Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Primary care, public health and the intersectoral management of health determinants : a realist inquiry Clair, Veronic 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2014_november_clair_veronic.pdf [ 5.44MB ]
Metadata
JSON: 24-1.0166068.json
JSON-LD: 24-1.0166068-ld.json
RDF/XML (Pretty): 24-1.0166068-rdf.xml
RDF/JSON: 24-1.0166068-rdf.json
Turtle: 24-1.0166068-turtle.txt
N-Triples: 24-1.0166068-rdf-ntriples.txt
Original Record: 24-1.0166068-source.json
Full Text
24-1.0166068-fulltext.txt
Citation
24-1.0166068.ris

Full Text

   PRIMARY CARE, PUBLIC HEALTH AND THE INTERSECTORAL MANAGEMENT OF HEALTH DETERMINANTS:  A REALIST INQUIRY  by  Veronic Clair   M.Sc., Montréal University, 1999  M.D., Montréal University, 2001 Certificant, Family Medicine, University of British Columbia, 2004 Fellow, Royal College of Physician and Surgeons of Canada, Public Health and Preventive Medicine, University of British Columbia, 2006  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Population and Public Health)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  October 2014   © Veronic Clair, 2014  ii   Abstract Problem: International bodies such as the World Health Organization call on nation-states to more adequately address health determinants (HD) and strengthen primary healthcare through intersectoral action for health (IAH). Despite many leaders expressing the desire to strengthen IAH, in many countries, practitioners and policy-makers struggle to succeed. This study aims to deepen the understanding of how to maximize the impact on health and HD through collaborations between primary care providers (PCP), public health professionals (PHP), and representatives of sectors other than the health sector (ROS).    Method: This research includes a supplementary analysis of a mixed-methods case study on Cuban IAH to clarify and compare the roles of the various actors of interest (PCP, PHP, ROS) in managing HD, including the contexts, mechanisms, and outcomes (CMO) in which IAH occurs in Cuba when involving those particular actors. This is followed by a knowledge synthesis of IAH interventions internationally, providing a variety of different contexts, which enables systematic comparison of the various CMO configurations extracted from those interventions, following a method inspired by realist synthesis. Data for the final CMO analysis and demi-regularity are drawn from both the Cuban case study and the other examples of IAH from around the world that involved PCP, PHP, and ROS.  Conclusions: IAH involving PHP, PCP, and ROS can lead to significant, positive health outcomes through the management of HD. A key context in which significant improvement in HD and health outcome occurs is when the IAH are carefully planned based on prior evidence and best practices related to partnership building and public health. Key mechanisms of those interventions include: 1) systematic attention to infrastructures, and activities that successfully increase social capital; 2) which in turn supports the negotiation of complementary and synergistic roles between PCP, PHP and ROS, and 3) using cycles of adjustment based on best practices of quality improvement which enable cumulative and reinforcing synergies over time (years and decades), as projects unfold in complex changing policy and practice environments, and as the multiple actors increase their social capital and experience in dealing with health determinants.    iii  Preface  This study could not have taken the shape it has without building extensively on the work carried out by the University of British Columbia (UBC) - and Cuban-based research team of which I was a part. Along with colleagues and other researchers from UBC and Cuba, including Jerry Spiegel, Milagros Alegret, Nino Pagliccia, Barbara Martinez, Mariano Bonet, and Annalee Yassi among others, I participated in an in-depth case study of Cuban management of health determinants at the municipal level. From the development of the initial concept to the publication of our results, I played various roles and therefore performed a variety of tasks (primarily during 2004-2006), including collecting relevant literature and contributing to the development of the Spanish-language survey and focus group guide for that case study. In close collaboration with the other team members, I contributed significantly to the qualitative analysis of the main case study’s focus groups to ascertain a correct interpretation of the transcripts. I participated in two field trips associated with the study, as well as in several workshops to collect the case examples and reach a final interpretation of the case study findings, under the leadership of Drs. Yassi and Spiegel, who are two members of my thesis committee. It should be noted that our team’s approach was inspired, in part, by a similar study that explored how Canadian public health institutions address health determinants (Frankish et al., 2007).  The findings of the case study, of which I am a co-author, are presented in “Intersectoral action for health at a municipal level in Cuba,” published in the International Journal of Public Health (Spiegel et al., 2012). I contributed to approximately 15% of the design, analysis, and writing of that study, under the leadership of Drs. Spiegel and Yassi. This article is quoted extensively in several places in the current document (with appropriate permissions obtained) as it provides useful information on the broad contexts, mechanisms, and outcomes of intersectoral collaboration to manage health determinants in Cuba.   The differing focus of my PhD thesis (as opposed to the full case study) was established prior to the obtainment of ethics approval, and I am acknowledged as a PhD student and a co-investigator on both the pilot study and the full case study. Both received ethics approval from the UBC Behavioural Research Ethics Board, respectively as (H04-80624) B04-0624 - Developing Methodology to Understanding the Cuban Paradox: A Pilot Study on the Determinants of Health in iv  Cuba and (H06-80184) B06-0184 - Inter-Sectoral Collaboration in Securing Excellent Health Outcome: A Multi-Methods Study.  Although my contribution to the case study forms an integral part of the work carried out in this thesis, it must be understood that in my supplementary analysis of the Cuban data, the frame of reference was shifted from a focus on the overall analysis of IAH (with some contrast of how it is practiced in rural communities and in urban communities), to a focus on the nature of the interaction between the practitioners, in an urban setting. I owe a great debt to the other team members, and am appreciative for their openness to using part of our collective findings in this thesis, which is appropriately cited, mainly in Chapter 3, section 3.1. Subsequent analysis of the data, mostly presented in section 3.2 to 3.4, were carried out by myself with much-appreciated advice from my thesis advisory committee members, Dr. Jerry Spiegel, Dr. Annalee Yassi, Dr. Robert Woollard, and Dr. Marjorie MacDonald.   The design of the supplementary analysis was developed to assess the similarities and differences between the roles of PCP, PHP, and ROS in managing HD through IAH in Cuba, as well as to determine in what context those collaborations occur, through what mechanisms, and with what outcomes. The supplementary analysis was mostly designed and performed by myself.  The statistical plan and analysis presented in the article mentioned in the previous paragraph were designed and performed by Nino Pagliccia, a PhD statistician who worked with the Global Health Research Program until shortly before I started my own analysis plan for my PhD thesis (Spiegel et al., 2012). That analysis inspired part of my own statistical analysis plan. I also consulted with Jonathan Berkowitz, a PhD statistician affiliated with UBC’s Department of Family Practice. He provided advice on the analysis strategy based on my draft analysis plan and research questions, helped to clean and reformat the data, and wrote some of the commands in the SPSS syntax files. Some of the statistical analyses were inspired by the analyses that were done as part of the full case study published in the article mentioned previously. However, due to a mistake in coding some of the categories and decisions to group some previously separated categories, I ended up re-running all the analyses. For the qualitative analysis, I consulted Dr. Guenther Krueger, a qualitative design and NVivo consultant recommended to me by my advisory committee. The analytic design for the supplementary analysis of the Cuban case study was discussed during thesis committee meetings, as well as during Canada-Cuba research team meetings, which did not result in any major changes.  v  Table of contents Abstract ............................................................................................................................................... ii Preface ................................................................................................................................................ iii Table of contents ................................................................................................................................ v List of tables...................................................................................................................................... vii List of figures ................................................................................................................................... viii List of symbols and abbreviations ................................................................................................... ix Glossary ............................................................................................................................................. xi Acknowledgements ......................................................................................................................... xiv Chapter 1: Introduction .................................................................................................................... 1 1.1 Health determinants ...........................................................................................................................1 1.2 Primary care .......................................................................................................................................9 1.3 Intersectoral collaboration ...............................................................................................................16 1.4 Why Cuba’s achievements matter ...................................................................................................19 1.5 Research question and purposes ......................................................................................................23 Chapter 2: Methodology.................................................................................................................. 25 2.1 Evolution of the research process and researcher reflexivity ..........................................................25 2.1.1 Secondary use of qualitative data and qualitative methods .........................................................27 2.1.2 Addition of a realist inquiry lens .................................................................................................29 2.2 Supplementary analysis of the Cuban case study ............................................................................30 2.2.1 Focus group and survey participants ...........................................................................................33 2.2.2 Focus group methods ..................................................................................................................37 2.2.3 Survey analysis ............................................................................................................................42 2.2.4 Integration of the supplementary analysis of the Cuban case study ............................................45 2.3 Realist synthesis ..............................................................................................................................46 2.3.1 Search strategy ............................................................................................................................56 2.3.2 Inclusion and exclusion criteria ...................................................................................................58 2.3.3 Extraction tool .............................................................................................................................62 Chapter 3: Cuba’s management of health determinants ............................................................. 63 3.1 Relevant information from the case study primary analysis including document review and field visits 63 3.2 Results of the focus groups ..............................................................................................................76 3.2.1 Conceptualization of health determinants ...................................................................................78 3.2.2 Priority in managing health determinants ...................................................................................85 3.2.3 Management of health determinants ...........................................................................................87 3.2.3.1 Strategies to manage health determinants from the various professional groups ...............87 3.2.3.2 Internal actions and intersectoral collaboration to manage health determinants ................90 3.2.3.3 Decision-making, evaluation, successes, and challenges ...................................................96 3.3 Quantitative findings and the integration of survey and focus group results ................................107 3.3.1 Prioritization of health determinants .........................................................................................108 3.3.2 Intensity of participation in internal actions to manage health determinants ............................112 3.3.3 Actions within participants’ own organization mentioned most frequently as important in addressing each health determinants ......................................................................................................116 3.3.4 Intersectoral collaboration .........................................................................................................122 3.3.4.1 Intersectoral collaboration by health determinants and professional groups ....................122 3.3.4.2 Intersectoral collaboration by health determinants, sectors, and professional groups ......123 vi  3.3.4.3 Intensity of collaboration ..................................................................................................128 3.3.4.4 Frequency of intersectoral collaboration ..........................................................................130 3.4 Building a framework to describe the system of intersectoral collaboration involving primary care providers and public health personnel with other sectors to address health determinants in Cuba ............132 3.4.1 Context ......................................................................................................................................132 3.4.2 General mechanisms .................................................................................................................134 3.4.3 Mechanism of contributions of various groups: primary care, public health, municipal, and local leaders. 137 3.4.4 Outcomes of the health creation system ....................................................................................144 Chapter 4: Realist review results.................................................................................................. 147 4.1 Articles included in the realist review and the health determinants they address .........................147 4.2 CMO of larger initiatives initiated by organizations with expertise in IAH and public health .....156 4.2.1 Context ......................................................................................................................................156 4.2.2 Mechanisms ...............................................................................................................................159 4.2.3 Outcomes...................................................................................................................................169 4.3 CMO of smaller projects ...............................................................................................................177 Chapter 5: Conclusion ................................................................................................................... 179 5.1 Discussion of key findings from the re-analysis of the Cuban case study and the realist review .179 5.1.1 Common and divergent contexts ...............................................................................................180 5.1.2 Common mechanisms ...............................................................................................................186 5.1.3 Common outcomes ....................................................................................................................201 5.2 How rigour was ensured in the research process ...........................................................................205 5.3 Limitations .....................................................................................................................................207 5.3.1 Limitations of the Cuban case study re-analysis .......................................................................207 5.3.2 Limitations of the realist review ................................................................................................210 5.4 Implications for future research and practice ................................................................................221 Bibliography ................................................................................................................................... 226 Appendix A Focus group and survey instruments...................................................................... 250 A.1 Focus group guide .....................................................................................................................250 A.2 Final survey questionnaire ........................................................................................................253 A.3 Definition of terms provided to participants in the focus groups ..............................................257 Appendix B Details of the search strategies ................................................................................ 259 Appendix C Extra data tables and analysis from the survey priority ranking question ........ 262 Appendix D Detailed statistics on internal action to manage health determinants ................. 273 Appendix E Detailed statistics on intersectoral collaboration ................................................... 285 Appendix F Details of the studies included in the realist review ............................................... 293  vii  List of tables Table 1 Public health and primary care collaboration barriers and facilitators adapted from Valaitis & al..................................................................................................................................................... 14 Table 2 Specific databases searched .................................................................................................. 58 Table 3 Inclusion and exclusion criteria for the realist synthesis ...................................................... 61 Table 4 Vertical and horizontal integration of Cuban government and health institutions ............... 71 Table 5 Organizations represented in the Santa-Clara focus group ................................................... 77 Table 6 Health determinants priorities by focus groups .................................................................... 86 Table 7 Strategies used to manage health determinants .................................................................... 89 Table 8 Actions listed in the local leaders focus groups .................................................................... 93 Table 9 Actions listed in the municipal leaders focus groups ........................................................... 94 Table 10 Actions listed in the primary care and public health focus groups ..................................... 96 Table 11 Intersectoral actions most important to further improve the health of the population ..... 103 Table 12 Mean ranking of health determinants by professional groups .......................................... 111 Table 13 Integration of priority health determinants across focus groups and survey .................... 111 Table 14 Mean level of internal action on each health determinant by professional groups........... 115 Table 15 Actions most important in managing health determinants internally ............................... 119 Table 16 Intersectoral collaboration listed ....................................................................................... 123 Table 17 Sectors involved in addressing each health determinant by professional group .............. 126 Table 18 Intensity of collaboration by professional groups ............................................................. 129 Table 19 Frequency of collaboration by professional groups .......................................................... 131 Table 20 Summary of social mechanisms facilitating Intersectoral Action for Health Determinants, involving PC, PH and OS, in the Cuban model. .............................................................................. 143 Table 21 Projects and references included in the realist review ...................................................... 152 Table 22 REACH outcomes............................................................................................................. 172 Table 23 Mechanisms in intersectoral action on health determinants (ISA HD) using complexity concepts............................................................................................................................................ 200 Table 24 Summary of research questions associated with gaps in knowledge, methods, and findings.......................................................................................................................................................... 204 Table 25 Normality tests, all participants ........................................................................................ 262 Table 26 Normality tests by professional groups ............................................................................ 263 Table 27 Details statistics for priority ranking of health determinants by professional groups ...... 264 Table 28 Tests of normality, all participants ................................................................................... 266 Table 29 Tests of normality for each professional group ................................................................ 267 Table 30 Kruskal-Wallis analysis of health determinants prioritization ......................................... 268 Table 31 Descriptive statistics of the level of internal action on health determinants by professional groups ............................................................................................................................................... 270 Table 32 Test of normality of intensity of internal action for each health determinant .................. 272 Table 33 Kruskal-Wallis analysis of internal action level ............................................................... 273 Table 34 Intensity of internal action by health determinant and professional group ....................... 283 Table 35 Organizations participating in intersectoral collaboration, by sector, for each health determinant ...................................................................................................................................... 285 Table 36 Full and frequent collaboration by sector, health determinant, and group ....................... 287 Table 37 Interventions retained in the realist review and the health determinants they addressed . 293 Table 38 Selected articles’ contexts, mechanisms, and outcomes for large initiatives with organizations experienced in public health and/or intersectoral partnership ................................... 299 Table 39 Projects part of the small initiative CMO configurations ................................................. 310  viii  List of figures  Figure 1 CSDH conceptual framework................................................................................................ 2 Figure 2 GDP per capita and life expectancy at birth .......................................................................... 4 Figure 3 Trends in GDP per capita and life expectancy at birth .......................................................... 4 Figure 4 Female under 5 years old mortality rate by GDP for countries in the Americas ................ 19 Figure 5 SciELO database search results combining public health and primary care keywords ...... 32 Figure 6 Examples of intersectoral actions flows for various health issues in Cuba ......................... 69 Figure 7 Mean ranking of health determinants by professional groups ........................................... 110 Figure 8 Level of internal action for each health determinants by professional group ................... 114 Figure 9 Proportion of full collaboration, when collaboration is identified .................................... 128 Figure 10 Cuban health creation model to ensure its citizens’ right to health ................................. 133 Figure 11 Realist review literature selection process ...................................................................... 149 Figure 12 Demi-regularity of projects started as part of larger initiatives ....................................... 176  ix  List of symbols and abbreviations  ALBD:  Active Living by Design bf: Breastfeeding CDC:  Centre for Disease Control and Prevention (US) CIDA:  Canadian International Development Agency CIHR:  Canadian Institute for Health Research CMO: Context mechanism and outcome CNP:  Community Network Program CSDH:  WHO Commission on Social Determinants of Health EPODE:  Ensemble Prévenons l’Obésité des Enfants [Together, let’s prevent childhood obesity] HIC : High Income Countries IAH: Intersectoral Action for Health ISA HD:  Intersectoral Action for Health Determinants INHEM:  Instituto Nacional de Higiene, Epidemiologia and Microbiología [National Institute of Hygiene, Epidemiology and Microbiology] IOM: Institute of Medicine ISA HD: Intersectoral action to manage health determinants KP: Kaiser Permanente LMIC: Low and Middle Income Countries MICY: Maternal, Infant, Child and Youth NBCCEDP: National Breast and Cervical Cancer Early Detection Program NC: North Carolina NCPP: North Carolina Prevention Partners NIH: National Institute of Health NYACTY: New York Asset Coming Together for Youth OS:  Other sectors PAHO:  Pan American Health Organization PBRN:  Practice-based Research Network PC: Primary care PH: Public health REACH:  Racial and Ethnic Approaches to Community Health Rx for Health: Prescription for Health x  RWJF:  Robert Wood Johnson Foundation SC: South Carolina SCMJ: South Carolina Medical Journal SCCDCN: South Carolina Cancer Disparities Community Network SES: Socio-economic status WHO:  World Health Organization  xi  Glossary The glossary below is an attempt to promote a common understanding of how certain terms are conceptualized in this research.    Health Equity: “Equity means fairness. Equity in health means that people’s needs guide the distribution of opportunities for well-being…This implies that all people have an equal opportunity to develop and maintain their health, through fair and just access to resources for health.” (Nutbeam, 1998) p. 355  Health Inequality: “[Health] Inequality refers to health differences that may be possible to reduce but not eliminate, such as those related to genetics or aging; inequity refers to differences that are unfair and preventable. Governments cannot necessarily fix all inequalities, but they can take action to reduce inequities.”  (Health Council of Canada, 2010)  Health Promotion and the Ottawa Charter: The first International Conference on Health Promotion led to the publication of the landmark Ottawa Charter for Health Promotion. It defines health promotion as: “the process of enabling people to increase control over, and to improve, their health. To reach a state of complete physical, mental and social wellbeing, an individual or group must be able to identify and to realize aspirations, to satisfy needs, and to change or cope with the environment. Health is therefore seen as a resource for everyday life, not the objective of living. Health is a positive concept emphasizing social and personal resources, as well as physical capacities. Therefore, health promotion is not just the responsibility of the health sector, but goes beyond healthy life-styles to wellbeing.” (WHO, 1986) p.1 The Charter proposed the following set of strategies to improve population health:  Build healthy public policy  Create supportive environments  Strengthen community actions  Develop personal skills  Reorient health services  Moving into the Future  xii  Intersectoral Action for Health: “a recognized relationship between part or parts of the health sector with part or parts of another sector which has been formed to take action on an issue to achieve health outcomes, (or intermediate health outcomes) in a way that is more effective, efficient or sustainable than could be achieved by the health sector acting alone.” (WHO, 1997) p.3  Neoliberalism: “the economic and political model known as “neoliberalism” (for its emphasis on “liberalizing” or freeing markets) or the “Washington consensus” (since its main proponents – the US government, the World Bank and the International Monetary Fund – are based in Washington, DC)… The core of the neoliberal vision was (and is) the conviction that markets freed from government interference “are the best and most efficient allocators of resources in production and distribution” and thus the most effective mechanisms for promoting the common good, including health. Government involvement in the economy and in social processes should be minimized, since state-led processes are inherently wasteful, cumbersome and averse to innovation.” (WHO, 2010a) p. 14.  Public Health: A recent article aiming to define global health provides us with a good review of the definitions of public health over time: “Farr, Chadwick, Virchow, Koch, Pasteur, and Shattuck helped to establish the discipline on the basis of four factors: (1) decision making based on data and evidence (vital statistics, surveillance and outbreak investigations, laboratory science); (2) a focus on populations rather than individuals; (3) a goal of social justice and equity; and (4) an emphasis on prevention rather than curative care… The definition of public health that has perhaps best stood the test of time is that suggested by Winslow almost 90 years ago: “Public health is the science and art of preventing disease, prolonging life and promoting physical health and efficacy through organized community efforts for the sanitation of the environment, the control of communicable infections, the education of the individual in personal hygiene, the organization of medical and nursing services for the early diagnosis and preventive treatment of disease, and the development of social machinery which will ensure every individual in the community a standard of living adequate for the maintenance of health; so organizing these benefits in such a fashion as to enable every citizen to realize his birthright and longevity.” xiii  The US Institute of Medicine (IOM) … described public health in terms of its mission, substance, and organisational framework, which, in turn, address prevention, a community approach, health as a public good, and the contributions of various partners. The IOM report defined the mission of public health as “fulfilling society's interest in assuring conditions in which people can be healthy”. In the Dictionary of epidemiology (2001), Last defined public health as “one of the efforts to protect, promote and restore the people's health. It is the combination of sciences, skills and beliefs that is directed to the maintenance and improvement of the health of all the people through collective or social actions.”” (Koplan et al., 2009) p.1993.  More recently, as part of its 1998 glossary on health promotion, WHO defined public health as: “The science and art of promoting health, preventing disease, and prolonging life through the organized efforts of society.” (Nutbeam, 1998) p. 352.  Public Health Professionals: Considering the definitions above, as well as the Bangkok Charter of Health Promotion, which states that health promotion is a core function of public health, the rest of this document will use the terms ‘public health’ and ‘public health professionals’ to encompass the field and the professionals with knowledge and expertise in improving the health of populations through a variety of methods, including health promotion (Bangkok Charter, 2006). Public health practice can be seen as the efforts to improve population health through strategies that act on health determinants at various levels. Such strategies include, but are not limited to, health promotion, health protection, prevention, and the provision of healthcare services. Public health professionals also strongly rely on evidence as part of their paradigm, including the evidence generated from epidemiology.  xiv  Acknowledgements  The Canadian Institute of Health Research originally funded the Cuban case study. The secondary analysis and knowledge synthesis were self-funded.  I would like to thank all the Cubans and Canadians who donated their time to the original Cuban case study, as well as for agreeing to and facilitating the secondary analysis as part of this thesis.  A special thank you to my spouse, whose support, friendship, and companionship has made the completion of this thesis possible in the middle of starting a family.   I am also very thankful of my thesis supervisor’s and committee members’ guidance and commitment to see this thesis through to completion, despite all the obstacles encountered in the process.   This thesis would not have been written in proper English without the copy-editing work and suggestions of Michelle Everton and Rick Marcuse. A big thank you to both for the detailed work they have done.  1 Chapter 1: Introduction In a significant publication that commemorated the 30th anniversary of the Alma-Ata International Conference on Primary Health Care, the World Health Organization 2008 World Health Report proposed to maximize the impact of investment in health and health systems by strengthening intersectoral action on health determinants (WHO, 2008). The recommendations of that report (titled Primary Health Care – Now More Than Ever) call for reducing global inequities in health outcomes by improving access to high-quality, comprehensive primary care services acting in collaboration with other sectors to address health determinants.   Traditionally, public health professionals have advocated for and engaged in intersectoral collaboration. However, little is known about primary care practitioners’ contributions to intersectoral collaboration that addresses health determinants. Recognizing the need for research in this area, this study aims to add to the understanding of the nature of collaboration acting on health determinants when the collaboration involves primary care practitioners, public health professionals, and representatives of sectors other than the health sector.    This introductory chapter provides context for the research on intersectoral collaborations and health determinants. It lays out: a summary of the key literature on health determinants, primary care, public health, and intersectoral collaboration; a rationale for the selection of Cuba as central to this study; and the specific research questions and goals of the thesis.   1.1 Health determinants Factors that have a strong influence on health, such as lifestyle, environment, human biology, and healthcare services, are known as health determinants (Evans & Stoddart, 1994). Over time, cultural, social, and economic factors have been added to the list of health determinants. This has led to a useful, expanded definition of health determinants as “The range of personal, social, economic, and environmental factors that influence the health status of individuals or populations.” (Nutbeam, 1998) p. 354.  Healthcare systems can influence both more immediate (or proximal) health determinants (e.g., personal position and socio-economic status, social cohesion, psychosocial factors, behaviours and biological factors) and more distant (or distal) health determinants (e.g., governance;  2 macroeconomic, social, and health policies; and cultural norms and values), which are elements in broad socioeconomic and political settings (CSDH, 2008a). The framework developed by the WHO’s Commission on Social Determinants of Health, (CSDH, 2008a) p. 43, is reproduced here, as Figure 1.  Figure 1 CSDH conceptual framework   It cannot be surprising that endless studies and reports suggest that maximizing improvement in the health of individuals – and entire population groups – is a worthwhile goal. More telling, however, is that many studies propose that the goal is achievable. Arguments presented include the following :   Health is valued by populations worldwide (Commission on Macroeconomics and Health, 2001; CSDH, 2008a)  Health has been recognized as a human right (United Nation, 1948)1                                                  1Article 25 states: “(1) Everyone has the right to a standard of living adequate for the health and well-being of himself and of his family, including food, clothing, housing and medical care and necessary social services, and the right to security in the event of unemployment, sickness, disability, widowhood, old age or other lack of livelihood in circumstances beyond his control.  (2) Motherhood and childhood are entitled to special care and assistance. All children, whether born in or out of wedlock, shall enjoy the same social protection.” (United Nations General Assembly, 1948).   3  Improvements in population health are achievable even for low-income countries and even in conditions of zero economic growth (Commission on Macroeconomics and Health, 2001; WHO, 2008)  Several countries with a low gross domestic product (GDP) have achieved better health outcomes than countries with higher GDP (Commission on Macroeconomics and Health, 2001; WHO, 2008) (CSDH, 2008a).   As might be inferred from the previous point, investment in health systems, including some healthcare provision, can act as a powerful economic growth factor (Commission on Macroeconomics and Health, 2001; Nazeem ud et al., 2014; Sachs, 2006; Sachs & Malaney, 2002). In situations, however, where the focus is on economic growth only, there is not necessarily an improvement in health outcomes. This is illustrated in the figures that follow. The first figure shows that for a given GDP of about $1000, countries have a life expectancy at birth anywhere from less than 40 years old to more than 70 years old.  The figure also shows that several countries with a much higher GDP (such as Namibia, South Africa, Botswana and Swaziland) have a lower life expectancy than would be expected, given the trend on average (redline).    Figure 3 confirms that the relationship between GDP and health is not straightforward. Different groups of countries have different starting point of life expectancies for a given GDP, and the evolution over time or as ain function of GDP changes can follow many different paths. For example, the Russian Federation group of countries (line g, in purple) saw an increase in GDP and a decrease in life expectancy, following the break-up of the Soviet Union, while most developing countries outside of Africa (line d, in turquoise) have seen increases in life expectancy with increased GDP. Those low-income countries even reached a higher life expectancy at a lower average GDP (end of line d, in turquoise) than middle-income countries had reached at that same level of GDP (beginning of line b, light green). However, those low-income countries remained at a lower life expectancy than China had several years ago with a lower GDP (turquoise arrow of line d, in turquoise, versus line h of China in red). Figures 2 and 3 were both reproduced with the permission of WHO from (WHO, 2008) p. 4 and 5 respectively.        4 Figure 2 GDP per capita and life expectancy at birth  Figure 3 Trends in GDP per capita and life expectancy at birth     5 When attempts are made to measure the relative contribution of different types of health determinants, the data seem to indicate that health determinants not related to the provision of healthcare services are key drivers of population health.  That is, what are often referred to as social determinants of health, (e.g., ethnicity2, occupation, education) are more important in determining health status than the healthcare system (McKeown, 1979). This theory has been debated, but a comprehensive review of the literature on estimates of the contribution of various health determinants shows that clinical care provision is responsible for less than 50% of the determination of mortality and morbidity as indicator of health status. Most methodologies used to estimate the contribution to population health find that between 10 to 27%  of population health can be attributed to the healthcare system (Booske, Athens, Kindig, Park, & Remington, 2010).   Another area of study, attempting to assess the link between increase in healthcare budget around the world and improvement in health indicators, shows that, in fact, there are few such increases in healthcare budgets that have resulted in significant improvements in the health of the population (WHO, 2008). That World Health Report of 2008 on primary healthcare also shows that the increased investments in healthcare are rarely directed toward the implementation of cost-effective interventions or policies that would maximize the health return of those investments, and are most often directed toward specialized type of health services with the least evidence of impact on population health (WHO, 2008).  Along these lines, advocates of a social determinants approach to improving population health have sometimes criticized investment in healthcare and medical professionals as being excessively costly (e.g., (Muntaner, Sridharan, Solar, & Benach, 2009)). Against a backdrop of conflicting claims and competing political perspectives, it must be remembered that the ability of health systems to respond to the healthcare needs of clients is shaped by the same broad health determinants that affect the distribution of health results. Furthermore, these broader determining factors significantly affect the distribution of health outcomes among various groups, often significantly reducing health equity (CSDH, 2008a).                                                   2 Ethnicity is sometime viewed as a non-modifiable health determinant (therefore not part of social determinants), as one ethnicity is determined from birth. Others suggest that ethnicity is a social determinant of health as it relates to how various ethnic groups are treated in a given society. This thesis considers ethnicity to be a social determinant of health, as is the case in studies of health determinants in the United States, which form a large part of the studies included in the systematic review.  6  On the other hand, I have observed3 that healthcare practitioners, often armed with supportive research on effective health services, routinely ask for more investment in clinical personnel and interventions, in response to the clinical needs of individual patients who present to their doors. Other commentaries seem to relate similar observations (Robinson, 2004; Syme, 2004). The practitioners’ request for investment in clinical interventions is understandable, as explained by Black in his review of levels of evidence and levels of healthcare decision-making (Black, 2001). However, one needs to be careful to use clinical evidence to make population health and health systems investment decisions. It is essential to understand that many factors other than clinical evidence, population health evidence, or any other forms of evidence, legitimately influence decision-making (Black, 2001). One such source of influence is the advocacy by professional groups, which is legitimate, but often creates an environment in which evidence of population health impact is rarely considered or rarely a determining decision factor (Black, 2001).   A more unified perspective has been advocated by grassroots movements for rights to both health and healthcare (such as the People’s Health Movement), in line with the vision for primary healthcare described in the Declaration of Alma-Ata (Narayan & Schuftan, 2008). The People’s Health Movement recognized the imperative both to treat those who are sick, and to improve health through action on other health determinants (People's Health, 2002; People's Health Movement). This view was supported by the report of the WHO Commission on Social Determinants of Health (CSDH, 2008a), which underscored that the provision of healthcare has global implications and, further, that it has been recognized as a human right – and not simply as a right to health, but a right to healthcare. Other authors recognize that health and healthcare are drivers of economic and human development (Audibert, Mathonnat, & De Roodenbeke, 2003; Commission on Macroeconomics and Health, 2001; Das & Samarasekera, 2011).  Despite countering claims and philosophical disputes, there is mounting evidence that health systems based in primary care result in better health outcomes at lower costs, across income levels, when compared to other types of healthcare systems (Macinko, Starfield, & Erinosho, 2009).  This conclusion is also supported by the Wanless review, a comprehensive review of healthcare spending in relation to long-term health outcomes conducted for the UK government (Wanless, 2002), and in                                                  3 This is from my experience in various administrative positions in the healthcare system in Canada and participation in various international conferences.  7 a follow-up report on progress with its implementation and the importance of investment in public health (Wanless, 2004).   After a long period of disinterest in the topic, there is now renewed interest in re-orienting health systems to address health inequities and population health, including through primary care approaches (Wise & Nutbeam, 2007).  According to Wise and Nutbeam, most countries have neither the infrastructure nor an optimal system to deliver the health promotion programs and services that have been proven effective (Wise & Nutbeam, 2007). Considerable evidence suggests that even within the health system’s contribution to health expectancy, a significant proportion is related to preventive services. In this regard, Wise and Nutbeam cite a Swedish review of the impact of health services which shows that health services might have contributed to a 5-year increase in the life expectancy of Swedes, but that clinical preventive services were responsible for 1.5 years of those 5 years, even with a smaller proportion of the healthcare budget (Wise & Nutbeam, 2007). Furthermore, rare is the health system that equitably reaches the entire population for promotion, prevention, or other healthcare services (Wise & Nutbeam, 2007). (Wise & Nutbeam, 2007), p. 24-25, point out that some of the effective yet frequently missing health promotion and prevention services include:   Patient-education programs on effective self-management of chronic disease   Patient engagement in care planning and delivery   Brief interventions relating to tobacco, alcohol and or physical activity   Interventions to prevent and treat depression and other mental illnesses  Maternal and child health interventions, including immunisation and screening programs.   On this subject, Wise and Nutbeam sum up matters with this conclusion: “It is hard to find a country in which there is universal access to all the health promotion programs and services that have proven to be effective; an optimal system for the delivery of quality healthcare that includes the benefits of health promotion and that reaches, equitably, the whole population; and an optimal infrastructure for public health/health promotion. Indeed there has been continuing difficulty in obtaining and/or maintaining investment in health promotion.” (Wise & Nutbeam, 2007) p. 25.    8 As the 2008 World Health Report noted, most increases in health budgets in countries around the world have failed to result in policies maximizing the health return of those investments: “Health systems are developing in directions that contribute little to equity and social justice and fail to get the best health outcomes for their money. Three particularly worrisome trends can be characterized as follows:  Health systems that focus disproportionately on a narrow offer of specialized curative care;  Health systems where a command-and-control approach to disease control, focused on short-term results, is fragmenting service delivery;  Health systems where a hands-off or laissez-faire approach to governance has allowed unregulated commercialization of health to flourish. These trends fly in the face of a comprehensive and balanced response to health needs. In a number of countries, the resulting inequitable access, impoverishing costs, and erosion of trust in healthcare constitute a threat to social stability.” (WHO, 2008)) p. xiii.  As a result of the evolving knowledge on health determinants and based on extensive review of the literature with expert and stakeholder consultations, the CSDH made three overarching recommendations to improve population health through addressing social determinants of health:  “1. Improve daily living conditions  2. Tackle the inequitable distribution of power, money and resources   3. Measure and understand the problems and assess the impact of action” (CSDH, 2008a) p.2.  In summary, changes in the social determinants of health have a major impact on the health of populations – and perhaps a larger impact than improvements resulting from direct investment in the healthcare system alone. This is especially true when focused investments in specialized healthcare take place to the detriment of investments in primary care and in other sectors, such as housing, education, employment insurance and other critical areas affecting the social determinants of health (CSDH, 2008a; WHO, 2008). Overwhelmingly, the literature supports an integrated approach to addressing health determinants – through macro- and micro-level interventions, including public policies, health promotion, clinical preventive services and access to comprehensive, high-quality primary care services. Furthermore, the literature indicates that such an approach appears to be both desirable and achievable, even for lower-income countries. In addition,  9 an integrated approach can lead to economic development and can increase productivity and wellbeing, even if the healthcare system is not the health determinant with the greatest impact on population health. Importantly, certain types of healthcare systems (particularly those based on primary care) have more impact on population health than other types of healthcare systems, and in addition, they cost less than systems based on more specialized clinical care provision.   1.2 Primary care The seminal Declaration of Alma-Ata (1978) puts primary healthcare at the centre of the healthcare system. The Declaration defined primary healthcare as essential healthcare to be made universally accessible, constituting the first level of contact with the national health systems for individuals, families and communities in a continuum of care process (Declaration of Alma-Ata, 1978). The Declaration of Alma-Ata further noted that primary healthcare:  Is reflective of a country’s socio-cultural and political context  Addresses the major health problems of the community, including access to treatment and essential drugs  Invests in education and, disease prevention and control programs   Promotes adequate food supplies and nutrition   Provides access to safe water and basic sanitation   Engages in intersectoral collaborations   Is built as a sustainable system with integrated referral mechanisms, prioritizing those most in need, with progressive improvement towards comprehensive healthcare for all   Supports and requires the participation of different individuals and community resources, building on community strengths  This historic declaration on primary healthcare clearly includes both initial care and comprehensive care. Essential characteristics linking primary care services to positive impacts on the health of the population include offering primary care services as the first level of patient contact with the healthcare system and covering comprehensive care needs (as opposed to a single disease-oriented program) (Starfield, Shi, & Macinko, 2005b). Those characteristics are clearly part of the Alma-Ata description of primary healthcare. The Alma-Ata declaration also included elements traditionally associated with public health: the understanding of the relevance of the socio-political context; the control of epidemic diseases; immunization; public sanitation and safe water and food  10 supplies; community participation; and intersectoral collaboration to address health issues (Declaration of Alma-Ata, 1978). Building on the steps laid out in the Alma-Ata Declaration, the World Health Report 2008: Primary Healthcare - Now More than Ever called for returning to a focus on primary healthcare, and proposed a series of reforms:   “Reforms that ensure that health systems contribute to health equity, social justice and the end of exclusion, primarily by moving towards universal access and social health protection - universal coverage reforms;   Reforms that reorganize health services as primary care, i.e. around people’s needs and expectations, so as to make them more socially relevant and more responsive to the changing world while producing better outcomes – service delivery reforms;  Reforms that secure healthier communities, by integrating public health actions with primary care and by pursuing healthy public policies across sectors – public policy reforms;  Reforms that replace disproportionate reliance on command and control on one hand, and laissez-faire disengagement of the state on the other, by the inclusive, participatory, negotiation-based leadership required by the complexity of contemporary health systems – leadership reforms” (WHO, 2008), p. xvi.  The WHO report on primary healthcare called for more intersectoral collaboration:  “In many regards, the responses of the health sector to the changing world have been inadequate and naïve. Inadequate, insofar as they not only fail to anticipate, but also to respond appropriately: too often with too little, too late or too much in the wrong place. Naïve insofar as a system’s failure requires a system’s solution – not a temporary remedy. Problems with human resources for public health and healthcare, finance, infrastructure or information systems invariably extend beyond the narrowly defined health sector, beyond a single level of policy purview and, increasingly, across borders: this raises the benchmark in terms of working effectively across government and stakeholders” (WHO, 2008) p. xiii.  In this report cited above, as well as in the Alma-Ata declaration, primary care was considered a subset of primary healthcare that focuses on the organization of health services at the point of entry to the healthcare system (Declaration of Alma-Ata, 1978). It is care that is continuous, comprehensive, and organized to respond to peoples’ needs and expectations based on an enduring relationship of trust between people and their healthcare providers (Declaration of Alma-Ata, 1978).  11 This view is shared by Starfield, who proposed that intersectoral collaboration is part of the responsibility of primary care (Starfield, 1998). In this thesis, I am looking at the role of primary care practitioners (rather than the role of primary healthcare practitioners, which might include public health professionals). However, the distinction is sometimes arbitrary and it is difficult to delineate where primary care becomes more akin to primary healthcare, as many authors do not distinguish between the two. Therefore, this thesis is not seeking to establish a firm distinction between the two, other than through the criteria for inclusion of primary care in the intersectoral interventions of the realist review, as defined in the methodology section.   To add further complexity to the delineation between primary care and primary health care, the WHO report further stated that primary care has a responsibility to tackle determinants of ill health (which might have been conceived more as a primary health care responsibility, since primary health care has a broader set of services and aims than primary care) (WHO, 2008). The WHO report went on describing primary care as a driver of specialized public health or social services, through referral and commissioning (WHO, 2008). However, the examples in the report were geared toward the clinical care systems responding with appropriate healthcare services to those who are rendered more vulnerable by various health determinants. The report did not contain examples of other types of interventions by public health or by other sectors, which are key in addressing health determinants.  Furthermore, those clinical care systems’ interventions were not aimed at tackling the social determinants of health as outlined in the report of the WHO Commission on Social Determinants of Health (CSDH, 2008a).   At the same time, the report cautions against certain pressures on healthcare systems that can undermine the effectiveness of comprehensive universal access to primary care. Those pressures include economic and political crises, blurring of public and private service provision, and power differentials in expression of entitlement and rights in relation to healthcare (WHO, 2008). This raised questions as to the appropriate role of primary care practitioners in relation to collaboration with public health professionals and representatives of other sectors also influencing health determinants.   Two independent research groups have recently published scoping reviews of public health and primary care collaborations. The scoping review by Levesque et al. (2013) focuses on functional roles and organizational models that bridge individual and population perspectives, while the one by  12 Martin-Misener et al. (2012) focuses on the topics of collaboration, the types of activities conducted, and reported barriers and facilitators of collaboration. The scoping review by Martin-Misener et al. (2012) covers collaborations only in Canada, the U.S., Western Europe, Australia and New Zealand (Martin-Misener & Valaitis, 2008; Martin-Misener et al., 2012; Valaitis & al., 2012; Valaitis et al., 2012), while there were no such limitations in the review by Levesque et al. (2013); however, the latter still included only articles published in English or French.   The main finding from the review by Levesque et al., in terms of the functional roles of public health and primary care, was that the roles overlap significantly. Public health contributes to more effective primary care, and primary care contributes to a population health approach in various ways (Levesque et al., 2013). Contributions of public health to primary care included:  Data analysis to understand population needs, develop interventions and assessment tools, and evaluate the impact of medical practices;   Investigation of outbreaks with the provision of clinical interventions;  Creation of partnerships to promote health and well-being, including policies and shared responsibility in relation to the prevention of diseases, injuries, and social problems;   Support for screening, immunization, and early detection; and  Acting as a moral authority to promote equity, quality, and access to primary care, and promotion of evidence based and efficient practices (Levesque et al., 2013). Contributions of primary care to public health included:  Being a source of knowledge and data related to the needs of the community;   Monitoring and treating communicable diseases;   Contributing to health promotion;   Decreasing mortality and morbidity through providing primary care; and  Engagement of local community and other primary care stakeholders in interventions and advocacy related to social issues and inequalities, as well as to patients’ interests (Levesque et al., 2013).   Levesque et al. (2013) also document several different models of public health and primary care interactions in: community health centres, community oriented primary care policies, multidisciplinary health clinics, centres for health and social services, the integration of general practitioners with special interest and care provisions, patient-centered medical homes, etc.  13 However, the review by Levesque et al. (2013) does not demonstrate the impact of those models, as it remains descriptive of the roles and organizational models found in the literature.  Martin-Misener et al.’s (2012) scoping review had findings similar to those of Levesque et al. (2013), in terms of types of collaboration. Martin-Misener et al. (2012) showed that the most frequent types of collaborations found in the articles reviewed aimed to improve the quality and cost-effectiveness of care by applying a population perspective to primary care (22% of the collaborations described), or used clinical practices to identify and address community problems (17% of the collaborations described) (Martin-Misener et al., 2012). Similarly to Levesque et al. (2013), Martin-Misener et al. (2012) mention several papers related to the integration of the various components of the health system (mostly in the UK), while collaborations to improve access to care for the uninsured were only reported from the U.S. Martin-Misener et al. mentioned relatively frequent collaboration with academic partnerships to improve students’ experience and service delivery. Academic collaboration was not discussed by Levesque et al., but was found to be relatively common by Martin- Misener et al. (2012). The scope of the activities carried out in those collaborations was reported as vast, and did include some community engagement and some multi-sectoral involvement in some cases, although there is no report of the proportion of the collaborations that included multiple sectors, nor whether these collaborations were of a different nature than the other collaborations.   The Martin-Misener et al. scoping review details are reported in the final research report to the funding agency by Valaitis et al. (2012). The final research report addresses the findings from the whole program of research on primary care and public health collaboration, and not just the scoping review by Martin-Misener et al.). Valaitis et al. (2012) collated the major facilitators of and barriers to collaborations described in the various publications they reviewed, and classified them in terms of system, organizational, and interactional levels. Those are presented in Table 1, which I adapted from the research report p.14-15.     14 Table 1 Public health and primary care collaboration barriers and facilitators adapted from Valaitis & al.4  Major Facilitators Major Barriers System Level • Government involvement, including the ‘fit’ of support for collaboration between PC and PH, and the endorsement of the value and benefits of collaboration in the community. • Relevant policy development (e.g., the reorganization of fiscal and structural resources). • Technical, informational, and financial support for the purpose of promoting integration, such as adequate funds for administrative functions and project implementation. • Sustained government funding. • Pooling and sharing of resources, volunteer and in kind contributions. • Professional education emphasizing a “system wide” approach for training and working collaboratively in public health. • Health care reform where national priorities take precedence over local priorities, and where reform causes uncertainty with how PC and PH sectors would function within newly created structures and governance processes. • A lack of stable funding, intermittent or one-off funding, for collaborative projects. • Separate, entrenched bureaucracies for PH and medical services. • Lack of an information structure, which limits the ability to adjust practice to the underlying risk of populations, and limits the ability to share or compare data. • Lack of population health needs assessments, relevant clinical data, and an evidence base for health promotion and cost-effective PH interventions, including effectiveness of collaborations. Organizational Level • Leadership development of community-based committees or boards with diverse membership to facilitate joint planning. • Involvement of multiple types of professionals. • Structures and processes that support team communication, autonomy, minimizing of competition, and opportunities for nurses and NPs to function at their full scope of practice. • Contractual agreements, parallel reporting, and common governance structures. • Use of a standardized, shared system for collecting data and disseminating information, and linked electronic records to support effective interdisciplinary care. • Shared protocols for multi-disciplinary, evidence based practice and quality assurance;  • Dissemination of information and evidence-based toolkits and decision support tools. • Lack of a common agenda or vision as well as dominating and competing agendas. • Divergent focus of sectors (e.g., individuals and short term results in PC versus populations and long term outcomes in PH) and devaluing of key PH activities. • Deterrents to buy-in from PC, including physician workload issues, lack of joint planning, and challenges with multiple stakeholder engagement. • Role confusion in PH, and overall lack of clarity and variation in PH roles between sites. • Resource limitations, including human (resources for team building and change management), time (required for community mobilization or evaluation), financial, and space resources. • Lack of knowledge and skills, including the capabilities of management to manage diverse teams, and deficiencies in expertise related to PH skills in PC. Interactional Level • Clear roles and responsibilities for all partners. • Better knowledge of each other’s roles, skills and agencies, enhancing the speed and nature of decision making among teams. • Positive relationships including trust, tolerance, and respect of partners. • Effective team communication, including regular staff meetings, involvement of the whole team, consensus building and joint planning and listening to community partners. • Stereotypical views of PC and PH roles and a lack of trust or belief in the value of PH activities. • Resistance to change and refusal to participate in planned activities. • Lack of understanding of PH roles and interdisciplinary teamwork. • Competing priorities and agendas. • Poor rapport between PC and PH and communication issues.                                                   4 The data in this table was directly extracted from lists in Valaitis et al (2012) p. 14-15.  15   Unfortunately, the Martin-Misener review did not address which facilitators and barriers were associated with various types of outcomes. Nonetheless, the reviewers attempted to extract markers of successful collaborations, which they described as sparingly found in the literature, and very broad. Markers of successful collaborations included elements such as new and sustained programs, improved access to health services, improved health knowledge, attitudes, behaviours or outcomes, team work, and increased capacity and expertise (Valaitis & al., 2012). These markers of successful collaboration were also described as outcomes of successful collaborations, creating confusion as to their nature, and confirming that this type of review is not equipped to distinguish what works, when, why, and with which outcomes (confusing markers of success - or markers of what works -  with outcomes). The authors described some negative outcomes of the collaborations, without distinguishing what might have been pre-existing challenges not addressed by the collaborations. Negative outcomes included anxiety related to skills mix, fear of being marginalized in integrated teams, few health-related gains, skills spread too thin, loss of time for patient care, and loss of critical mass of public health staff needed to have impact, by spreading them throughout primary care teams with different paradigms and responding to different pressures.  Unfortunately, that review did not extend to collaborations with sectors outside of public health and primary care nor from publications outside of a very restricted group of high income, westernized countries. It is uncertain whether the findings apply to collaborations that are broader than public health and primary care, or to collaborations in LMIC or even to other HIC outside of the high-income, Westernized countries included. Furthermore, by nature, a scoping review does not attempt to explain what works, when, and why. It describes what is found in the selected documents, but does not describe how much agreement there is between the various documents, nor why there is agreement or not.   As alluded to previously, the Martin-Misener et al. (2012) scoping review was part of a larger program of research, which included: an environmental scan of public health and primary care collaboration in three Canadian provinces; a collation of key informants’ perceptions on structures and processes influencing collaboration and the nature of existing collaboration; a synthesis of key Canadian stakeholders views regarding primary care and public health collaboration; and a series of 10 case studies of public health and primary care collaboration in two Canadian provinces. The  16 overall findings contributed to the development of a multilayered model of public health and primary care collaboration. The model included the level and factors outlined in Table 1, except with a change from an interactional level to intra- and inter-personal levels. The overall approach of Valaitis et al.’s (2012) program of research remained descriptive. It did not tease out the different outcomes arising from different types of collaboration in different contexts, or whether those outcomes are due to different mechanisms being present or absent (what works when, how, why, and with which outcomes). This is also the case of the review by Levesque et al. Both groups mentioned that most of the documents included were editorial or opinion papers, descriptive case reports, or reviews (Levesque et al., 2013; Valaitis et al., 2012). Therefore, there are still many gaps in knowledge around how intersectoral collaborations involving primary care, public health, and other sectors work and what outcomes arise from various types of collaborations. Key literature on intersectoral collaboration will therefore be reviewed in the next section.  1.3 Intersectoral collaboration Intersectoral health collaboration is “a recognized relationship between part or parts of different sectors of society, which has been formed to take action on an issue to achieve health outcomes or intermediate health outcomes in a way which is more effective, efficient or sustainable than might be achieved by the health sector acting alone.” (Nutbeam, 1986).  Several reports from governments and other institutions called for more comprehensive intersectoral action on health determinants (CSDH, 2008a; PHAC, 2007; PHAC & WHO, 2008; Romanow, 2002; Subcommittee on Population Health, 2009; WHO, 2008). However, such calls are rarely followed by concrete actions (Health Council of Canada, 2010; Raphael, 2003a, 2003b, 2008). The complexity of intersectoral action was recognized via a systematic review of 18 countries engaged in addressing health determinants (PHAC & WHO, 2008). This review emphasized that how intersectoral action for health is developed and how it is implemented matter as much as what actions are taken.  To assist governments enacting intersectoral actions, several organizations have created toolkits and supporting materials. The Health Council of Canada5 created a practical document for decision makers on the rationale and the “how to” for government promotion and implementation of intersectoral action (Health Council of Canada, 2010). The document was produced based on an extensive literature review, key informant interviews, and expert advisory panel input. The Health                                                  5 The Health Council of Canada is a non-profit organization created by federal and provincial governments’ accords, with a mission to inform and strengthen the healthcare systems.  17 Council of Canada concluded that effective intersectoral action on health determinants would necessitate (Health Council of Canada, 2010): 1) An overriding philosophy and long-term commitment backed by adequate resources that governmental initiatives and decisions should be viewed through a population health lens, under the leadership of those at the top (including the prime minister). This long-term commitment should involve capacity building of both elected representatives and civil servants of all relevant ministries on the health implications of various programs and policies of various ministries, and on effective cross-sector interventions of those in government; 2) The use of evidence (even if not yet conclusive), data, and other information in a systematic fashion to create clear, identifiable, and measurable goals, with realistic timelines, support tools, and evaluation and communication of progress towards reaching the goals; 3)  Establishing the appropriate infrastructures for society’s participation in the initiatives, with an independent authority to coordinate activities across ministries and departments. Building the appropriate infrastructure and human capacity should include cross-ministry structures and processes. Such structures and processes need to enable various external stakeholder contributions, including sharing leadership, accountability, and rewards among partners, as well as balancing central direction and local discretion.  The need for appropriate infrastructure was also discussed by O'Neill, Lemieux, Groleau, Fortin, and Lamarche (O'Neill, Lemieux, Groleau, Fortin, & Lamarche, 1997). They examined how coalition theory can inform intersectoral action for health and summarized the main literature on coalition theory that was most relevant to the context of intersectoral action for health. O’Neill et al. (1997) used Gamson’s parameters of functioning coalitions to illustrate the domains of infrastructure needed for successful collaboration. These parameters include: the effective distribution of resources; meaningful anticipated rewards as a result of being part of the coalition; support for building positive ties with the other players or sectors; and formal agreement on rules and decisions. They also noted that the power of the participating actors in the collaboration is important, and the formal and informal ties between the actors play a significant role in the sustainability of the coalitions. O’Neill et al. added another parameter relevant to sustainable collaboration, which is the ‘organizational context’, given that the organization environment plays a role in the behaviours and interactions of the members of the collaboration.    18 Furthermore, as noted by O’Neil et al. (1997) in their review of the literature on intersectoral collaboration for health, despite occasional suggestions on how to work in an intersectoral context, most authors concede that, in fact, this type of work fails more often than it succeeds.  They attribute these failures to two factors. The first factor was that health professionals operate in prestigious sectors of society and often expect other sectors to “buy-in” to health-related issues without regard for the legitimate mandate and agenda of other sectors. The other reason advanced by O’Neil et al. was that recommendations for intersectoral action on health are usually based on lessons derived from trial and error, without rigorous scientific methodology or theory guiding the knowledge acquisition. O’Neil et al. concluded that ideology needs to be transformed into actual practice (O'Neill et al., 1997).   O’Neil et al. (1997) contributed significant theoretical advancement in that regard, using case studies to further develop the coalition theories as they apply to intersectoral action for health (O'Neill et al., 1997). As mentioned previously, they added organizational context as a factor that supports intersectoral collaboration, based on their literature review. Despite advances in building a sound theoretical basis for intersectoral action for health, O’Neil et al. called for more research, as their work was still preliminary and exploratory, based on three case studies and a review of the literature outlining the lack of sound theoretical understanding of why intersectoral action sometimes works, and often fails.    Intersectoral action promotes awareness of differing perspectives of the health consequences of current practices and policies (Harris, Services, Health, & Promotion, 1995; Nutbeam & Harris, 1995). Theoretically, physicians and nurses have important insights about the impact of policies and practices on health. In practice, however, physicians, nurses, and other clinicians have only a partial understanding of various health determinants and their interactions or how to improve health by acting on health determinants (PHAC, 2007). This is frequently compounded through differing ideological perspectives, such as those described by Raphael - for example, in terms of personal responsibilities versus societal responsibilities in the adoption of healthier practices (Raphael, 2001, 2003a, 2003b, 2009, 2011, 2013a, 2013b; Raphael, Curry-Stevens, & Bryant, 2008). The common lack of awareness among clinicians regarding their partial understanding and ability to act on health determinants can impede the health sector’s participation in intersectoral collaboration. Even though intersectoral collaborations have benefits, maintaining a collaborative environment has its share of difficulties (Bauld et al., 2005; Bauld & Langley, 2010), such as: limited time for collaboration and  19  Haiti Bolivia Honduras Ecuador Cuba Peru Brazil Mexico Chile Argentina Canada USA Costa Rica Guatemala Venezuela Colombia 0 20 40 60 80 100 120 1000 10000 100000 GDP (international dollars) Mortality rate short-term projects; limited resources to invest in the collaboration itself (rather than in individual organizations’ deliverables); and a lack of intersectoral collaboration skills.   Therefore, to fully implement the vision of primary care participation to broadly address health determinants (as recommended by WHO), there is a need for more than a broad statement of the responsibilities of primary care and public health professionals to address health determinants with intersectoral action. There is a need to develop and test how those collaborations work, how they are created and sustained, and what impacts are associated with the various ways those collaborations are created, managed, and sustained.  To summarize, the literature shows the scarce examination of specific contexts, mechanisms, and outcomes that determine how in ersectoral collaboration can successfully address health determinants. This is especially true of collaborations involving primary care practitioners, public health professionals (the health sector’s stakeholders most often perceived as having a prominent role to play in managing health determinants), and representatives from sectors other than the health sector. The next section will explain why the Cuban example is relevant to learn how to make these types of collaborations work, and how to fulfill the promises of the Ottawa Charter, the WHO Commission on Health Determinants and the WHO call for primary care renewal.    1.4 Why Cuba’s achievements matter Cuba, a relatively poor country, has achieved excellent health outcomes (Pietroni, 2001; World Economic Forum, 2003).  Considerable evidence indicates that Cuba is a distinct outlier when standard population health indicators are considered in relation to per capita Gross Domestic Product (GDP).  As shown in Figure 4 (reproduced from Spiegel J.M., Yassi A. (2004) p. 87, with permissions) Cuba’s female under-5 mortality rate is comparable to what is being achieved in developed countries such as the US and Canada (Spiegel & Yassi, 2004b). This female mortality rate in Cuba is well below that of Latin American countries with similar levels of economic performance, such as Bolivia, Honduras, and Ecuador. As pointed out by Spiegel and Yassi (2004), a similar success story emerges for Cuban life expectancy and other health outcome indicators (PAHO; Spiegel & Yassi, 2004b; World Bank Group, 2003).   Figure 4 Female under 5 years old mortality rate by GDP for countries in the Americas  20           Cuba was not always an outlier (MacDonald, 1999).  Prior to 1959, Cuba’s health indices were similar to other poor countries. Then, from 1959 to the late 1980s, there was steady improvement in health indices in Cuba. For example, in 1959 the infant mortality rate was 60 per thousand live births; ten years later, this rate had dropped to 46.7, and by 1983 was 16.8 and steadily decreasing. In 1991, the rate was 10.7 per thousand live births. This dramatic improvement did not take place in other poor countries in the region (MacDonald, 1999). Cuba’s disease patterns have in fact evolved to a stage where they are today more similar to developed than to developing countries, with the proportion of deaths from chronic non-communicable diseases, such as heart disease and cancer, far outweighing mortality from infectious and parasitic disease. This prompted the observation that “[Cubans] live like the poor and die like the rich” (Macintyre & Hadad, 2002).   Cuba has succeeded in having primary care providers and public health professionals contribute to intersectoral actions in a variety of ways, some of which described in the following paragraphs (Alegret, Yassi, Spiegel, & Rodriguez, 2003; Barcelo, Guzman Pineiro, Spiegel, & Rodriguez, 2003; Bonet et al., 2007; Castell-Florit Serrate, 2007; Castell-Florit Serrate, Carlota Lausanne, Mirabel Jean-Claude, & Cabrera Gonzalez, 2007a; de la Torre, Lopez Pardo, Gutierrez Muniz, & Rojas Ochoa, 2004; Loucks et al., 2003; Munoz et al., 2002; Pagliccia et al., 2010; Sanchez et al., 2009; Saney, 2003; Spiegel et al., 2003; Spiegel, Yassi, Mas, & Tate, 2002; Spiegel et al., 2007; Spiegel et al., 2008; Spiegel & Yassi, 2004b; Whiteford & Branch, 2007; Yassi et al., 2003; Yassi et al., 1999). An example of this occurred during the “Special Period” of the late 1980s, following the collapse of its Soviet Bloc trading partners.    21 During the “Special Period”, Cuba faced a drastic decrease in trade and aid, with severe economic repercussions. In response to the potentially severe impact of the crisis on its population, Cuba sustained an intensification of comprehensive community participation strategies that had been set in motion in advance of the crisis to manage the situation (Saney, 2003). This resulted in increased investments in primary care and social services, rather than massive cuts to and/or dismantlement of the publically funded healthcare system, which has occurred in other countries facing fiscal crisis. As a consequence, rather than experiencing a decline in health status like that observed in many former members of the Soviet Union ((WHO, 2008) p. 4), Cuba suffered little impact on the health of its population and rapidly resumed its improving trend (Spiegel, Labonte, & Ostry, 2004). Overall, not only did Cuba achieve great health outcomes on average, but it achieved great equity in doing so across Cuba, in both rural and urban areas, both industrial and services-based economic areas, and both urban core and suburban areas (Alegret et al., 2003; Lopez Pardo, 2004).    Cuba’s achievements, which have been widely acknowledged (World Economic Forum, 2003), have often been attributed to political will (Spiegel et al., 2004). While of course true, this does not explain the mechanisms and processes whereby the health sector and various other sectors collaborate to address health determinants. Despite recognition of Cuba’s accomplishments, global scientific and public policy communities are scarcely learning what they could from this experience (Spiegel et al., 2004).   Cuba’s success, in terms of improved mortality and morbidity, is sometimes attributed in part to its comprehensive, universal health coverage based on primary care services as the point of entry into the health system (Spiegel & Yassi, 2004a). Spiegel and Yassi (2004) described the Cuban health care system as a publicly financed, comprehensive health care system oriented toward primary care and preventive activities. They explain that the point of entry to the health system is through the local neighbourhood clinics, which are staffed with a family physician and a nurse. They described those clinics as being responsible for about 120 families each. They also explained that the doctor and the nurse live in the community they serve (they are frequently provided a house in that community) and they are therefore geographically very accessible in case of emergency. The responsibilities of the local clinic include providing medical care, as well as creating a health culture by supporting health promotion activities frequently delivered by other community organizations.    22 Spiegel and Yassi described another important element of the system: a strong vertical integration from the local clinic to the local polyclinic (with some basic specialists trained in internal medicine, pediatrics, obstetrics and gynecology, and public health, for example), to the local hospitals, and to the secondary and tertiary levels of services (hospitals, institutes, special programs), with strong attention paid to preventing hospitalization and re-hospitalization though intense, proactive community follow-up, even if patients do not present to the clinics’ doors. However, the article by Spiegel and Yassi (2004) lacks details on how primary care became successfully integrated in multisectoral actions to address health determinants, how public health and primary care collaborate with other sectors and what outcomes, successes, and challenges they have encountered; and why.   Some might doubt the ability of countries with different political systems to learn from Cuba, because it is operating under a socialist system governed by a communist party, as explained in the Cuban Constitution (Cuba, 1976). However, the fact that several countries are expanding their primary care systems through components and approaches similar to Cuba somewhat counteracted this argument (Macinko et al., 2009). In devoting a special issue to examining what developed countries can gain by "learning from developing countries", the BMJ explicitly draws attention to the importance of conducting studies such as the one conducted for this thesis:  “The link between expenditure on health and health outcomes is not straightforward. Despite burgeoning health budgets, few countries in the developed world can claim to be delivering universally high quality, equitable healthcare. Could they have something to learn from less developed countries, whose meagre resources have long ensured that cost effectiveness is a dominant consideration?” (Richards, 2004) p. 239.   Our team (led by Dr. Spiegel and including Drs. Alegret, Paglicia, Bonet, Yassi, various others in Canada and Cuba, as well as myself) sheds some light on how health determinants are managed at the municipal level. We, showed public health and primary care involvement based on the primary analysis of the case study on intersectoral collaboration to manage health determinants in Cuba (Spiegel et al., 2012). The relevant findings from our case studies that contribute to answering the research questions of this thesis are presented in Chapter 3:, section 3.1. However, as will be explained in more detail in the methodology chapter, a supplementary analysis of the original data is necessary to ascertain how primary care practitioners, public health professionals, and local and municipal leaders representative of various sectors each conceive of, prioritize, and manage health determinants. This is because in our original case study, the analysis amalgamated the answers from  23 all the stakeholder groups, describing as a whole the municipal management of health determinants (Spiegel et al., 2012).   Coalition theory specifies that differing actors have differing yet compatible goals, contributions, and rewards, and that effective collaboration depends on an actor’s ability to be an effective part of the coalition.  To learn more fully from the Cuban example, there is a need for further research on how primary care practitioners, public health professionals, and members of other sectors are able to participate together in intersectoral collaboration. This was echoed by Frenk and Chen (2011), who called for more research on addressing the divides between clinical and public health approaches, between public and private approaches, and between knowledge gained and action implemented:   “The way forward will entail a reinvigoration of research-generated knowledge as a crucial ingredient for global cooperation and global health advances. To do this we will need to overcome daunting gaps, including the divides between domestic and global health, among the disciplines of research (biomedical, clinical, epidemiological, health systems), between clinical and public health approaches, public and private investments, and between knowledge gained and action implemented.” (Frenk & Chen, 2011) p.1.  The next section will summarize the research questions and purposes that have arisen from this review of the key literature on the main concepts and domains central to this thesis.   1.5 Research question and purposes Despite the calls to implement intersectoral action to address health determinants (Bauld et al., 2005; Health Council of Canada, 2010; WHO, 2008), barriers remain (Bauld & Langley, 2010; Raphael, 2003a, 2003b, 2008, 2009, 2011, 2013a, 2013b; Raphael et al., 2008).  There is a need to generate more practical knowledge on how to implement intersectoral action to address health determinants. Within the health sector, the role of management of health determinants seems to fall heavily on primary care and public health. Many sectors, apart from the health sector, impact health determinants, with various degrees of purposeful attention to the issue (Harris et al., 2005; Nutbeam & Harris, 1995). The literature is unclear regarding what works in terms of collaboration involving public health, primary care, and other sectors to address health determinants. Therefore, the research questions of this study are as follows. In situations where there is collaboration between primary  24 care and public health professionals within broader intersectoral collaborations to address health determinants, including in the exemplary case of Cuba: - What are the contexts in which those situations happen? - What are the mechanisms leading to actions on health determinants? - What are the outcomes of those collaborations?  This study has two main purposes. One purpose of this study is to assist decision-makers in understanding and contributing to create conditions that have led to successful collaborations in addressing health determinants, including primary care, with enough detail to adapt the findings to their particular context. The other purpose is to assist stakeholders, especially primary care practitioners, in effectively participating in those collaborations to address health determinants. The next chapter provides details of the research methodology.    25 Chapter 2: Methodology The research questions outlined in the previous chapter are complex. This is, in part, because similar interventions with either slightly different mechanisms of actions, or those in slightly different contexts, might lead to different outcomes (Pawson, Wong, & Owen, 2011). To cover the complexity of the topic adequately, this study involves two main components: (a) an inquiry into the Cuban experience through the re-analysis of an existing Cuban case study of the management of health determinants, and (b) a realist synthesis of the world literature on intersectoral collaboration to address health determinants by primary care, public health, and other sectors’ representatives, This chapter will begin with a discussion of the rationale for this research approach, followed by the details of the specific methods used in the Cuban case-study re-analysis, and the realist review.  2.1 Evolution of the research process and researcher reflexivity Researcher reflexivity is the explicit positioning of the researcher as part of the context of the study (Creswell, 2013). The researcher’s biases, personal experiences, beliefs, and opinions are explicitly stated so that the validity of the results is upheld. Researcher reflexivity is particularly important in qualitative studies because of the subjectivity in the data analysis process. By explicitly stating the impetus for this research and my level of involvement in the topic of this research, I intend to increase the validity of this study.   As a family physician who has worked extensively in multidisciplinary primary care settings in Canada and other countries, my experience is that some practitioners and leaders in such settings fail to fully grasp the expertise of the public health community or to comprehensively address health determinants. This occurs, even if they work in multidisciplinary teams, and even if they are well-intentioned. This seems to be, in part, because of a lack of understanding by primary care providers of the vast field of public health, as well as their operating under a different paradigm and constraints than public health professionals. From my discussions with primary care colleagues, I learned that many also believe that they have the knowledge and skills to manage collaborations that addresses health determinants. However, in practice, I have not seen this demonstrated by these colleagues.  As a public health and preventive medicine specialist who has practiced this specialty concurrently with a clinical practice in family medicine, I have seen the power of public health  26 action and public health policies on the health determinants affecting my patients and on the health status of communities. I have also witnessed several times that when public health resources are combined with primary care resources under primary care leadership, it can lead to the use of public health professionals and resources to provide more one-on-one care and fewer public health interventions on public policies, health promotion, or surveillance.   Similarly, I have also witnessed friction, power imbalance, limited resources, and time constraints affecting collaboration between primary care providers and community organizations or community leaders. Overall, I would echo one of the findings of the literature review in the introduction of this thesis - despite many calls to action, addressing health determinants through intersectoral collaboration is difficult to pursue and to sustain. Furthermore, it seems even more difficult to study the impact of such interventions in an appropriate fashion, considering the complexity of those intersectoral interventions.  From my perspective, primary care practitioners, public health professionals, and representatives of various other sectors are all contributing relevant lenses with which to view issues. Nevertheless, more often than not, it is unclear how those groups can work effectively in intersectoral collaborations to address health determinants. It is currently unclear how best to best help with the issues I have observed and how to assess and ensure meaningful outcomes from intersectoral collaborations to manage health determinants. To prevent my perspective from unduly influencing the research process, I did not use those issues, as I have observed them, as the “a priori” organizing principle for the data extraction. Instead, I sought other models to extract data that might provide insight on how to manage health determinants through intersectoral collaboration involving primary care, public health, and other sectors’ representatives. I also discussed each step of the research process with my thesis advisory committee to seek different perspectives and verify that the research process appears sound, and that the conclusions of this thesis are solidly anchored in the data collected. Those procedures to increase validity are based on the discussion of the different elements that strengthen validity in qualitative studies, including researcher reflexivity from Creswell and Miller (Creswell & Miller, 2000).  The idea to conduct a more specific exploration of primary care participation in collaboration with public health and other stakeholders originated when I had the opportunity to participate in a workshop in Cuba in 2004 as part of a Canadian International Development Agency (CIDA)  27 University Partnerships in Cooperation and Development Tier 2 grant held by Drs. Spiegel and Yassi. It became evident from reviewing thesis proposals of the Cuban students (with whom we worked) that primary care practitioners were intricately involved in intersectoral management of health determinants in Cuba, including community-wide surveillance, health promotion, and health status reporting, and so on. Furthermore, the actors from the various sectors involved seemed to work well together, collaborating with public health professionals and other sectors in a very harmonious and productive fashion. While I contributed to two research projects funded by the Canadian Institute of Health Research (CIHR) in relation to the intersectoral management of health determinants in Cuba (Dr. Spiegel, principal investigator), there were no specific research questions regarding the involvement of primary care and public health as separate yet complementary stakeholders within the health sector. Therefore, this aspect became the focus of my PhD.  Since my PhD research question originated from my participation in the case study, it seemed rational to seek to understand in depth how Cubans carried out this type of collaboration (involving primary care, public health, and other sectors’ representatives) was fostered in Cuba, based on the information collected as part of the original Cuban case study. This was supported by the fact that the Cuban case study collected information related to the role of various stakeholders involved, including primary care and public health representatives, with local and municipal leaders of various other sectors. The next section will discuss methods related to the secondary use of data previously collected in qualitative research.  2.1.1 Secondary use of qualitative data and qualitative methods The secondary use of qualitative data carried in this thesis would be classified as a supplementary analysis of the original case study, as it took an emerging theme (the participation of primary care providers with public health professionals in a larger, intersectoral collaborative process) and sought to understand it better, although it was not the focus of the original study (Heaton, 2004). In her book, Heaton (2004) tries to bring more rigour to the typology and application of qualitative data re-analysis. As Heaton explains, the secondary analysis of qualitative and quantitative data, including mixing them to answer a new research question, to verify a primary analysis or for other purposes is becoming a more mainstream approach (Heaton, 2004). Heaton (2004) and others advocate the advantages and appropriateness of mixing methods, choosing and adapting them in innovative and creative ways based on the availability of information, feasibility, strengths, and limits of various  28 methods and sources of information, to better answer complex research questions (Frost, 2011; Heaton, 2004; Pope, Mays, & Popay, 2007; Sparkes & Smith, 2013; Taylor & Francis, 2013; Todd & et al., 2004).  This is exactly what this thesis is doing: using various available data, analysing them through different methods, and then combining the findings and generating more insight from all the components of this research together. This thesis combines the findings of the different sources of data in a manner that is closely related to constant comparison.   Constant comparison has been described as:  “The method of comparing and contrasting is used for practically all intellectual tasks during analysis: forming categories, establishing the boundaries of the categories, assigning the segments to categories, summarizing the content of each category, finding negative evidence, etc. The goal is to discern conceptual similarities, to refine the discriminative power of categories, and to discover patterns.” (Boeije, 2002) p. 392-3.  In other words, constant comparison techniques proceed through systematic comparison and categorization of all the data elements with each other, contrasting them and finding merging patterns, as well as data elements that do not fit these patterns (Frost, 2011; Heaton, 2004; Taylor & Francis, 2013; Todd & et al., 2004).   Boeije recognizes that comparison can be done in a variety of ways, and that the researcher has an essential role to play in selecting the comparisons that contribute to meaningful interpretation through theoretical sampling. She continues by explaining that the new data collected based on the theoretical sampling process is analysed and compared with the previously collected data, which is also re-analyzed taking into consideration the new data (Boeije, 2002). This thesis follows this approach by comparing the data elements of each of the sources in relation to similarities and differences by types of health professionals or other sectors’ representatives, until no more new comparisons are possible and no new insight is gained. It then repeats the process, comparing the data from different sources of information to the original Cuban case study, and creating an extensive document describing the details of all those comparisons.   At that point, following the supplementary analysis of the Cuban data, I felt that the findings might not have been as meaningful or insightful as I wished they would be to support the implementation of intersectoral action to manage health determinants by decision makers around the world. Therefore, I decided to continue my research to find other examples of intersectoral  29 collaboration to manage health determinants that involved public health, primary care, and other sectors’ representatives. As discussed later in this chapter, the method of realist review seemed the most appropriate to fulfill the purposes of this thesis.   2.1.2 Addition of a realist inquiry lens The addition of a realist inquiry lens to the analysis of interventions (on health determinants through intersectoral collaboration involving primary care, public health, and other sectors’ representatives) led me to re-immerse myself into the Cuban data and to re-categorize the findings in terms of context, mechanisms, and outcomes. This led me to derive more coherent insight from all the data pieces and resulted in the creation of the framework explained in section 3.4. The development of this framework occurred in parallel with the analysis of the data collected through the realist review. A final process of comparison between the findings of the supplementary analysis of the Cuban case study and the most important findings of the realist review took place in an attempt to generate a meaningful conclusion from the two major parts of this thesis.   Traditionally, constant comparison methods have been used in grounded theory approaches, originally developed by Glasner and Strauss in the 1960s (Frost, 2011). This thesis did not use a grounded theory approach per se. Rather, I borrowed from and modified such approach, especially the constant comparison technique, combining it with other techniques, mostly using a realist approach, to better fulfill the purposes of this research. My aim was to generate as much insight as possible based on available data, generating some new theoretical understandings without necessarily generating a new theory.   The combination of a systematic review with the supplementary analysis of a case study is a very strong methodology. This is because of the complementarity of the two methods: the case study enables subtle, in-depth understanding in one specific context (Khairul, 2008), while systematic review provides the ability to study a wide range of interventions in a variety of contexts, accounting for the variability in context. Furthermore, it combines a classic and well defined methodology to study intersectoral collaboration to address health determinants, the case-study method (Atkinson, Cohn, Ducci, & Gideon, 2005; Borg & M.B, 2002; Jansson & Tillgren, 2010; Koller et al., 2009; PHAC, 2007; PHAC & WHO, 2008; Plochg, Delnoij, Hoogedoorn, & Klazinga, 2006; Signal & Durham, 2000; Spiegel et al.), with an innovative method (that of a realist review) to  30 further advance the methodological aspect of studying intersectoral collaboration to address health determinants. The next two subsections will describe in detail the data sources and methods of the supplementary analysis of the Cuban case study and of the realist review.  2.2 Supplementary analysis of the Cuban case study As noted previously, this part of my thesis research was conducted as a subset of a program of research on public health in Cuba, carried out through the UBC Global Health Research Program and Cuba’s National Institute of Hygiene, Epidemiology and Microbiology (INHEM). The original case study of Cuban intersectoral collaboration to address health determinants was pursued in the context of more than 15 years of collaboration between the Canadian and Cuban researchers including members of and advisors to my thesis committee. The original case study sought to improve the understanding of how Cuba manages health determinants at the municipal level, by describing and comparing the intersectoral management of health determinants of two municipalities in one Cuban province.  The original Cuban case study included: a document review, site visits, key informant interviews, focus groups, and surveys on the management of health determinants at the municipal level in Cuba (Spiegel et al., 2012). The original raw data were analyzed in this thesis with a focus on differences and similarities between the representatives of various sectors in relation to how they conceptualize, manage, and prioritize health determinants. Despite the original case study collecting information from various stakeholder groups, it did not originally seek to assess whether the various stakeholders had distinct patterns of conceptualization, prioritization, internal action, or intersectoral action in relation to health determinants.   As a member of the research team, I participated in all aspects of the research, and am a co-author in the synthesis article that is extensively quoted in the introduction and in my first results chapter of this thesis, as part of the broad context of management of health determinants in Cuba (Spiegel et al., 2012). For the purpose of this thesis, subsequent chapters will detail the understanding of the management of health determinants by various professionals, leaders, and representative of various sectors, rather than a comparison between municipalities.    31 I obtained the documents for the review from several sources between 2004 and 2010. One of those sources was a broad literature search in Medline using ‘Cuba’ and ‘health determinants’ as keywords. A second source of documents came from performing a literature search using SciELO (Scientific Electronic Library Online) a database of Spanish and Portuguese language publications from various Latin American countries. The database was searched with keywords such as ‘health determinants’, ‘primary care’, and ‘primary healthcare’. The term ‘public health’ was not used, as the goal was to find information on public health and primary care collaboration. The search using ‘primary care’ as a search term was sufficient to retrieve articles that included collaboration with public health. The search with the term ‘public health’ generated a large volume of articles that had no mention of primary care, while searches with both ‘public health’ and ‘primary care’ as keywords together systematically led the database to return no articles on the Scielo.sdl.cu (the Cuban-specific database) – see screen shot in Figure 5. When the search was attempted in the larger Latin American database (Scielo.org), the database would stay idle and not produce results (several trials for the combined search was done both from Cuba and from Canada over 2 years (2006-2008) before the combined search was abandoned). However, the search with the public health AND primary care keywords combined would have yielded fewer articles than the search with the keyword primary care alone, supplemented by a search with the keywords health determinants alone. As both those other searches were performed and were both fully reviewed, it did not appear necessary to pursue this combination search further. It seems that the database is not set up to handle this type of combined search. The research team was confident that the search with the other keywords alone and the request for relevant articles from key experts (described below) yielded the most relevant articles.              32  Figure 5 SciELO database search results combining public health and primary care keywords    The third source of documents was through asking key informants in Cuba about relevant documents, including case examples of intersectoral collaboration in Cuba. The key informants interviewed included local primary care practitioners as well as provincial and national leaders in primary care, public health, and intersectoral collaboration in Cuba. I participated in collecting some of those documents during two field trips to Cuba (in 2004 and 2006). Along with other team members, I reviewed all the documents and case examples collected, and participated in summarizing the findings, which were reviewed by the Cuban team before publication. I paid particular attention to the documents related to the inclusion of primary care and public health professionals in intersectoral collaboration in Cuba.  As part of the two field trips, I engaged in key informant discussions of the roles of primary care practitioners, public health professionals, and members of other sectors. The field trips included  33 visits to organizations deemed by the local Cuban team to be relevant examples of primary care, public health, and other sectors’ joint engagement in the management of health determinants. Those field visits and key informant discussions were recorded through extensive note-taking, followed by team discussions including both Cubans and Canadians to ensure adequate understanding of the examples. The key informants included: two primary care physicians deemed to be typical of primary care physicians in the region (one more in the city centre, and one more on the periphery); a primary care physician seen as a local leader in intersectoral collaboration to manage health determinants; the director of the provincial healthcare management team; and two municipal public health specialist physicians. The sites visited included: two family physicians’ clinics; the intersectoral analysis unit of Santa-Clara; the regional hospital; the regional center for public health and epidemiology; a polyclinic (in which was located a public health physician); a maternal health centre; a centre for the elderly; a school with a community library; and a training facility for public health technicians.  The focus groups and surveys of the full case study were conducted in two municipalities: Santa-Clara, an urban city, and Camajuaní, a rural municipality. This thesis focuses on the management of health determinants in urban settings, due to feasibility considerations for the realist review and literature that seem to indicate that intersectoral collaboration in rural areas might follow different mechanisms than in urban settings, as mentioned in the realist review method section. Therefore, I excluded the data from Camajuaní from this re-analysis. The following sections describe the focus groups, survey participants, settings, and specific methodology used to re-analyze the data in relation to the main research questions of this thesis.  2.2.1 Focus group and survey participants The Canada-Cuba research team chose Santa Clara, a municipality in the central province of Villa Clara, Cuba, for the in-depth study. The rationale for that choice was that Santa Clara is a typical Cuban urban and industrial setting, with a population of more than 238,000 (Alegret et al., 2003), while Havana, the capital, is unique and would not have informed the stakeholders on intersectoral collaboration in a typical urban setting. Santa Clara was also a city with several researchers interested in studying health determinants in Cuba (Alegret et al., 2003).   34  Scientists and practitioners in Cuba and Canada contributed to adapting the survey and focus group instruments from a similar study that had been conducted in Canada (Frankish et al., 2007). Two workshops held with senior scientists and decision makers from the Cuban health sectors refined the research methodology and instruments, and reviewed them for face and content validity (Global Health Research Program, 2008a, 2008b).    In Santa Clara, participants and participating institutions were identified through purposive sampling based on their known involvement in managing health determinants as local and municipal leaders, primary care practitioners, and public health professionals. With ethical approval granted in Canada and Cuba, focus groups and questionnaires were conducted during the first quarter of 2007. An experienced member of the Cuban research team who was familiar with the topic facilitated each focus group and ensured informed consent from participants in the study. Another person was responsible for recording the discussion. Each focus group lasted about two hours, and was conducted immediately after participants completed the surveys.  All the participants were asked to answer the surveys individually. The Santa Clara case study included six such meetings, all with purposive invitation (purposive sampling) to participants, in locations deemed generally typical of the Santa Clara health system by the Cuban members of the research team. Participants included: long-standing decision-makers overseeing primary care in that region; local primary care, public health, and preventive medicine specialists very familiar with the region’s intersectoral stakeholders; and national leaders and researchers in the area of intersectoral collaboration.    Two such meetings were conducted with local leaders from four different “Consejos del Poder Populares”, or Popular Power Councils. Within the Cuban government, Popular Power Councils are the lowest level of organization - the neighbourhood or circumscription level. The members of the Popular Power Councils are elected through local elections held every three years. Two other focus groups combined primary care practitioners, including family physicians and community nurses, from the local primary care clinic, with members from the area’s polyclinic. The polyclinics have a mandate to support family physicians and are staffed with social workers, specialists in paediatrics, internal medicine, obstetrics, and gynaecology, and public health professionals.    35 One focus group was conducted with municipal leaders from the municipal assembly. The municipal assembly is composed of elected members from each of the Popular Power Councils, as well as non-elected representatives of Cuban ministries, institutes, sectors, and large civil organizations. The municipal assembly members are also members of the municipal health council, a structure responsible for addressing health issues at the municipal level. This council is composed of members of the municipal assembly as well as health services decision makers and clinicians.  The last focus group was conducted with members of the Municipal Unit of Hygiene and Epidemiology, and included public health professionals and decision makers. There are Municipal Units of Hygiene and Epidemiology in all of Cuba’s 169 municipalities. They report (along with all the other health and healthcare institutions of the municipality) to the Municipal Health Directorate, which is accountable to the municipal level of government, as well as the Provincial Health Directorate, which in turn is accountable to the provincial level of government and the Ministry of Public Health (Castell-Florit Serrate et al., 2007a) .   As we explained in Spiegel et al. (2012) p. 17:  “All recruited participants responded positively and participated fully. The only exception was the second focus group with the Popular Power Council members in Santa Clara. Due to an unforeseen local event needing the immediate attention of some committee members, only one person from each Popular Power Council attended this focus group. All focus group participants completed the questionnaire.”  After obtaining informed consent, but prior to answering the survey, a document was distributed to all participants that defined the terms ‘health determinants’ and ‘intersectoral collaboration.’ These were based on the Public Health Agency of Canada definition (PHAC, 2006)  and the WHO International Conference on Intersectoral Action for Health  proceedings (WHO, 1997), but adapted to the Cuban context based on the recommendations of the Cuban research team. The adaptation by the Cuban team mentioned that political will, social organizations, and the health system can all influence health, but are not the only influences. The document explained that other factors determine health, such as life circumstances (e.g., physical and social environment, personal health practices, etc.). The final document in Spanish is reproduced in Appendix A.3. The goal of providing this document was to promote a common understanding of these terms, and to give examples to facilitate answering the survey.   36  The survey form collected the following identification information: focus group number; the  participants’ organization; and their location (municipality and local area). Below, I have summarized the main survey and focus group questions. The final versions of the focus group guide, survey, and terminology document were produced in Spanish. They are reproduced in section Appendix A. The health determinants in the survey were those mentioned as key health determinants by the Public Health Agency of Canada (PHAC, 2006), which were also used in the study conducted by Frankish et al. on the management of health determinants in Canada (Frankish et al., 2007). To facilitate understanding of the content of the documents, below is an English-language summary of the questions asked in the focus groups and survey.   There were six main questions that guided the discussion in the focus groups:  1) What constitutes a health determinant in Cuba?  2) How are health determinants prioritized?   3) Who is responsible for addressing health determinants; how is it done internally; and how is the intersectoral collaboration achieved?  4) Which of the strategies of the Ottawa Charter on health promotion are used to manage health determinants? (The list of the Ottawa Charter strategies are from (WHO, 1986)) 5) How does the healthcare sector prevent diseases and promote and protect the health of the population?  6) What are the successes and challenges of intersectoral collaboration, and how are they evaluated?   The four main survey question themes were:  1) What are the five most important health determinants for your organization (rank them in order of priority, from 5 to 1, 5 being the most important)?  2) For each health determinant, what is the level of action carried by your organization?  1) no action; 2) recognition of the relevance of the health determinant; 3) planning action; 4) action implemented recently; and 5) long standing program of action. 3) For each health determinant, please indicate what is the most important action conducted by your organization. 4) For each health determinant, please list the most important organizations with which your organization has collaborated in the last year and indicate the frequency and the level of  37 collaboration reached between the two organizations. The options for the frequency included: 0-2; 3-6; more than 6; unknown. The levels of collaboration option were: 1) no collaboration; 2) minimum informal collaboration; 3) collaboration through formal mechanisms of exchange of information, without joint action; 4) joint planning of action; and 5) joint planning, implementation and evaluation.    The following sub-sections will describe the analytic process followed to extract meaning from the focus group transcripts, followed by statistical methodology to analyse the survey answers. I chose this order as the focus group provides rich information that can then be supplemented by the survey answers, which contribute to their interpretation.   2.2.2 Focus group methods Focus groups consist of interviewing a small group of individuals, emphasizing the interaction between the participants and the researcher(s), to gain important insights and information about a specific phenomenon (Krueger, 1994b). The main feature of focus groups that differentiates them from other types of group interviews is that the researcher takes a central role in encouraging focus group participants to interact with each other (Barbour & Kitzinger, 1999). Through interactive discussion, important insights and information can emerge, underscoring the importance for the researcher to encourage participant interaction (Krueger, 1994b). Even though focus groups encourage divergent views and opinions (Powell & Single, 1996), Krueger argued that because participant interaction involves discussion among the participants, the information forwarded by the group can be indirectly validated since everybody has the chance to confirm or challenge any part of the discussion.    The use of focus groups has both advantages and disadvantages.  One advantage of using focus groups is the possibility of discovering a wide range of information because divergent views are encouraged (Powell & Single, 1996). Moreover, because focus groups are rooted in participant interaction, Powell noted that the discussion in focus groups is rich, and can contain deep elaborations.  Because there are several participants, Powell also suggested that the possibility of neglecting several components of a topic becomes less likely in focus groups. Moreover, Morgan and Krueger (1993) noted the synergistic effect of a group discussion that is both dynamic and productive.  It is expected that a comprehensive discussion will be produced in focus groups  38 because participants question each other and have an opportunity to agree or disagree with each other (Morgan & Krueger, 1993).  Regarding the disadvantages of the focus group research method, one area that can be considered a weakness is the possibility for a group effect (Carey & Smith, 1994). Even though divergent views are encouraged in focus groups, sometimes participants may resort to group thinking, and individual participants may not express his or her true opinions on a subject for fear of contradicting the group (Powell & Single, 1996). In addition, even though the moderator has the central role in encouraging productive discussion among the group, moderators can sometimes be a hindrance to a more productive focus group by disrupting conversations (Powell & Single, 1996). Finally, another possible weakness of focus groups is the possibility that some topics may not be easily discussed within a group setting, because some participants may find a topic sensitive (Krueger, 1994b; Morgan & Krueger, 1993). Thus, the comprehensiveness of the discussion could be limited when some participants decline to take part in the discussion.   The limitations described above are generic limitations of focus groups. In the context of performing a supplementary analysis of a case study that was conducted with a slightly different goal, without the availability of the researchers who conducted the focus groups, and without full transcripts, it is difficult to ascertain how those limits might apply specifically to this study.  However, the Cuban research team members who conducted the focus group had extensive training and experience in conducting focus groups as per best practices, and sought to create a climate in which discussion, including agreement and disagreement with other participants’ views, was encouraged. The experience of the Canadian members of the research team, who were part of other focus groups conducted in collaboration with the Cuban researchers, confirmed that it was indeed the case that discussion was encouraged, and open challenge or support to the views expressed was frequent. This could, in part, be ascertained through the analysis, as it was clear that a variety of opinions were expressed and examples were given.  The Cuban research team provided the data from the focus groups in two documents. One document was labeled “resumen” (which will be referred to as the summary transcript), and the other one was labeled “relato” (which can be translated literally as ‘story line’ although does not refer to traditional story line extraction of focus group as understood in Canada). The summary transcript contains the participants’ answers to the main focus group questions, in the form of lists  39 of themes and key expressions for each of the questions discussed. Within the document, lists of themes and key expressions were grouped in three broad focus group categories: (1) local leaders, (2) municipal leaders, and (3) primary care and public health professionals. The document labeled “relato” by the Cuban team is referred to as the general transcript in this thesis. It contains the transcript of the discussion in the form of quotes.  It does not attribute quotes to any one participant, or to any one focus group. Instead, it has all the focus group results in one document, divided by the broad questions discussed. However, it is not divided by focus group categories. The Cuban research team was not able to provide full transcripts of the focus group discussions.   Furthermore, by the time this secondary analysis occurred, the Cuban team was not able to provide a separate set of answers from the two primary care focus groups versus the focus group consisting only of public health professionals. However, for each of the original individual focus groups, the Cuban research team was able to provide tables with the group’s consensual ranking of the health determinants. The general transcript has some sections labeled as responses from the healthcare sector, which had merged the answers from the public health and the primary care focus groups. There are some limitations arising from having only two documents describing the content of the six focus groups, with both documents appearing to fall short of full narrative transcripts. I discuss these limitations in the concluding chapter.  Within this thesis, the analysis of the focus group transcripts in combination with other qualitative data, such as the field visit, document review, and key informant interviews, is a hybrid between constant comparison analysis and thematic analysis. Thematic analysis seeks to extract the core patterns in a life story, as described by Sparkes and Smith, (2013) based on Riessmans’ (2008) book titled ‘Narrative methods for the human sciences’ (Sparkes & Smith, 2013). In this thesis, the intersectoral management of health determinants involving primary care, public health and representatives of other sectors in Cuba is the subject of this ‘life story’. Thematic narrative analysis is performed through immersion in the data, followed by identification of key themes, making connections across the evolution of the ‘life story’, in this case, across the evolution of intersectoral collaboration in Cuba. This supports the extraction of patterns and meanings constructed over time, by asking questions such as “‘What is going on here?’, ‘What does this theme mean?’, ‘What are the assumptions underpinning it?’, ‘What are the implications of this theme?’, ‘What conditions are likely to have given rise to it?’, and ‘What is the overall story the different themes reveal about the topic?’” (Sparkes & Smith, 2013) p. 133. I used this type of inquiry to gain further insight from the  40 various sources of information, asking questions such as why environmental health concerns seem to be a lower priority in one set of answers, while in another section, they seem to be considered a priority area to address.   However, because the data available is not in the form of a life story per se, much of the understanding of various practitioners’ roles and of how intersectoral collaboration functions in Cuba was derived through more classical constant comparison approaches. The method of constant comparison looks at patterns, themes, and perspectives of the participants, enabling the researcher to assess similarities and differences between the various actors (Boeije, 2002; Sparkes & Smith, 2013). In our case, the four groups of actors were the local leaders of various sectors, municipal leaders of various sectors, the primary care practitioners, and the public health professionals. To produce a systematic analysis of patterns, I alternated between the various categories, properties, and core roles and specific conceptualization and prioritization of health determinants of the various groups that seem to be emerging from the focus groups, as recommended by (Bowling, 2009).   Constant comparison analysis has been used extensively in grounded theory methodology (Strauss & Corbin, 1990). It is also frequently used with a more descriptive and interpretative aim (Thorne, 2000), which is an approach more appropriate to this study’s goals and purposes. The more descriptive and interpretative approach assists in establishing the similarities and differences between the various stakeholder groups, in terms of their broad conceptualization of, prioritization of, and respective roles in the management of health determinants. This, in turn, provides relevant insight for the interpretation of how the contextual elements, mechanisms, and outcomes of collaboration in Cuba are configured. This is relevant to the later part of integrating the Cuban findings with those of the realist review, which is based on analysis of configurations of contexts, mechanisms, and outcomes, as explained in section 2.3.  Both the summary transcript from the professional group and the general transcript fall into the category of semi-structured data, following the questions in the interview guide very closely. Therefore, data groupings (or nodes) were created for each health determinant, each question, and each emerging sub-theme and important concept discussed. Those were linked and compared across professional groups, with extensive use of memos and annotations. The analytic process, including refinement of categories, emerging themes, similarities, and differences was done using NVivo 9 from April to July 2011.  41  The end of the process of constant comparative data analysis occurs when data saturation is achieved, wherein the constant comparison of data elements does not yield new categorizations or insight, a signal that adding new data can no longer improve the findings (Strauss & Corbin, 1990). Data saturation pertains to both the data collection and the data analysis processes. During the focus groups, the Cuban researchers sought to reach data saturation by using sub-questions, probing questions, and contrasting questions about what they were understanding from various participants, until they determined that the discussion had reached its full potential and was not yielding any new insight (Barbour & Kitzinger, 1999; Krueger, 1994a, 1994b; Rabiee, 2004; Strauss & Corbin, 1990).   There are several limitations in terms of the level of saturation that can be reached in the supplementary analysis performed as part of this thesis. It was not possible to organize more focus groups, or to ensure that more probing questions were asked in relation to the roles of the various stakeholders, or even to have the full transcripts. However, based on the data I had at hand, I believe I have reached saturation in the analysis. In discussion with both Canadian and Cuban team members, I was not finding new insights, and in fact, many of the details of the differentiation of categories I established toward the end of the comparative analysis were deemed too detailed by other team members, leading to a state of “losing the forest for the trees”.   Furthermore, during the analysis process, many questions arose in relation to some of the specific answers in the focus group. This was also true of the survey, but to a lesser extent. I wanted to ensure I properly understood the answers. I had extensive discussions with the Cuban researchers as well as with key informants6, during my two field trips to Cuba in 2006 and 2009, through e-mail and a few international teleconferences, and in person during the multiple occasions that members of the Cuban research team came to UBC between 2006 and 2011. At several points during the analysis and interpretation of the findings, including while the team was writing the summary article by Spiegel et al. (2012), the Cubans had the opportunity to review my drafts and provide further comments. In addition, most of the information exchanged during the discussions and reflections                                                  6 Dr. Pastor Castell-Florit Serrate, a preeminent Cuban researcher in the field of intersectoral management of health determinants; Dr. Barbara Martinez, who was the Director of the Provincial Centre of Hygiene and Epidemiology of Villa Clara, and had been the head of the municipal health directorate, the body responsible for the health system service delivery in Santa Clara; Miriam Concepcion and Mariano Bonet of the National Institute of Hygiene, Epidemiology and Microbiology; as well as the several family physicians I encountered during the field, though unfortunately, I did not take note of their names.  42 that arose in the course of reviewing the literature, preparing the methodology and various instruments, and performing the analysis, have been recorded in field notes and diaries. All field notes and diary entries were reviewed toward the end of the thesis redaction to ensure trustworthiness, rigour, and comprehensive inclusion of information, as well as saturation of the insight gained through the analytic processes.  2.2.3 Survey analysis The supplementary analysis aimed to establish whether there were similarities or differences in the pattern of answers of the various stakeholders of interest (public health professionals, primary care practitioners, local leaders, and municipal leaders) for the four main question themes listed on page 30. The comparison of the various stakeholders’ answers was performed with the hope that the various patterns of answers would provide insight on how health determinants are managed in Cuba. The analysis plan was determined in consultation with a biostatistician affiliated with UBC’s Department of Family Practice. All analyses were performed with SPSS v.19. Both the statistician and I contributed to re-coding the data by professional group, type of organization, and type of action, as well as by intensity and frequency of engagement in intersectoral collaboration. These groupings were discussed with the Cuban research team to ensure accuracy.   For the purpose of this thesis, I use ≤.05 as the threshold for statistical significance. Due to the small sample size in each professional group, and since this is a re-analysis of a data set that was originally meant to compare municipalities (not professionals within a given municipality) I also use p<0.1 as the threshold for a tendency toward significance. Ultimately, both these levels are arbitrary, and the full p-values are displayed in the results.   For survey theme #1, respondents were asked to rank the five most important health determinants from the list provided (from 1 to 5, 5 being the most important). For survey theme #2, each respondent was also asked to rank their organization’s level of participation in the management of each health determinant on a Likert scale from 1 (‘no action’) to 5 (‘our organization has been acting on this health determinant for a long time’).   For the purposes of answering the questions posed in this thesis, for each health determinant, I averaged the rank given by each professional group, for each of the research themes. To determine  43 which statistical test was appropriate to assess differences in these rankings, the normality of the distribution had to be assessed. This was done by visual inspection of the distribution and by comparison of mean and median values, as well as through the standard normality test available in SPSS (Kolmogorov-Smirnov and Shapiro-Wilk) (SPSS, 2010). The visual assessment, the comparison of the mean and the median, and the normality test all revealed that the data were not normally distributed.  When normality can be assumed, the differences between the various groups can be assessed with a one-way analysis of variance (ANOVA) (Dawson-Saunders & Trapp, 1994; SPSS, 2010). If the data are not normally distributed, they can be analyzed using a Kruskal-Wallis test (Dawson-Saunders & Trapp, 1994), the non-parametric equivalent of the one-way ANOVA (Dawson-Saunders & Trapp, 1994). This test is appropriate for comparing several groups with different sample sizes, variances, and distributions, as it transforms the data to compare ranks of the whole distribution, rather than the original numerical values in each group (Dawson-Saunders & Trapp, 1994). The transformation involves ranking all the answers of all the respondents, regardless of professional group membership. Any tied value is assigned the average of the rank they would have been assigned had they not been tied, and the calculation is then adjusted for the number of ties. When the null hypothesis is rejected, it means that at least one of the group’s distributions is statistically significantly different than the others, but it does not specify which one.  If a difference in distributions is found, a pair-wise comparison, still using Kruskal-Wallis test, can be performed to determine between which groups lays the statistically significant difference(Dawson-Saunders & Trapp, 1994; SPSS, 2010).   For survey theme #3, participants were asked to name the most important action done by their organization to manage each health determinant. For purposes of this thesis, the actions listed for each health determinant were regrouped in various categories of interventions, and by each professional group. This enabled a description of the frequency of various types of actions. However, the small number of times each action was mentioned by each of the groups (even regrouped in categories) did not enable further meaningful statistical analysis by action.  For survey theme #4, each respondent was asked to name up to three organizations with which their organization engages in intersectoral collaboration, and identify the corresponding intensity and frequency of collaboration for each of those organizations. At first, the general pattern of  44 collaboration by health determinants, for all groups together, is described through simple counts and proportions. It was not necessary to carry out advanced statistical testing when the goal was a simple general description of the total pattern. However, to analyse whether there were differences between the professional groups, in accordance with the goal of this thesis, statistical comparisons were performed. Since departures from normality can occur, and because most of the intersectoral categorical variables are more akin to nominal categories than numerical ones, the differences between the various groups in terms of patterns of collaboration were assessed through the Chi-square test. The Chi-square test is the most appropriate statistical test to determine whether there are differences in frequencies or proportions between three or more independent groups (i.e., the various professional groups in this study) (Dawson-Saunders & Trapp, 1994; SPSS, 2010).   Overall, the respondents listed 73 organizations. Removing duplicates reduced the list to 35, which is still too many to enable a meaningful comparison of patterns of intersectoral collaboration between the various groups. To obtain conceptually sound groupings of those organizations and to remain consistent with the study’s goal related to intersectoral (not inter-organizational) collaboration, all organizations were classified into sectors. This regrouping into seven sectors also ensured there were no small frequencies expected, which could inflate the Chi-square value (Dawson-Saunders & Trapp, 1994).   The final classification was based on descriptions of the various organizations in Cuban documents and Cuban governmental webpages, as well as through discussion with the Cuban research team. Each of the organizations was grouped into one of seven sectors: education; community organizations; communication; political organizations; public work and governmental environmental institutions; health and social work; and commercial, industrial and financial organizations (details of the grouping from the original data can be made available). Organizations overlapping a few categories were categorized in the one sector judged to be the most pertinent by myself and another research team member very familiar with Cuba7. For example, the Association of University Students was grouped with the other organizations of the educational sector. Some organizations might appear to be more political, but were clearly described by Cubans as more akin to civil society/NGO functions than political organizations.                                                   7 Nino Pagliccia, a UBC statistician part of the Cuba-Canada team studying intersectoral management of health determinants, and who has worked with Cuba for several decades in a variety of professional and volunteer capacities.  45  Even after reducing the sectors to the groupings mentioned above, the distribution of intersectoral collaboration by sector, health determinant, and professional group produced some cells with small frequencies. The expected frequencies were rarely below five, which indicates that the Chi-square is still a valid statistical test. Furthermore, supplemental statistical analysis performed on a grouping of organizations in an even smaller number of sectors and with the merging of the Santa Clara and the Camajuani data sets, still showed statistically significant differences in patterns of intersectoral collaboration (analysis not shown8). This supports the robustness of the Chi-square with the original sectorial grouping described in the prior paragraph.    The intensity of collaboration was measured on a scale of 1 (no collaboration) to 5 (full collaboration, described as long-term joint planning, execution, and evaluation of a program or action to address this health determinant). The ‘level of collaboration’ variable was transformed into two categories (full collaboration versus all other levels of collaboration) to account for a ceiling effect observed in the answer, as most answered at the level of full collaboration. This variable thusly became dichotomised. Similarly to assessing the sectors with which each organization collaborates, Chi-square tests are the most appropriate to assess whether there are differences between the professional groups in terms of intensity of collaboration, in total or by sector. The frequency of collaboration is measured through a categorical scale, from 0-3, 3 to 6, or more than 6 formal interactions per year. For each health determinant, the differences in frequency of collaboration between the professional groups are also best assessed through the Chi-square test.  2.2.4 Integration of the supplementary analysis of the Cuban case study The re-analysis of the Cuban case study aimed to synthesize the diverse data sources in a manner that allowed for identifying the main contexts, mechanisms, and outcomes configurations. This iterative process started in 2011 and ended in 2013. It involved several other researchers, who reviewed the framework and the findings, and commented on their appropriateness until a framework that appeared valid and newly contributing to the world knowledge was created. This framework is presented in Chapter 3, section 3.4.                                                    8 Those analyses were not adding any more meaning or insight, and the amount of data and detailed analysis in the appendix is already substantial.  46 The approach of creating a framework representing the Cuban system based on categorization by contexts, mechanisms, and outcomes was chosen to facilitate bridging the findings of the realist synthesis of the literature with the findings from the Cuban re-analysis. As mentioned earlier, because the Cuban experience is likely not the only relevant one to consider in relation to how public health, primary care, and other sectors’ representatives manage health determinants. Thus, and to provide further insight into how to manage health determinants through that particular type of intersectoral collaboration, I decided to add a knowledge synthesis of the world literature to the Cuban supplementary analysis. This also increases the ability of this thesis to generate knowledge that will be useful in settings other than Cuba. Overall, the knowledge synthesis and the re-analysis of the Cuban case study data will contribute to creating a uniform realist inquiry to answer the research questions listed in the previous chapter.   The following section will present the methodology of the knowledge synthesis, which will also provide important insight on the rationale for using patterns of contexts, mechanisms, and outcomes to draw practical and relevant recommendations for decision-makers who implement interventions in a variety of contexts.  2.3 Realist synthesis This section will describe the rationale for conducting a systematic review, inspired by realist synthesis methods, followed by a detailed description of the steps performed in this systematic review based on realist review standards (Wong, Greenhalgh, Westhorp, Buckingham, & Pawson, 2013a).   Conducting a knowledge synthesis, especially based on realist review standards, is an appropriate methodology to deepen the understanding of how complex interventions work in a variety of circumstances (Pawson, Greenhalgh, Harvey, & Walshe, 2005). It is also a very timely process, since no systematic literature reviews have been conducted on intersectoral collaboration to manage health determinants before, let alone with the stakeholders of interest in this research, or with a realist approach.  Other methods considered for the systematic review include the use of a scoping review methodology, as was done to describe collaboration between primary care and public health  47 (Martin-Misener & Valaitis, 2008; Martin-Misener et al., 2012; Valaitis et al., 2012). Scoping review methodology could describe what the literature is reporting on the topic, through thematic analysis (Levac, Colquhoun, & O'Brien, 2010), however it does not possess the analytic process to determine theoretical structures, to determine how to best support collaboration, or to manage health determinants in various contexts (Dixon-Woods, Agarwal, Jones, Young, & Sutton, 2005). Scoping reviews of public health and primary care collaboration and various knowledge network reports express the need for increased research on the theoretical basis of collaboration between public health and primary care and on the management of health determinants (Health Systems Knowledge Network, 2007; Kelly, Morgan, Bonnefoy, Butt, & Bergman, 2007; Martin-Misener & Valaitis, 2008; Pfeiffer, 2003). These needs arise because of the paucity of publications in that area of research, and the hope to be able to design better interventions once a theory is built in order to explain why some interventions are more or less successful when they involve the stakeholders of interest of this research (Martin-Misener, 2009) (Pawson, 2002).    Grounded theory research approaches have been proposed as an avenue for synthesising research and generating a sound theoretical basis in emerging fields (Dixon-Woods et al., 2005). However, these approaches lack procedural transparency, which is important in systematic reviews, and they suffer from what some researchers call ‘methodological anarchy’, as there is a variety of different ways grounded theory is produced, without consensus on appropriate methods for various uses (Dixon-Woods et al., 2005).   Another method mentioned by Dixon-Wood et al. is Miles and Huberman’s cross-case technique (Dixon-Woods et al., 2005). WHO, the Public Health Agency of Canada, and many of the knowledge networks that fed into the CSDH report used case studies and cross case-study comparison to provide guidance on how to address health determinants (CSDH, 2008a; PHAC, 2007; PHAC & WHO, 2008). However, those reports did not generate very much analytic insight of what works when, and why or why not in various circumstances. Those reports also did not have systematic review methods to find the cases, using mostly expert-written cases studies, supplemented by a narrative approach to reviewing the literature. Furthermore, they suffer the limitations of such approaches in terms of lack of transparency in how searches are conducted, and how information is selected for inclusion or not (Dixon-Woods et al., 2005).    48 At the other end of the synthesis spectrum, meta-analyses - which attempt to quantitatively synthesize the impact of interventions - have limited utility for the study of the impact of intersectoral collaboration on health outcomes, due to the paucity of randomized or even control trial methods to assess the impact of intersectoral collaboration to improve health (Hayes et al., 2010). It is even more difficult to produce a systematic, quantitative comparison of outcomes across different interventions on health determinants, since those interventions are usually long-term, complex interventions prone to changing over time, and rarely amenable to exact replication when implemented in different settings, or even at different times in the same setting (Pawson, 2002; Pawson, Greenhalgh, Harvey, & Walshe, 2004; Pawson et al., 2005). Realist synthesis method appears to be the most appropriate type of systematic literature review for the questions asked in this thesis, considering realist reviews are more structured and purposefully take into account the issues in synthesising evidence from complex interventions to guide decision-making in changing circumstances (Pawson, 2002; Pawson et al., 2004; Pawson et al., 2005; Shepperd et al., 2009).  A realist review methodology is better suited to the research question posed by this thesis, by systematically assessing the context, mechanisms, and outcomes (CMO) configuration of interventions (Pawson et al., 2004). As explained in the Realist Synthesis Training Material, “A CMO configuration is a statement, diagram or drawing that spells out the relationship between particular features of context, particular mechanisms and particular outcomes. In a sentence, they take the form of “In ‘X’ context, ‘Y’ mechanism generates ‘Z’ outcome.”” Those material also cite an explanation provided by Jagosh et al (2012) “CMO configuring is a heuristic used to generate causative explanations pertaining to the data. The process draws out and reflects on the relationship of context, mechanism, and outcome of interest in a particular program. A CMO configuration may pertain to either the whole program or only certain aspects. One CMO may be embedded in another or configured in a series (in which the outcome of one CMO becomes the context for the next in the chain of implementation steps). Configuring CMOs is a basis for generating and/or refining the theory that becomes the final product of the review.”(Jagosh et al., 2012) p.316.   As Pawson explains, “In common sense terms, this amounts to saying that more consequential lessons are to be learned if we try to test out the same policy idea by seeing how it turns out in diverse settings.” (Pawson, 2002) p. 344. Furthermore, a realist review can assist in the creation of a sound theoretical basis for the field by extracting demi-regularities from the CMO configurations (Pawson, 2002; Pawson et al., 2004; Pawson et al., 2005; Pawson et al., 2011).  Demi-regularity are  49 defined as “semipredictable patterns or pathways of program functioning. The term was coined by Lawson (1997), who argued that human choice or agency manifests in a semipredictable manner—“semi” because variations in patterns of behavior can be attributed partly to contextual differences from one setting to another.”(Jagosh et al., 2012) p.317.   An important feature of realist synthesis is that it can help to build a framework representing the repetitive patterns found across data sources, which can then inform decision-makers and professionals as well as to help create or refine a middle-range theory. Middle-range theory “is an implicit or explicit explanatory theory that can be used to assess programs and interventions. “Middle-range” means that it can be tested with the observable data and is not abstract to the point of addressing larger social or cultural forces (i.e., grand theories).” (Jagosh et al., 2012) p. 316.  Through the use of demi-regularity from the Cuban supplementary analysis and the realist review, a realist inquiry approach provides a way to interpret and integrate those two study component in a middle-range theory and discuss the findings in relation to the rest of the world literature on the topic (Pawson, 2002; Pawson et al., 2004; Pawson et al., 2005; Pawson et al., 2011; Wong et al., 2013a).   Examining the causality of a phenomenon is the root of a realist synthesis, but the process is different from the causality arising from quantitative studies (Pawson et al., 2004). Causality is determined in realist synthesis through a successionist model (Pawson et al., 2004). In a successionist approach to causality, it is the repetition of finding certain mechanisms leading to a certain outcome across some, but not all contexts, that helps describe the true relationships between the variables. For instance, X can be concluded to cause Y, within context A, B and C; while in different contexts (F, G and H), X leads to Z, and does not cause Y. Furthermore, whereas X can be the mechanism to lead to Y in context A, it may actually serve as the outcome (Y) in context B. Therefore, realist inquiries enable the researcher to situate the outcomes within a more specific frame, which is not normally sought through other systematic review methods, such as meta-analysis (Shepperd et al., 2009).   In other words, realist synthesis assesses the relationship of two events, taking into account a specific context and the specific mechanisms needed to occur in order for the relationship to exist (Pawson et al., 2004). Bas