Are Integrated Approaches Working to Promote Healthy Weights and Prevent Obesity and Chronic Disease?

Transcription

1 Are Integrated Approaches Working to Promote Healthy Weights and Prevent Obesity and Chronic Disease? A Review and Synthesis of the Literature with Suggestions and Recommendations for Policy and Decision Makers Report Written by: Lindsay McLaren, Alan Shiell, Laura Ghali, Diane Lorenzetti, Melanie Rock, and Susan Huculak Additional Collaborators: Judy Birdsell and Mark Lowey Financial contribution from the Health Policy Research Program, Health Canada 2004 Centre for Health & Policy Studies, Dept Community Health Sciences, University of Calgary The views expressed herein do not necessarily represent the official policy of Health Canada 1

2 Project Investigators Alan Shiell, PhD Phone (403) ; Lindsay McLaren, PhD* Phone (403) ; Melanie Rock, PhD Phone (403) ; Diane Lorenzetti, MLS Phone (403) ; Laura Ghali, PhD Phone (403) ; Affiliation for all authors: Centre for Health and Policy Studies, Department of Community Health Sciences University of Calgary G230 Health Sciences Centre 3330 Hospital Drive, N.W. Calgary, Alberta, T2N 4N1 Fax: (403) * Please address correspondence to Lindsay McLaren 2

3 Acknowledgements We would like to thank the following Research Assistants for their invaluable contributions to this project: * Susan Huculak * Almaymoon Mawji * Alison Stuart * Stefan Ali * Beth Debruyne * Mathew Farrell * Karoline Guelke * Shariq Khoja * Giovanna Longhi * Rosemary Perry We would also like to thank the following individuals who participated in our project workshop, held in October 2003, and/or who provided feedback on an earlier version of the report: * Susan Crawford * Ann Ellis * Roxanne Felix * Elizabeth Gyorfi-Dyke * Dexter Harvey * Michelle Hooper * Bonnie Hostrawser * Pauline Poon * Cynthia Smith * Jayne Thirsk Finally, we would like to acknowledge our funding support: * Health Canada, Health Policy Research Program * Alberta Heritage Foundation for Medical Research * Canadian Institutes of Health Research * Centre for Health and Policy Studies * Institute of Health Economics * Markin Chair Program in Health and Society 3

4 Table of Contents Executive Summary.. Page 5 Background.. Page 5 What we found. Page 5 Recommendations for research and funding.. Page 6 Recommendations for policy. Page 7 Résumé de la Direction Page 10 Renseignements généraux Page 10 Nos constatations Page 10 Recommandations en matière de recherche et de financement. Page 11 Recommandations en matière de politiques Page 13 Main Report. Page The Problem Page How We Did Our Research. Page Results. Page How Integration is Presented in the Literature Page Descriptive Models of Integration - Obesity.... Page Examples of Vertically Integrated Interventions.. Page Examples of Horizontally Integrated Interventions. Page Healthy Public Policy... Page Implications for Policy and Decision Makers. Page Trends: People and Places. Page Something is Not Working: the Nature of Evidence.. Page The Nature of Community. Page Existing Evidence and Gaps.. Page Does Integration Work?.. Page Could Integration Work?. Page Where to Next?. Page 29 References.. Page 32 4

5 Background Executive Summary Chronic non-infectious illnesses, such as cardiovascular disease, type 2 diabetes, and cancer, are the leading cause of death and disability in Canada and worldwide. One risk factor common to these health problems is excess body weight, often described as obesity. The proportion of Canadians who are overweight or obese has increased markedly in recent decades, making this an important contemporary public health problem in Canada. Various studies show that: the prevalence of obesity among Canadian adults increased from 9.7% in the 1970s to nearly 15% in 1998; among Canadian children, the prevalence of obesity increased from 2% in 1981 to nearly 10% in 1996; the direct cost of treating and managing obesity in Canada has been estimated at $1.8 billion (1997 dollars), representing 2.4% of total health care expenditures in Canada. While the bulk of existing research has envisioned excess weight as a characteristic of individuals, it is more appropriately viewed as a problem of populations. This is because the primary contributors to excess body weight physical activity and diet are not behaviours that are chosen by individuals in isolation. Rather, they are behaviours whose nature and frequency are nested within and constrained by various circumstances. These include: social (e.g., social norms), economic (e.g., the distribution of income across a population), cultural (e.g., cultural practices around food and the body), political (e.g., agricultural and taxation policies), and physical (e.g., nearby facilities and resources) circumstances. As a result of such circumstances, entire populations have become heavier over time, and the causes of these trends are best understood as population-wide, rather than as characteristics of individual persons. Further, because of the multi-faceted, multi-level, multisector, and population-wide nature of risk factors for excess body weight and related chronic diseases, these health outcomes lend themselves to an integrated approach to research and intervention. To examine whether existing integrated approaches among populations are effective in this context, we conducted an extensive review of the literature on integrated approaches to the prevention of excess weight and chronic diseases among populations. In other words, we reviewed the literature on integrated strategies to promote healthy body weights, and thus prevent obesity and related chronic diseases, in entire populations (including schools, worksites, and whole communities). What we found: Discussion of integration in the literature was mainly conceptual or theoretical, as opposed to taking the form of a specific action plan. This makes it difficult to understand the feasibility of implementing an integrated strategy, and to understand its impact. Integration took different forms in the literature, including vertical integration (i.e., a focus on a number of different levels of influence such as individual-level knowledge or attitudes, and the physical environment); and horizontal integration (i.e., integration across organisations, or sectors, designed to increase capacity, maximise resources, and minimise duplication of effort). The majority of interventions detected in our literature review were of the vertical 5

6 integration sort, and took place in discrete settings such as schools, worksites, or communities. Interventions involving horizontal integration were infrequently encountered, and when they were encountered, they were unlikely to have included an outcome evaluation. Evaluation of integrated interventions can be tricky. Many integrated interventions do not lend themselves to a randomised control trial (RCT) design (considered a powerful design for gathering evidence), and even when a control group is included, it is difficult to discern an intervention effect. While we should not abandon the RCT format, we need to remain aware of this issue. There is a large amount of evidence available on vertical integration. This evidence is mixed, and overall there is no consistent, compelling portrait in favour of vertical integration. In contrast, we found evidence that some single-component (non-integrated) interventions can be quite effective. It is not possible to draw a conclusion about the impact or effectiveness of horizontal integration, since evidence on this type of integration is limited. However, the available evidence suggests that achievement of partnerships and collaboration alone has not halted continued weight gain among populations. We need to go beyond creating organisational capacity, to conduct and evaluate sufficiently resourced interventions (which themselves may or may not be vertically integrated). Upstream factors (i.e., social, economic, political, cultural circumstances) are rarely incorporated into intervention strategies, and interventions that target these influences are virtually absent in the health literature. Strong evidence of the importance of these factors is found in correlational research, and there is an emerging consensus that incorporation of these factors will be crucial to improving our current health profile. Overall, it must be noted that our collective efforts to promote healthy weights are not working, given the significant trend towards a heavier Canadian population. Recommendations for research and funding: An investigation into the non-health literatures, to understand the impact on health of policies and practices in other sectors. Social policy interventions (e.g., housing, child care, income supplement) and private sector practices (e.g., marketing, management) are unlikely to be found in the traditional published health literatures; in order to engage these public and private sectors, we need evidence of their impact which is likely to be found outside of the health literatures. An investigation into economic incentives for diet and physical activity, including food taxation policies; subsidised facilities, services, and equipment; and disincentives for driving such as congestion charges and higher parking fees. This investigation should consider the feasibility of implementing and maintaining such initiatives, and potential consequences. Sufficient funding to ensure proper evaluation of health promotion interventions. While evaluation should be considered for all health promotion interventions, the evaluation scope needs to be tailored to what the outstanding questions are (e.g., distribution of effects, unintended consequences of intervention, cause and effect). Depending on the nature of the project, this may require a long term commitment of funds. The necessity of accommodating long term evaluation funding is consistent with the life course perspective, which emphasises the influence of factors in early life upon health in adulthood. Evaluation of social policy interventions. Along with the need to evaluate interventions 6

7 conducted by practitioners in the health sector, we need to monitor the health impact of actions in the non-health sector such as those mentioned above: social policy interventions (e.g., housing, child care, income supplement) and private sector practices (e.g., marketing, management). A requirement that evaluation includes an assessment of distribution of impact. A desirable aim for population health interventions is to yield an equitable impact, as opposed to an effect that is concentrated in particular population subgroups (usually those of higher social status). At present, we don t know whether this is the case, as most interventions do not report this information. Continued funding for synthesis research, but with greater clarity and communication regarding the dissemination strategy. In the early stages of research, the researchers and potential end-users should work together in a formalised, resourced manner such as a teleconference, to identify the intended audience of the work, and the appropriate dissemination strategy; for example, alignment with national and/or provincial chronic disease prevention alliances. Establishment and maintenance of continuous, long term population-level surveillance for key variables at the individual and environmental level. This surveillance is essential to the accumulation of evidence on macro- and micro-level factors and their impact on changes in population health. Various sources of data should be exploited, including agriculture and transport sectors, and findings should be communicated to the public. Key priorities for surveillance include: o Individual self-report data on physical activity and, particularly, dietary intake (the latter is extremely infrequent in Canada), including information on car ownership, driving time, frequency of restaurant patronage, frequency of walking / cycling to work and school, and details of home food preparation. o Macro-level trends and policies that have implications for food and nutrition, including the actions of the food industry (e.g., food product content, portion size, advertising strategies, restaurant / outlet proliferation), and agricultural sector policies. o Macro-level trends and policies that have implications for physical activity, including transport, urban design, and land development. Funding for research-policy placements such as policymaker participation in the research process and vice versa, as a means of improving understanding of the different cultures, and reducing the intimidation that may be felt by working outside one s area of expertise and comfort. As a template for this interchange, one might draw from Health Canada s voluntary sector policy internship ( or the Canadian Health Services Research Foundation s researcher decision maker partnership ( An investigation into how to facilitate intersectoral integration in government. This could take place in the context of the research-policy placements, or it could be an opportunity available to students or trainees interested in the politics of health. This investigation could begin by looking at case studies of governments (regional, provincial, national) in which an incentive system to encourage cross-sector engagement has already been attempted (if any exist). Recommendations for policy: Action, with a commitment to evaluate. We have selected the following 7

8 recommendations on the basis of two considerations: first, on the basis of consistent correlational or causal evidence from etiological or intervention research; and second, on the basis of their potential to have an impact that is equitable across the population rather than concentrated among groups of higher social status. Thorough evaluation of the impact of these initiatives is essential and needs to be fully resourced. Because the following recommendations are based on different types of evidence (a defining characteristic of the literature), we assign an equal weight to each in terms of priority. o Regulation of advertising and promotion of foods to children. There is strong evidence in favour of an association between time spent watching television and excess body weight among children, and research suggests that the most compelling mechanism for this association relates to exposure to the promotion of nutritionally poor food items. Both excess body weight and time spent watching television are higher among children from lower income families, and efforts to reduce television consumption need to take this socioeconomic gradient into account. Advertisements during children s television programs predominantly feature highly processed, non-nutritious foods; these foods are in turn highly desired and over-consumed. Thus, we assert that targeting food advertisements may have a large and equitable impact, as opposed to admonishments to watch less television, which are likely to disproportionately reach families of higher socioeconomic means. o Improvement of the walkability of neighbourhoods A growing body of research supports an association between the walkability of neighbourhoods (e.g., presence of sidewalks, pedestrian/car separation, safety, aesthetic dimensions such as trees, variety of destinations, amenities such as benches) and both higher levels of walking, and lower body mass index. We assert that it is time to act on the basis of these associations, to make our neighbourhoods safer and more pleasant for walking (and other exercise), playing, and socialising. One suggestion is to devise and conduct a health impact assessment of major planning proposals, to ensure their attention to this issue. o Fiscal policies to facilitate healthy lifestyle There is evidence showing that price has important implications for food purchasing behaviour, and research suggests that, on a calorie per cost basis, less nutritious foods are cheaper. Suggestions for fiscal policies include subsidisation of healthier foods (e.g., fruits and vegetables, whole grains, low fat milk) as well as recreation or sporting opportunities; and taxation of less healthy foods (e.g., highly processed food items, foods containing trans fatty acids) as ways of facilitating a healthier diet, equitably across the population. o Whole school interventions to facilitate health Schools represent a means of accessing nearly every child, and therefore are ideal venues for shaping healthy development. Interventions should focus on whole school efforts, including policies to ensure nutritious, affordable, and appealing food in cafeterias and vending machines (examples: mandate availability of fruit and vegetables; restrict availability of foods containing trans fatty acids or high levels of saturated fat; consider issuing school food service contracts to local growers); ensuring regular physical activity and education around food preparation and time management skills; and providing opportunities for student involvement and engagement both in the classroom and in after-school activities so that students feel connected to each other and their environment. As a suggestion for monitoring, progress towards improvement in nutrition, physical activity, and social environment could be incorporated into school inspection criteria. o Whole worksite interventions to facilitate health 8

9 Worksites similarly represent a means of accessing a large proportion of the adult population. Interventions should be of a workplace-wide nature, including policies to enable flexible work hours, provision of child care facilities, provision of exercise and shower facilities, improvement in cafeteria food quality and price, and opportunities for employee involvement and engagement. An incentive system may need to be created, whereby worksites of all kinds are rewarded (e.g., with positive publicity) and compensated financially, for adoption of such structural changes. o Incentives for intersectoral integration in government. There is plenty of evidence in support of an association (correlation) between non-health sector factors (e.g., education, housing and employment conditions) and health, and the policy recommendations above will require intersectoral integration to occur. Currently, working outside of one s mandate is discouraged, and efforts to reverse this tendency are necessary. One suggestion is to implement a reward system for cross-sectoral engagement (e.g., agricultural sector is rewarded for aligning policies with public health goals); another suggestion, in line with recommendations in the UK, 1 is to appoint a specific public health committee to monitor health targets (related to obesity and chronic disease) across all relevant government sectors. 1 House of Commons Health Committee report on obesity: 9

10 Renseignements généraux Résumé de la Direction Les maladies chroniques et non infectieuses, comme la maladie cardiovasculaire, le diabète de type II et le cancer, constituent les principales causes de décès et d invalidité au Canada et dans le monde entier. L excès de poids, souvent qualifié d obésité ou d embonpoint, est l un des facteurs de risque qui caractérise chacun de ces troubles de santé. Au cours des dernières décennies, la proportion de Canadiens et de Canadiennes faisant de l embonpoint ou souffrant d obésité s est accrue de manière considérable, ce qui en fait maintenant un important problème de santé publique au Canada. Selon diverses études : le taux de prévalence de l obésité chez les adultes du Canada a augmenté pour passer de 9,7 % dans les années 1970 à près de 15 % en 1998; chez les enfants du Canada, le taux de prévalence de l obésité s est accru pour passer de 2 % en 1981 à près de 10 % en 1996; le coût direct lié au traitement et à la gestion de l obésité au Canada a été évalué à 1,8 milliard de dollars (en dollars de 1997), soit 2,4 % des dépenses totales en soins de santé enregistrées au Canada. Même si la plupart des études considèrent que l excès de poids est une caractéristique des particuliers, il s agirait davantage d un problème lié à la population. Cela s explique par le fait que les principales causes de l excès de poids soit l activité physique et le régime alimentaire ne sont pas des comportements qui sont choisis par seulement certaines personnes. La nature et la fréquence de ces comportements relèvent plutôt de diverses circonstances. Cela comprend des circonstances sociales (soit les normes sociales), économiques (soit la répartition du revenu au sein d une population), culturelles (soit les pratiques culturelles en ce qui a trait à l alimentation et au corps), politiques (soit les politiques en matière d agriculture et de fiscalité) et physiques (soit les installations et les ressources qui existent dans la région). De telles circonstances ont fait en sorte que des populations complètes font face au problème de l embonpoint depuis un certain temps, et il est plus facile de comprendre les causes de ces tendances quand on s attarde à la population entière et non pas aux caractéristiques des particuliers. De plus, puisque les «facteurs de risque» engendrant l embonpoint et les maladies chroniques connexes comportent diverses facettes, se retrouvent à plusieurs niveaux, touchent de nombreux secteurs et concernent toute la population, ces enjeux de santé se prêtent à une méthode intégrée de recherche et d intervention. Afin de déterminer si les méthodes intégrées actuelles qui sont employées au sein des populations portent fruits dans ce contexte, nous avons effectué un tour d horizon complet de la documentation déjà publiée sur les méthodes intégrées visant à prévenir l embonpoint et les maladies chroniques au sein des populations. Autrement dit, nous avons passé en revue la documentation portant sur les stratégies intégrées et visant à favoriser un poids sain et par conséquent, à prévenir l obésité et les maladies chroniques connexes et ce, au sein de populations complètes (ce qui comprend les écoles, les lieux de travail et des collectivités entières). Nos constatations Au lieu de prendre la forme d un plan d action particulier, la documentation portant sur l intégration était surtout de nature conceptuelle ou théorique. Par conséquent, cela permet plus difficilement de comprendre la faisabilité de la mise en œuvre d une stratégie intégrée et de comprendre les incidences d une telle stratégie. 10

11 Dans la documentation, l «intégration» prenait diverses formes, notamment l intégration verticale (c est-à-dire que l accent est mis sur divers niveaux d influence comme les connaissances ou les attitudes personnelles, et le milieu physique); et l intégration horizontale (c est-à-dire l intégration entre les organisations ou les secteurs et ce, dans le but d accroître la capacité, de maximiser les ressources et de minimiser le chevauchement des efforts). La plupart des interventions visées par la documentation examinée relevaient de l intégration verticale et prenaient place dans des milieux discrets comme des écoles, des lieux de travail et des collectivités. Nous avons pris connaissance de peu d interventions relevant de l intégration horizontale et les quelques interventions d intégration horizontale que nous avons pu examiner étaient peu susceptibles d inclure l évaluation des résultats. L évaluation des interventions intégrées peut s avérer difficile. Bien des interventions intégrées ne se prêtent pas à une étude sur échantillon aléatoire et contrôlé (ce genre d étude est considéré comme une excellente méthode de rassemblement des éléments probants) et même lorsqu il y a un groupe témoin, il est difficile de discerner l effet d une intervention. Cela ne signifie pas qu il faut laisser cette formule de côté, mais plutôt, qu il faut être conscient des restrictions qu elle comporte. Il existe beaucoup d éléments probants sur l intégration verticale. Ces éléments probants sont variables et en général, il ne se dresse pas de portrait cohérent et convaincant en faveur de l intégration verticale. Par contre, nous avons constaté que certaines interventions ne comprenant qu une seule composante (donc, une méthode non intégrée) peuvent être assez efficaces. Il n est pas possible de tirer une conclusion sur les incidences ou l efficacité de l intégration horizontale, car il existe peu d éléments probants concernant ce type d intégration. Cela dit, les éléments probants qui existent laissent croire que la réalisation de partenariats et la collaboration, à elles seules, n ont pas réussi à mettre un frein au gain de poids constant au sein des populations. Il faut donc aller au-delà de la création de la capacité organisationnelle et ce, dans le but de faire des interventions dotées de suffisamment de ressources et de les évaluer (dont leur intégration peut être verticale ou non). Il est rare que les facteurs «en amont» (c est-à-dire les circonstances sociales, économiques, politiques et culturelles) soient intégrés aux stratégies d intervention, et il n y a à peu près pas mention des interventions qui visent ces influences dans la documentation sur la santé. Par contre, nous avons pu retracer l importance de ces facteurs dans la recherche corrélationnelle et de plus en plus, il y a un consensus selon lequel l incorporation de ces facteurs jouera un grand rôle dans l amélioration de notre profil de santé actuel. Dans l ensemble, il y a lieu de remarquer que les efforts collectifs qui sont déployés dans le but de promouvoir un poids sain ne portent pas fruits, puisque la population canadienne a fortement tendance à afficher un excès de poids. Recommandations en matière de recherche et de financement Dépouiller la documentation «ne se rapportant pas à la santé» afin de comprendre les incidences des politiques et des pratiques d autres secteurs sur la santé. Il est peu probable que la documentation qui est publiée sur la santé fasse mention des interventions en matière de politique sociale (c est-à-dire en ce qui a trait à l habitation, aux soins à l enfant, aux programmes de supplément de revenu) et des pratiques du secteur privé (c est-à-dire en matière de marketing, d administration); afin de remporter la faveur de ces secteurs publics et privés, nous devons prouver leurs incidences, ce qui se retrouve probablement ailleurs que dans la documentation sur la santé. 11

12 Effectuer une recherche sur les stimulants économiques en matière de régimes alimentaires et d activité physique, ce qui concerne notamment les mesures relatives aux taxes sur les aliments; la subvention d installations, de services et d équipement; et les moyens pris pour dissuader les gens de se déplacer en voiture, comme des frais pour les véhicules qui occasionnent une congestion et des frais de stationnement plus élevés. Cette recherche devrait aussi tenir compte de la faisabilité de mettre en œuvre et de garder en vigueur de telles initiatives, ainsi que des conséquences possibles. Financer suffisamment l évaluation adéquate des interventions visant à favoriser la santé. Même s il faudrait considérer l évaluation de toutes les interventions visant à favoriser la santé, il y a lieu de déterminer l étendue des évaluations à la lumière des enjeux à l étude (par exemple, la répartition des effets, les conséquences non intentionnelles de l intervention, la cause et l effet). Selon la nature du projet, cela pourrait se traduire par un financement à long terme. La nécessité d accorder du financement à long terme cadre bien avec la perspective du déroulement de la vie, qui fait ressortir l influence qu exercent certains facteurs en jeune âge sur la santé à l âge adulte. Évaluer les interventions en matière de politique sociale. En plus de la nécessité d évaluer les interventions réalisées par les praticiens du secteur de la santé, nous devons surveiller les incidences sur la santé des mesures prises dans des secteurs autres que celui de la santé, comme ceux dont il est question ci-dessus : les interventions en matière de politique sociale (c est-à-dire en ce qui a trait à l habitation, aux soins à l enfant, aux programmes de supplément de revenu) et des pratiques du secteur privé (c est-à-dire en matière de marketing, d administration). Faire en sorte que l évaluation comprenne également l évaluation de la répartition des incidences. Il est désirable que les interventions en matière de santé destinées à la population se traduisent par des incidences équitables, et non pas par un effet concentré auprès de segments particuliers de la population (habituellement la population faisant partie d un rang social plus élevé). Pour l instant, nous ne savons pas si tel est le cas, car la plupart des interventions ne font pas mention de cette information. Continuer de financer la recherche de synthèse, mais en mettant davantage l accent sur la clarté et les communications en ce qui a trait à la stratégie de diffusion. Aux premiers stades de la recherche, les chercheurs et les usagers éventuels devraient travailler de concert de manière formelle et à l aide de ressources, comme dans le cadre d une téléconférence et ce, pour déterminer quel est le public visé par le travail, ainsi que la stratégie de diffusion appropriée; par exemple, s aligner avec des alliances nationales et (ou) provinciales s occupant de prévention des maladies chroniques. Établir et maintenir une surveillance continue et à long terme de la population pour ce qui est de variables importantes sur le plan personnel et le plan environnemental. Cette surveillance est essentielle à la collecte d éléments probants concernant les macro-facteurs et les micro facteurs ainsi que leurs incidences sur les changements caractérisant la santé de la population. Il y a lieu de dépouiller diverses sources de données, notamment celles des secteurs de l agriculture et du transport, puis de communiquer les constatations qui en émanent au grand public. Voici les principales priorités en matière de surveillance : o les données d autoévaluation individuelle relativement à l activité physique et, plus particulièrement, à l apport alimentaire (ces dernières sont plutôt rares au Canada), ce qui comprend des renseignements sur la possession ou non d un véhicule automobile, sur le temps passé au volant, sur le nombre de fois que les personnes mangent au restaurant, sur le nombre de fois que les personnes se rendent au travail ou à l école à pied ou à bicyclette, et des détails concernant la préparation des aliments à domicile; o les tendances et les politiques de macro-niveau susceptibles d avoir des 12

13 incidences sur la nourriture et l alimentation, ce qui comprend les décisions prises par l industrie alimentaire (c est-à-dire le contenu des produits alimentaires, la grosseur des portions, les stratégies publicitaires, la prolifération des restaurants et des comptoirs d alimentation) et les politiques du secteur agricole; o les tendances et les politiques de macro-niveau susceptibles d avoir des incidences sur l activité physique, ce qui touche notamment les aspects du transport, de l aménagement urbain et de l aménagement du terrain. Financer les affectations de recherche et d établissement de politiques, comme la participation des preneurs de décisions au processus de recherche, et vice-versa et ce, dans le but de donner lieu à une meilleure compréhension des diverses cultures et de faire en sorte que les personnes qui ne travaillent pas dans leur domaine de spécialisation ou qui ne sont pas à l aise dans un autre domaine se sentent moins intimidées. À cette fin, il est possible de consulter le site du Partenariat avec le secteur bénévole de Santé Canada ( ou celui de la Fondation canadienne de la recherche sur les services de santé pour en savoir plus sur le partenariat entre les chercheurs et les preneurs de décisions ( Réaliser une enquête afin de déterminer comment favoriser l intégration intersectorielle au sein des gouvernements. Cette enquête pourrait se faire dans le cadre des affectations de recherche et de politiques ou encore, elle pourrait être effectuée par des étudiants ou des stagiaires qui s intéressent à la politique de la santé. Pour commencer, les enquêteurs pourraient examiner les études de cas des gouvernements (régionaux, provinciaux et national) pour lesquels un système d encouragement a déjà été mis en œuvre dans le but de favoriser la collaboration intersectorielle (le cas échéant). Recommandations en matière de politiques Agir, tout en s engageant à évaluer. Nous nous sommes arrêtés sur les recommandations suivantes en raison de deux considérations : premièrement, en raison de l information probante causale ou corrélationnelle qui est constamment ressortie de la recherche étiologique ou de la recherche sur les interventions; et deuxièmement, en raison de leurs effets possibles et équitables sur la population en général et non pas seulement sur les groupes de rang social plus élevé. Il est essentiel de faire l évaluation approfondie de l effet découlant de ces initiatives et de faire en sorte que les évaluations soient bien documentées. Puisque les recommandations suivantes sont fondées sur différents types d éléments probants (soit une caractéristique déterminante de la documentation), nous accordons la même priorité à chacune des recommandations. o Réglementation de la publicité et de la promotion entourant les aliments destinés aux enfants Bien des éléments probants portent à croire qu il existe une corrélation entre le fait de regarder la télévision et l embonpoint chez les enfants. Selon diverses recherches, cette corrélation serait surtout attribuable à la visualisation de publicité concernant des aliments de mauvaise qualité nutritive. L excès de poids et le temps passé à regarder la télévision sont plus courants chez les familles au revenu moins élevé, ce qui fait que tout effort visant à réduire le nombre d heures passé à regarder la télé doit tenir compte de ce facteur socioéconomique. La publicité visionnée pendant les programmes télévisés pour enfants met surtout en vedette des aliments hautement transformés et non nutritifs; ces aliments sont fortement désirés par les enfants, qui les consomment en grande quantité. Par conséquent, nous avançons que le fait de cibler la publicité d aliments pourrait avoir un effet considérable et équitable, au lieu de préconiser des remontrances aux jeunes pour qu ils regardent moins la télé, ce qui est plus susceptible 13

14 d atteindre les familles ayant des moyens socioéconomiques plus élevés dans une mesure disproportionnée. o Amélioration de la «marchabilité» des quartiers De plus en plus de travaux de recherche démontrent qu il existe une corrélation entre la «marchabilité» des quartiers (comme la présence de trottoirs, de séparations entre les chemins pour piétons et ceux pour voitures, la sécurité, les aspects esthétiques comme les arbres, la variété des destinations, puis les commodités comme des bancs) et le fait de marcher plus souvent et un indice de masse corporelle moins élevé. Nous avançons donc que le moment est venu d agir à la lumière de cette corrélation, de faire en sorte que nos quartiers soient plus sécuritaires et qu ils se prêtent mieux à la marche (et à d autres exercices), au jeu et à la socialisation. Une suggestion consisterait à concevoir et à réaliser une évaluation des incidences sur la santé des grands projets d aménagement et ce, afin d attirer l attention de leurs auteurs sur ce point. o Mesures financières favorisant un mode de vie sain Divers éléments probants laissent croire que le prix exerce une grande influence sur l achat des aliments, et selon diverses études, les aliments moins nutritifs coûtent moins cher par calorie. En matière de mesures financières, nous suggérons la subvention des aliments plus sains (comme les fruits et les légumes, les grains entiers, le lait à faible teneur en gras) et les occasions de faire des loisirs et du sport; puis, nous suggérons l imposition des aliments moins sains (comme les aliments hautement transformés, les aliments contenant des acides gras trans) et ce, afin de favoriser un régime alimentaire plus sain de manière équitable au sein de la population. o Interventions visant toute l école dans le but de favoriser la santé Grâce aux écoles, il est possible d atteindre presque tous les enfants. Il s agit donc d un endroit idéal pour favoriser un développement sain. Les interventions devraient viser toute l école, et prendre notamment la forme de politiques visant à assurer la vente d aliments nutritifs, abordables et appétissants à la cafétéria et dans les distributrices (par exemple, faire en sorte que des fruits et légumes soient vendus, restreindre la disponibilité d aliments contenant des acides gras trans ou ayant de fortes teneurs en gras saturés, considérer la possibilité d accorder des contrats de restauration aux producteurs de la région); à assurer la tenue d activités physiques régulières et à donner des cours sur la préparation des aliments et la gestion du temps; à donner l occasion aux élèves de participer à diverses activités en classe et après les cours afin qu ils puissent se sentir plus proches les uns des autres ainsi que de leur environnement. Une suggestion de suivi visant à améliorer l alimentation, l activité physique et le milieu social consisterait à tenir compte des progrès dans le cadre des critères d inspection de l école. o Interventions visant tout le milieu de travail dans le but de favoriser la santé De même, les milieux de travail constituent une bonne façon d avoir accès à une grande proportion d adultes. Par conséquent, les interventions devraient viser tout le milieu de travail et prendre la forme de mesures concernant des heures de travail flexibles, des garderies, des salles d exercice avec douches, la vente d aliments de meilleure qualité et à meilleur prix à la cafétéria, et la possibilité pour les employés de participer et de s engager. La faisabilité de mettre en œuvre un système d encouragement devrait être considérée, en ce sens que les lieux de travail de toutes sortes pourraient être récompensés (au moyen de publicité par exemple) pour avoir adopté des changements de cette nature. o Encouragements en vue de l intégration intersectorielle au gouvernement 14

15 De nombreux éléments probants portent à croire qu il existe une corrélation entre des facteurs liés aux secteurs ne relevant pas de la santé (comme les secteurs de l éducation, de l habitation et des conditions d emploi) et la santé. Pour que les politiques et les mesures recommandées ci-dessus portent fruits, il faut préconiser l intégration intersectorielle. En ce moment, il n est pas bien vu de dépasser les limites prévues par un mandat de travail, ce qui signifie que des efforts doivent être déployés pour rectifier cette tendance. Une suggestion consisterait à mettre en œuvre un système visant à récompenser l engagement intersectoriel (par exemple, le secteur agricole serait récompensé s il faisait en sorte que ses politiques cadrent avec les objectifs de santé publique); une autre suggestion, qui est conforme aux recommandations émanant d un rapport sur l obésité publié au Royaume-Uni, 2 consisterait à nommer un comité spécialisé en santé publique dans le but d assurer le suivi des objectifs en matière de santé (pour ce qui est de l obésité et des maladies chroniques) au sein de tous les secteurs pertinents du gouvernement. 2 House of Commons Health Committee report on obesity: 15

16 1. The Problem Main Report Chronic illnesses, such as cardiovascular disease, type 2 diabetes, and cancer, are the leading cause of death and disability in Canada and worldwide (Stein, 1997). One risk factor common to these health problems is excess body weight, or obesity (Bray, 2003). The proportion of Canadians with excess body weight has increased markedly in recent decades, making this an important contemporary public health problem in Canada. Various studies show that: the prevalence of obesity among Canadian adults increased from 9.7% in the early 1970s to nearly 15% in 1998 (Katzmarzyk, 2002); among Canadian children, the prevalence of obesity increased from 2% in 1981 to nearly 10% in 1996 (Tremblay et al., 2002); the direct cost of treating and managing obesity in Canada has been estimated at 2.4% of total health care expenditures in Canada (Laird Birmingham et al., 1999). While the bulk of existing research has envisioned obesity as a characteristic of individuals that may be explained by their chosen lifestyle, it is more appropriately viewed as a problem of populations. This is because the primary contributors to excess body weight physical activity and diet are not selected by individuals in isolation. Rather, the nature and frequency of these behaviours are nested within and constrained by various circumstances. These include: social (e.g., social norms), economic (e.g., the distribution of income across a population), cultural (e.g., cultural practices around food and the body), political (e.g., economic and agricultural policies), and physical (e.g., nearby facilities and resources) circumstances (Hill & Peters, 1998; French et al., 2001). As a result of such circumstances, entire populations have become heavier over time; the causes of these trends are best understood as population-wide rather than as characteristics of individual persons (Rose, 2001). Because of the multi-faceted, multi-level, multi-sector, and population-wide nature of risk factors for excess body weight and associated chronic disease, these health outcomes lend themselves to an integrated approach to research and intervention. By integrated approach, we mean one that incorporates more than one risk factor or disease outcome, more than one level of influence, more than one disciplinary perspective, more than one type of research method, or more than one societal sector, and which targets populations rather than individuals as a unit. By population, we mean a group that includes both healthy and less healthy individuals, all of whom must be targeted in order to have a large and sustainable impact on population health. A population can refer to the collection of persons within a particular setting, such as a school, a worksite, or a neighbourhood. An important implication of targeting populations is as follows: risk reduction across a population even if modest has been argued to yield greater and potentially more sustainable benefits, than larger improvements among a few individuals who are most at risk (Rose, 2001). Based on these understandings, we conducted a critical review and synthesis of the published scientific literature and of the grey literature (including government and organisational websites), on integrated, whole-population approaches to the prevention of excess weight gain and related chronic diseases. Our aims in conducting this work were: to better understand what is meant by integration; to examine the evidence about integrated strategies; and to provide recommendations for research funding and policy. 16

17 2. How We Did Our Research We chose to review studies conducted with human populations, which were published in either English or French. For the published scientific literature, we opted to retain documents from January 1990 through December 2002, in order to obtain a record of 12 full years from each database. For the grey literature, we included the period from January 1990 onwards, with the end point left open in order to detect as much recent information as possible. Prior to conducting the actual literature searches, we developed a procedure for deciding whether a document from the literature was potentially related to an integrated approach. To qualify, a document had to relate to: primary prevention and/or health promotion; and a chronic disease associated with excess weight gain (cardiovascular diseases, diabetes & various forms of cancer); and o multiple risk factors (two or more risk factors including diet/nutrition and poverty/deprivation, etc ); and/or o multiple levels (two or more levels of influence on health; e.g. personal behaviour and public policies); and/or o multiple societal sectors (two or more defined societal sectors, including education, agriculture, etc ); and/or o multiple disciplines (two or more, typically academic, disciplines); and/or o multiple research methods (studies incorporating the use of both quantitative and qualitative research methods); and/or o studies described as being integrated or comprehensive, or similar terms. We searched a total of 24 medical and social science electronic databases to identify peer-reviewed studies appropriate for inclusion in our review. We also searched the grey literature, including key organisation and government websites. Using a document-screening template created for this project, we looked for examples of interventions that used integrated strategies, and other documents that provided specific guidance or recommendations on what integrated strategies entail, and how they can be implemented with the overall goal of being able to inform programs and policies to promote healthy body weights, and thus prevent obesity and related chronic diseases, among entire populations. For the published scientific literature, all abstracts were first examined independently by two research assistants (RAs), who looked for evidence of multiplicity. Next, full-text documents were read in full by one RA to ensure that reference to multiplicity represented integration (i.e., recognition of the synergy or benefit of multiplicity). Finally, project investigators examined the retained documents, and eliminated those that did not represent a population health approach (to be considered a population health approach, the document had to target groups rather than individuals, and groups had to include healthy as well as less healthy persons). For the grey literature, documents that related broadly to integrated approaches to the prevention of obesity and chronic disease were identified. Next, one RA read through these documents and pulled out those representing specific case examples or an explicit model or framework of integration. Unlike systematic literature reviews, our review did not exclude documents based on research design or quality. The reason for this is that the topic of our review integrated approaches targeted at populations does not easily lend itself to study designs typically regarded as most rigorous, such as the randomised control trial. We wished to be inclusive, and thus all documents that adhered to criteria above were included. 17

18 After completing our literature search (but prior to beginning the review and synthesis), we held a one-day workshop with various potential end-users of this document. This workshop fostered discussion around the concept of integration, and ways in which our report could be shaped so it would be of most use to those in a decision-making role. Research synthesis was conducted in two different ways. For the settings-based interventions (programs taking place in schools, worksites, and communities), details of all studies were recorded in a table, and themes ( types of studies) that emerged from the table were discussed. Though study design and quality were not considered as exclusion criteria, in our discussion we drew attention to studies for which effects are most easily discerned for example, those with control groups. For the other integrated initiatives (e.g., provincial or national initiatives, frameworks, etc), all documents were read, and a few were selected for discussion as representing a cross-section of those obtained. Once our report had been drafted (April, 2004), participants of the end-user workshop were contacted again to see if they were interested and willing to read the report and provide feedback. All participants agreed, and each provided feedback via a semi-structured telephone interview (April June, 2004). Feedback comments were summarised and, to the extent possible, incorporated into this final version (August 2004). It is felt that a strong degree of rapport was established through this process of involving the end-users throughout the research process, and the comments themselves were invaluable to the research team. 18

19 3. Results 3.1 How Integration is Presented in the Literature We found integration to be a popular term that is used worldwide to describe promising strategies for improving health. The term has been used by the World Health Organisation ( Health Canada s Centre for Chronic Disease Prevention and Control ( the Integrated Pan-Canadian Healthy Living Strategy ( and the Alberta Healthy Living Network ( among others. We also found some discussion of integration in the peer-reviewed academic literature. In general, these academic documents describe models or frameworks for understanding health that explicitly include attention to a person s setting or context as an influence on health (Kelly et al., 1993). For the most part, discussion about integration in the literature is conceptual or theoretical; as opposed to providing a specific action plan. Overall, the review revealed interpretations of integration that, for simplicity s sake, can be identified in two broad categories: horizontal integration and vertical integration. These categories are not mutually exclusive; integration could certainly be both vertical and horizontal. Horizontal integration may be used to describe frameworks that emphasise efforts at partnerships and collaboration (ranging from the sharing of information, to formation of an enduring network or partnership) among organisations or institutes. For example, a horizontally integrated initiative might involve collaboration between the World Health Organisation, the United Nations Development Program, and the World Bank; or between the Ministries of Health, Education, and Agriculture. Vertical integration, on the other hand, may be used to describe integration that is multilevel. The visual image evoked in this type of integration has the individual at the bottom, with increasingly broader levels of influence stacked above (e.g., family/peer group; school/workplace; neighbourhood/community; broader society). The integration of vertical integration refers to efforts at the various levels being directed toward a common goal (i.e., exploiting the synergy of targeting multiple influences on a health outcome). For example, an integrated effort to improve nutrition in a school may include nutrition education incorporated into the curriculum, coupled with a healthier selection of foods in the cafeteria. By integrating efforts at multiple levels, it is hoped that better and longer-lasting results will be achieved, than by taking action at only one level. It is not clear, from conceptual discussions of integration, whether integration is viewed on a continuum with greater and lesser degrees, or whether integration describes a discrete subset of initiatives. It is likely that different types of integration are appropriate for different circumstances, and that no type of integration is inherently superior in all scenarios. These are outstanding issues that need to be discussed and critiqued, if integration is to be a useful construct for improving population health. 3.2 Descriptive Models of Integration, in the Specific Context of Obesity In the above section, we described how integration is presented in the health literatures, in a general sense. In addition to these general discussions, we also encountered integrated models or frameworks that were developed for (or applied to) the specific issue of obesity. In general, these descriptive models emphasise the multiple, interacting factors, operating at various levels, that influence weight gain among populations. Because of their multilevel nature, such models may be considered examples of vertical integration. Some examples follow. 19

20 In the ecological systems framework or model (Davison & Birch, 2001), an individual s weight status is understood in terms of the contexts or ecological niches in which that person is embedded: family, peer group, school, community, and society. Key features of this approach include: accommodation of bi-directional relationships between persons and settings, interactions among levels of influence, and both direct and indirect influences. The obesogenic framework (Swinburn et al., 1999) seeks to understand the extent to which surroundings, opportunities, or conditions of life promote weight gain in individuals or populations. This framework is situated within an ecological model of weight status, and is characterised by the features outlined above by Davison and Birch. According to its authors, the obesogenic framework is most useful at the needs analysis and problem identification stage of designing an intervention. Finally, the U.S.-based International Obesity Task Force has developed a systemsoriented causal web to illustrate the linkages among several contributors to excess weight gain ( This model is also consistent with an ecological framework, in that it outlines direct and indirect contributors operating at multiple levels: international (e.g. market globalisation, economic development, media and advertising); national/regional (e.g. transport, urbanisation, education, local media); community (e.g. public transport, public safety, local agriculture, local health care); and work/school/family (e.g. recreation facilities, food availability at worksite, school, and at home). As mentioned, these frameworks are predominantly multi-level, or vertically integrated. Also, they are explanatory or descriptive, rather than prescriptive. That is, they provide a framework for understanding and accounting for the various influences on weight gain, without providing specific guidance for application. Furthermore, because of their multifactorial, multilevel, and explanatory nature, these models have the effect of throwing everything into the mix without providing any indication of what contributors are more important than others. To their credit, these descriptive models draw attention to the environment as an important influence on weight gain. This is an advantage, in light of the historical emphasis on attributes measured and conceptualised at the individual level (e.g., diet, exercise behaviour), in the majority of the existing literature. However, the ecological metaphor evoked by these frameworks has been criticised (Krieger, 1994) for placing the individual at the centre of a diverse collection of influences, and thereby perpetuating the view that influences in closest proximity to the individual (i.e., downstream factors such as diet and physical activity behaviour, conceptualised at the individual level) are the most important influences to target. In contrast, we assert that the obesity epidemic is a problem of populations, and that upstream / distal factors (which are represented as furthest from the individual in the ecological metaphor) are actually the ones with the greatest leverage for impact. 3.3 Examples of Vertically Integrated Interventions Settings-Based Approaches In addition to descriptive models and conceptual discussion on integration discussed above, our review of the literature identified a broad array of integrated interventions that took place in discrete settings the main ones being schools, worksites, and whole communities. In general, these interventions are characterised by an acknowledgement of the health promotion principle of supportive environments: settings or contexts that are conducive to health. This healthy environments emphasis is based on the assumption that, in order to be most effective, interventions must entail a multifaceted and multilevel approach. Thus, many of the settingsbased interventions detected could be characterised as vertically integrated. For a complete list of settings-based studies reviewed, please contact the authors. A summary of these studies follows. 20