. This limited attention to the change process was ascribed to the daily time pressure in general practice.

Discussion Most of the respondents from general practice reported that facilitation visits had increased their knowledge, awareness, and skills in relevant areas as well as their motivation and confidence to change. These positive influences were ascribed to the facilitation approach (visits focused on their experienced challenges and a hands-on approach to knowledge dissemination), the facilitators’ background (endowing them with credibility and know-how), and the formal frames established by the intervention (protected time, recurring visits) with the latter sometimes appearing to be more influential for generating engagement and commitment than the specific content of the facilitators’ presentations. The respondents were generally quite satisfied with these aspects and influences of the intervention, and previous studies have found that GPs appreciate facilitation visits for some of the same reasons [6, 24-27]. However, despite the positive influences of facilitation in this study there were also some challenges which impeded the change process: a) some of the knowledge provided was redundant, inadequate, or forgotten; b) the visits were sometimes delayed or interrupted; c) not all practices prioritised change efforts between visits due to busy work schedules, and d) technical challenges with the DCM decreased motivation and wasted precious time. Further, the practices had diverse understandings and expectations in relation to the facilitation visits and in some cases these expectations were not fulfilled. This suggests that a more explicit initial dialogue about current knowledge and preferred facilitation approach should have been initiated early in the process. A study by Watkins et al likewise discussed the need for an introductory talk about objectives and rules of engagement [28].

Facilitators in general practice are most often nurses or practice assistants rather than GPs [1]. However, in this study most of the GPs found it to be important that the facilitators were peers, because this helped to establish the credibility of the facilitators and to increase the GPs’ perceptions of manageability and usefulness. This resembles the value that has often been ascribed to opinion leaders as change agents [29].

107

In both cases much of the influence of the change agent is linked to the legitimacy and credibility gained by having worked under similar conditions. Another facilitation study also indicated the importance of the facilitator being a fellow GP in order to establish trust and acceptance [25], but studies with non-peer facilitators have also reported that practices were satisfied with the visits [27, 30, 31]. Since these studies only had one type of facilitator it is difficult to draw solid conclusions about the advantages and disadvantages of using different professions as facilitators. We have only identified two studies that compared facilitation visits by a peer GP with visits by a non-peer and these indicated that peer facilitation is more effective [32, 33]. However, in both of these studies the groups lacked comparability due to differences in the facilitators’ education prior to conducting the visits and their prior facilitation experience.

Strengths and limitations Using interviews with practices as well as observations of facilitation visits is a strength of this study. Although rarely used in facilitation studies, observations provide a more nuanced picture of the facilitation process when combined with the practices’ reported experiences. Thus the observations made it possible to explore potentially less idealised versions of the facilitation process and to pose more nuanced and critical questions to the practices. It is also a strength that data was collected prospectively while the intervention was carried out since this reduced recollection bias among the participants and made it possible to explore the entire process. Potential limitations are that not all practices were interviewed twice and that the group interviews (where GPs and staff were interviewed together) might have inclined staff not to state conflicting opinions and made GPs more careful about criticising the peer facilitator. However, we deemed it important to give room for dialogue about a common experience between the various participants.

Regarding the transferability of the findings, complex interventions always take place in a specific context which put certain limits on transferability. However, some of the external and internal contextual conditions influencing the facilitation intervention described here have also been found to play a role in other studies, e.g. competing priorities and heavy clinical workload, areas with special political attention (in this case the DCM), IT problems, and limited traditions for formal quality improvement meetings [6, 25, 34, 35].

Conclusion In this study of practice facilitation in a real-life setting most of the participants from general practice experienced that facilitation had increased their knowledge in some areas of chronic care and changed their perceptions of the relevance and manageability of making changes in these areas. The participants pointed to several elements of the intervention which influenced the process positively such as the flexibility of the intervention (allowing participants to choose among several different topics), the provision of protected meeting time, the legitimacy and know-how of the peer facilitators, the focus on defining and delegating tasks, and the commitment associated with the deadlines set by recurrent visits. Despite the overall positive assessments of the participants, a number of internal and external factors were found to impede the facilitation process. Some of these challenges may be alleviated by thorough discussion of the needs, capabilities, and wishes of the involved practices from the outset, and by employing facilitators with diverse skills so that the different needs and starting points may be optimally matched and addressed. Furthermore, since knowledge, attitudes, and confidence are important but not sufficient factors for

108

generating behavioural change [36] future research in practice facilitation should explore the complex links between such cognitive influences of facilitation and sustainable changes in the practice organisations.

Abbreviations COPD: Chronic obstructive pulmonary disease GP: General practitioner DCM: Data Capture Module ICPC: International classification of primary care Competing interests The authors declare that they have no competing interests. Authors’ contributions TDD designed the study, collected the data, analysed and interpreted the data, and drafted the manuscript. FBW Participated in the interpretation of the data and the drafting of the manuscript. MBK participated in the interpretation of the data and the drafting of the manuscript. TT supervised the study, and participated in the design of the study, the interpretation of the data, and the drafting of the manuscript. All authors read and approved the final manuscript. Acknowledgments We thank the Danish Research Foundation for General Practice, The Health Foundation, and the Research Foundation for Primary Care in the Capital Region of Denmark for funding the study. Our thanks also go to the practices for participating in the study. References 1. Baskerville NB, Liddy C, Hogg W. Systematic review and meta-analysis of practice facilitation within primary care settings. Ann Fam Med. 2012;10(1):63-74. 2. Bidassie B, Williams LS, Woodward-Hagg H, Matthias MS, Damush TM. Key components of external facilitation in an acute stroke quality improvement collaborative in the Veterans Health Administration. Implement Sci. 2015;10:69. 3. Dogherty EJ, Harrison MB, Graham ID. Facilitation as a role and process in achieving evidence-based practice in nursing: a focused review of concept and meaning. Worldviews Evid Based Nurs. 2010;7(2):76-89. 4. Harvey G KA. Implementing evidence-based practice in healthcare: a facilitation guide. 1 ed. London: Routledge; 2015. 5. Kauth MR, Sullivan G, Blevins D, Cully JA, Landes RD, Said Q, et al. Employing external facilitation to implement cognitive behavioral therapy in VA clinics: a pilot study. Implement Sci. 2010;5:75. 6. Kotecha J, Han H, Green M, Russell G, Martin MI, Birtwhistle R. The role of the practice facilitators in Ontario primary healthcare quality improvement. BMC Fam Pract. 2015;16:93. 7. Liddy C, Laferriere D, Baskerville B, Dahrouge S, Knox L, Hogg W. An overview of practice facilitation programs in Canada: current perspectives and future directions. Healthc Policy. 2013;8(3):58-67. 8. Nagykaldi Z, Mold JW, Aspy CB. Practice facilitators: a review of the literature. Fam Med. 2005;37(8):581-8. 9. Harvey G, Loftus-Hills A, Rycroft-Malone J, Titchen A, Kitson A, McCormack B, et al. Getting evidence into practice: the role and function of facilitation. J Adv Nurs. 2002;37(6):577-88. 10. Liddy CE, Blazhko V, Dingwall M, Singh J, Hogg WE. Primary care quality improvement from a practice facilitator's perspective. BMC Fam Pract. 2014;15:23.

109

11. Stetler CB, Legro MW, Rycroft-Malone J, Bowman C, Curran G, Guihan M, et al. Role of "external facilitation" in implementation of research findings: a qualitative evaluation of facilitation experiences in the Veterans Health Administration. Implement Sci. 2006;1:23. 12. Due TD, Thorsen T, Waldorff FB, Kousgaard MB. Role enactment of facilitation in primary care– a qualitative study (unpublised). 13. Due TD, Thorsen T, Kousgaard MB, Siersma VD, Waldorff FB. The effectiveness of a semi-tailored facilitator-based intervention to optimise chronic care management in general practice: a stepped-wedge randomised controlled trial. BMC Fam Pract. 2014;15:65. 14. Pedersen KM, Andersen JS, Sondergaard J. General practice and primary health care in Denmark. J Am Board Fam Med. 2012;25 Suppl 1:S34-8. 15. Christiansen T. Organization and financing of the Danish health care system. Health Policy. 2002;59(2):107-18. 16. Bodenheimer T, Wagner EH, Grumbach K. Improving primary care for patients with chronic illness. JAMA. 2002;288(14):1775-9. 17. Bodenheimer T, Wagner EH, Grumbach K. Improving primary care for patients with chronic illness: the chronic care model, Part 2. JAMA. 2002;288(15):1909-14. 18. Forløbsprogrammer for kronisk sygdom - Generisk model og forløbsprogram for diabetes [Disease management programmes - generic model and Disease management programmes for Type 2 Diabetes]. Copenhagen: The Danish Health and Medicines Authority, 2008. 19. Forløbsprogram for Type 2 Diabetes - Hospitaler, almen praksis og kommunerne i Region Hovedstaden [Disease management programme For Type 2 Diabetes - Hospitals, general practice and municipalities in the Capital Region of Denmark]. Hilleroed: 2009. 20. Forløbsprogram for KOL - Hospitaler, almen praksis og kommunerne i Region Hovedstaden [Disease management programme for COPD - Hospitals, general practice and municipalities in the Capital Region of Denmark]. Hilleroed: 2009. 21. Schroll H, Christensen RD, Thomsen JL, Andersen M, Friborg S, Sondergaard J. The danish model for improvement of diabetes care in general practice: impact of automated collection and feedback of patient data. Int J Family Med. 2012;2012:208123. 22. Patton M. Qualitative Research & Evaluation Methods. 3 ed. United States of America: Sage publication; 2002. 23. Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3(2):77-101. 24. Bunniss S, Gray F, Kelly D. Collective learning, change and improvement in health care: trialling a facilitated learning initiative with general practice teams. J Eval Clin Pract. 2012;18(3):630-6. 25. Buch MS, Edwards A, Eriksson T. Participants' evaluation of a group-based organisational assessment tool in Danish general practice: the Maturity Matrix. Qual Prim Care. 2009;17(5):311-22. 26. Dalsgaard T, Kallerup H, Rosendal M. Outreach visits to improve dementia care in general practice: a qualitative study. Int J Qual Health Care. 2007;19(5):267-73. 27. Kousgaard MB, Thorsen T. Positive experiences with a specialist as facilitator in general practice. Dan Med J. 2012;59(6):A4443. 28. Watkins C, Timm A, Gooberman-Hill R, Harvey I, Haines A, Donovan J. Factors affecting feasibility and acceptability of a practice-based educational intervention to support evidence-based prescribing: a qualitative study. Fam Pract. 2004;21(6):661-9. 29. Locock L, Dopson S, Chambers D, Gabbay J. Understanding the role of opinion leaders in improving clinical effectiveness. Soc Sci Med. 2001;53(6):745-57. 30. Baskerville NB, Hogg W, Lemelin J. Process evaluation of a tailored multifaceted approach to changing family physician practice patterns improving preventive care. J Fam Pract. 2001;50(3):W242-9. 31. Huston P, Hogg W, Martin C, Soto E, Newbury A. A process evaluation of an intervention to improve respiratory infection control practices in family physician offices. Can J Public Health. 2006;97(6):475-9.

110

32. Petrova M, Dale J, Munday D, Koistinen J, Agarwal S, Lall R. The role and impact of facilitators in primary care: findings from the implementation of the Gold Standards Framework for palliative care. Fam Pract. 2010;27(1):38-47. 33. van den Hombergh P, Grol R, van den Hoogen HJ, van den Bosch WJ. Practice visits as a tool in quality improvement: mutual visits and feedback by peers compared with visits and feedback by non-physician observers. Qual Health Care. 1999;8(3):161-6. 34. Hogg W, Baskerville N, Nykiforuk C, Mallen D. Improved preventive care in family practices with outreach facilitation: understanding success and failure. J Health Serv Res Policy. 2002;7(4):195-201. 35. Tierney S, Kislov R, Deaton C. A qualitative study of a primary-care based intervention to improve the management of patients with heart failure: the dynamic relationship between facilitation and context. BMC Fam Pract. 2014;15:153. 36. Dixon A. Motivation and Confidence: what does it take to change behaviour? . London: The King’s Fund; 2008.

111

112

113

Appendix 5: Paper 3

The effectiveness of a semi-tailored facilitator-based intervention to optimise chronic care management in general practice: a stepped-wedge randomised controlled trial

Tina Drud Due, Thorkil Thorsen, Marius Brostrøm Kousgaard, Volkert Dirk Siersma & Frans Boch Waldorff

114

115

116

117

118

119

120

121

122

123

124

125