Percentages, Timing, and Characteristics of Abstracts Reviewed
A total of 1004 abstracts were reviewed for the meta-narrative mapping exercise, with 94% of the abstracts emerging from the HP (n = 641) and PH (n = 307) literatures combined. That the HC and UH abstracts would constitute only 6% of the overall sample of abstracts is not surprising for several reasons. Rather than an academic line of inquiry per se, HC is a worldwide health movement designed to empower communities and cities to take action on locally defined health concerns . The movement speaks to, and receives broad-based support from, governmental and non-governmental organizations alike, and consequently focuses its dialogue in the 'grey' literature that is not captured by academic databases [24, 25, 84, 113]. Meanwhile, the total number of UH abstract hits was smaller than the other bodies of literature, as UH did not emerge as a distinct field of research until the early 2000s [42, 114]. This small pool of abstracts, coupled with the fact that they most often did not fit the inclusion criteria (28% inclusion rate), generated a very small proportion of UH abstracts to be included in the meta-narrative mapping.
Publication activity in all four bodies of literature increased over time. The HP and HC abstracts dominated the first decade of the review, and the PH and UH literatures became more prolific over the second decade, mirroring the timelines of key developments in each of these fields of research. The publications of the Ottawa Charter and Epp Report in 1986 coincided with the emergence of HP [15, 90] and the birth of the HC movement in Canada and internationally [17, 115], while PH gained considerable momentum in the mid- to late-1990s [34, 116, 117], and UH emerged in the early 2000s [114, 118].
Abstracts of Canadian origin were especially high among the PH literature, likely reflecting the strong influence of Canadian scholars in the development of this discourse [34, 41, 119, 120]. Abstracts of European origin were most common among the HC literature, reflecting the fact that while the HC movement originated in Canada , Europe, facilitated by its support from the WHO regional office , has been at the forefront of HC policy interventions [21, 122, 123]. The greatest concentration of abstracts originating from developing countries were from the UH literature, as much of the current UH research focuses on detrimental health impacts of rapid urbanization [124–126]. Similar findings on the geographic origins of the health inequities knowledge base, especially of articles emerging from the HP and PH bodies of literature, have been observed elsewhere .
The epistemological traditions of the four bodies of literature likely account for the trends in study types and target populations that were observed. With their strong epidemiological roots [127, 128], population-based surveys are commonly employed in the PH and UH literatures, and accounted for the majority of study types in this review . Meanwhile, the orientation of the HC movement to community- and government-based action accounts for the preponderance of program evaluations. Finally, the relative diversity of study types and target populations among the HP literature likely reflects the age, maturity, and resulting diversity of research programs within this body of literature. It is also worth noting that the substantive scopes and methodological paradigms employed in studies from all four of these bodies would have been shaped, if not dictated, by the priorities and terms of funding agencies and requests for proposals.
Thematic Contents of Literature and Changes over Time
Four article themes were particularly prominent in the abstracts reviewed. 'Research-related' themes, constituting 13% of article themes, captured issues ranging from conceptual or theoretical concerns (e.g., debates between PH and HP), appropriate use of indicators, instruments, and methods (e.g., how best to measure income inequality), and assessments of knowledge gaps and translation (e.g., lack of program evaluations). The highest proportion of research-themed articles occurred in the first quarter, with a steady decline in the remaining 15 years of the review. That research themes were the most prominent, especially early on in the review timeframe, suggests early efforts to establish a coherent body of knowledge on health inequities, and ongoing challenges in this knowledge base to developing evidence-based policy.
The other three themes that occurred in roughly equal measure (≈8% each) were 'healthy lifestyles' (i.e., consumption of alcohol and tobacco, nutrition and physical activity, preventive screening, and vaccines), 'healthcare' (i.e., access and utilization, costs and expenditures, systems, delivery, primary care, and health human resources), and 'social policy' (i.e., social, public, health, urban planning or policy). The prominence of the 'healthy lifestyles' and 'healthcare' themes illustrate the ongoing tendencies - criticized decades earlier  - for researchers to fixate on issues and interventions of a 'behavioural' and 'biomedical' nature. The prominence of the 'social policy' article theme might have suggested that a broader academic dialogue on health inequities was taking place. However, timeframe analysis revealed that 'social policy' coverage waned over the 20 year period timeframe, while coverage increased and remained consistently high over the 20 years for 'healthcare' and 'healthy lifestyles', respectively.
Similar findings were observed for the SDOH profile of the literature. The three most commonly profiled determinants - personal health practices & coping skills, healthcare services, and social support networks - reinforce the individualistic perspectives on population health inequities that emerged in the article theme analysis. While broader determinants, such as 'social environments', 'income & social status', and 'physical environments', were profiled, they constituted only 10% to 15% of all SDOH coverage over the entire 20 year time period. In contrast, coverage of 'personal health practices & coping skills' was at or above 20% over the entire timeframe, and 'healthcare' coverage increased considerably over time (from 5% in the first quarter to nearly 15% in the last). Thus, while some health inequities scholars made consistent attempts to steer the discourse towards broader health determinants and related implications that may be politically unpalatable, it appears there was a greater propensity to fixate on health determinants with implications for more downstream interventions that are often more amenable to implementation.
Roles for Municipal Governments
Less than one-fifth (17%) of the abstracts implicated municipal governments in any way. The apparent inattention of the majority of health inequities researchers to municipal governments may be explained by a few reasons: they may simply not hold interests in this particular realm; they may struggle to access funding for research on municipalities; or they may recognize the limitations of municipal governments' capacities to address health inequities and consequently refrain from invoking municipalities' participation and/or target their recommendations to higher authorities. We are unable to discern from our study findings the extent to which any of these, or other factors, contribute to this observation.
Seven categories were established for potential municipal roles, responsibilities and activities to reduce population health inequities (Table 2). In the Canadian context, categories 1 and 2 deal with assessing health and social needs and delivering health-based services - assessments and service delivery that might typically fall outside the range and jurisdiction of municipal services [95, 130]. Categories 3 through 6 deal with relationships between the municipality and other governments, non-governmental organizations, and within the municipality itself , while category 7 captures the types of responsibilities over which Canadian municipalities have clear existing jurisdiction, such as zoning, by-law enforcement, public libraries, and fire protection .
While abstracts of Canadian origin implicated municipalities the most (n = 48), the proportion of these relative to all Canadian abstracts reviewed was relatively small (11%). This finding suggests that the overall Canadian contribution to the health inequities knowledge base has been minimal in terms of prescriptions for municipal activity on health inequities. In contrast, while small in number (n = 13), the majority of abstracts of Mexican, South & Central American origins (65%) implicated roles for municipalities. The municipal level focus in this region of the world is likely attributable to a few factors: 1) the 'local' nature of many health problems, whereby cities in Mexico, South & Central America are simultaneously lacking basic municipal infrastructure and services to facilitate sanitary living conditions [131–133], and face common Western-world health problems associated with rapid urbanization (e.g., pollution-induced asthma) and widespread adoption of sedentary lifestyles (e.g., obesity) [132, 134]; 2) the influence of the Pan American Health Organization (PAHO) which has played a key role in addressing population health inequities in Latin America, including conducting research on strategies for engaging municipalities in health promotion initiatives  and providing strategic direction for developing interventions to address 'neglected populations' ; and 3) higher investments in participatory community-based approaches to tackling local health and social issues, as well as a strong tradition of engagement with the Healthy Cities movement in these countries [137–141].
Considering both the total number (n = 41) and the proportion (33%) of abstracts implicating municipalities, it would appear that the European literature has made the most substantial contribution to the academic dialogue on prescriptions for municipal governments to address local health inequities. The emphasis placed on municipal governments by the European abstracts mirrors the importance placed on healthy urban planning and the prominence of the Healthy Cities movement in the European context [85, 123]. With comparable (if not superior) municipal infrastructures and population health profiles, prescriptions arising from European literature bear some relevance and utility to the North American context. Indeed, it is worth noting that 'joining or building on existing local health networks and partnerships' was the most commonly cited role in both the European and Canadian literatures, suggesting that similar challenges and contexts for municipal intervention exist in these distinct geographical regions.
The most important limitation of our study is in attempting to make generalizations about the applicability of potential municipal government interventions across diverse governmental forms and functions, and geographical jurisdictions. As discussed in the introduction, the scope of powers and responsibilities of municipal governments vary tremendously both across and within nations. The generalizability of the study findings was enhanced through the use of more generic terms to code the abstracts, and by synthesizing the full scope of the scholarly 'prescriptions' into seven broadly defined and internationally relevant categories; by employing this thematically broad codebook for extracting data from the abstracts, researchers and policy-makers are permitted greater latitude to conceptualize municipal interventions relevant to their own jurisdictions. Despite a rigorous methodological design, the nature of any meta-analysis requires readers need to be critical in applying study findings to the unique contexts in which they work.
While the literature search was international in scope, the priority placed on abstracts of Canadian origin and the exclusion of non-English language abstracts mars our findings with a 'Western hemisphere' or 'developed country' bias. The English language is predominant in primarily wealthy nations whose researchers have disproportionate access to research funding and success with publication; have well established municipal governance systems and sophisticated municipal infrastructure; and have high functioning acute-care medical systems, and public health sectors that deal increasingly with reducing chronic, rather than infectious, diseases. Meanwhile, there is tremendous international variation in the scope of, and patterns in, population health inequities, and no internationally agreed upon definition for 'population health inequities'. These characteristics have important implications for the nature of municipal governments' involvements in addressing population health inequities, and thus, likely influenced the scope of prescribed roles that emerged from the literature reviewed for this study. Had our language capacities facilitated it, this limitation could have been partly addressed by reviewing articles of non-English origin. However, we submit that this 'Western hemisphere' bias is not isolated to our study, but rather pervades academe in general, and is especially reflected in developing country researchers' inequitable access to research funding and publication acceptation in international journals. It is possible that other prescriptions for municipal-level involvement (likely focusing on developing basic infrastructure and provision of relief aid in partnership with non-governmental organizations) may have emerged if more abstracts been reviewed from researchers in developing countries. Given the implicit interests in this study in understanding potential roles for municipalities with established and operational governance structures, we feel that the breadth of data retrieved from the abstracts that were reviewed remains applicable and relevant to jurisdictions that may have been under-represented in our analysis.
Another limitation of this study was in restricting our analysis to the four bodies of literature chosen. As discussed, our decision not to include the policy sciences and social epidemiology, for instance, may have led our findings to under-represent dimensions of the health inequities knowledge base that focus on broader social welfare policies or more technically-oriented epidemiological studies documenting the scope of health inequities at the local level. As we were interested in uncovering scholarly prescriptions for municipal government interventions on health inequities, we feel that the breadth of the search strategy that was employed (in terms of scope of electronic databases and search terms) captured the abstracts of greatest relevance to the questions posed in this study.
A related limitation was in treating these four bodies of literature as discrete and mutually exclusive entities. These bodies of literature co-developed over the past two decades, and as with most academic disciplines with diverse perspectives, they rely on the same baseline information. Indeed, 51 of the abstracts reviewed appeared in more than one body of literature and accounted for a total of 103 abstract cases in the dataset. An analysis of the differences between the contents of the repeat abstracts and the total sample was performed (results not shown), to document any systematic differences in articles that permeate multiple literatures. While there were no significant differences in geographic origin or in the relative emphasis on categories for municipal roles, compared to the entire sample, these 51 abstracts were significantly more likely to focus on the 'social environment' determinant of health (p = 0.002) and to implicate municipal governments in the task of addressing health inequities (p < 0.001). Thus, while this approach was employed to ensure methodological transparency and the accurate depiction of the relative contribution of each body of literature, this sub-analysis reveals that the actual quantity of abstracts emphasizing broader health determinants and a role for municipal governments was slightly overestimated and that, in fact, health inequities scholars have been even less vocal on these issues than what our larger analysis suggests.
Having a second reviewer would have been beneficial for confirming the validity and reliability of the codebook, but this was not possible due to inadequate study funding. Similarly, not reviewing entire articles may have presented an analytical weakness in this study, as article abstracts typically provide only cursory information; the information requirements for abstracts vary considerably across journals; and relevant articles without abstracts would have been excluded. Reviewing entire articles would have revealed a more accurate picture of the nuances of the health inequities knowledge base, but the sample size would have necessarily been smaller to facilitate such an intense review. Because of the importance of abstracts in offering readers a "preview of what's to come" while emphasizing some issues over others, the more cursory approach of reviewing abstracts was the best way to track meta-narratives from this large and diverse body of knowledge over a twenty year timeframe.
Policy Implications: Prescriptions for Municipal Government Intervention
Overall, the health inequities knowledge base offered insufficient guidance to municipal governments in developing healthy public policy at the local level. Health was conceptualized in primarily 'behavioural' and 'biomedical' terms, providing little incentive for municipalities to consider, and act on, the full range of the SDOHs. If researchers, who have at their disposal voluminous evidence on the social determinants of health inequities, overwhelmingly defer to healthy lifestyles and healthcare services as the levers for improving health, then how can busy, and often uninformed, policy-makers be expected to conceptualize health any differently? The minimal attention paid to municipal governments in the health inequities knowledge base urges critical reflection on the subject areas and types of health research that funding agencies privilege, and highlights the need for increased funding and translation of interdisciplinary health inequities research that is relevant to policy-makers, especially at the municipal level where human resources devoted to exchange with research communities are in short supply.
The relative silence of the health inequities knowledge base on avenues for municipal action presents another challenge to developing healthy public policy at the municipal level. With less than one-fifth of the abstracts implicating municipalities in any way, and the tendency for those implications to originate from Europe, it is clear that health inequities researchers offer inadequate prescriptions for municipal policy-makers from other jurisdictions to draw from . Even if prescriptions were readily available, municipal policy-makers would justifiably have little faith in the effectiveness of such prescriptions, given the dearth of evaluations of programs targeting health inequities . At a minimum, though, paradigm shifts are needed in both the academic and policy domains to move the issue of population health inequities onto the municipal government agenda.
Future Research Directions
The findings from this study illuminate a number of potential avenues for future research. Given its explicit 'city' focus, it was not surprising that the Healthy Cities literature implicated municipal governments in the greatest proportion. One might have expected, however, to see more implications from the Canadian Healthy Communities literature, considering the role that Canadian scholars played in launching the Healthy Cities movement. What characteristics are present in Europe, that are not present in North America and other jurisdictions, that would explain the apparent uptake of Healthy Cities agendas by municipal policy-makers across the region (as evidenced by the various past and current Healthy Cities projects across Europe), as well as the high degree of support from the WHO regional office for Healthy Cities programs? What powers might European cities possess that facilitate the implementation of prescriptions from the Healthy Cities movement? These cross-jurisdictional differences signal the need for researchers to investigate the nature of urban health governance across diverse political systems, which could offer an explanation for the lack of action on health inequities at the local level in Canada and elsewhere.
While our findings suggest evidence of uptake of the Healthy Cities component of the health inequities knowledge base by municipal policy-makers, especially within European and Latin American jurisdictions, little is known of the extent of uptake of the other components of this knowledge base by municipal policy-makers. As the objective of this study was to survey the health inequities knowledge base for prescriptions for municipal government intervention, another logical direction for future research would be to assess the extent of awareness and utilization of this knowledge by municipal policy-makers in jurisdictions around the world, as well as the perceptions held by municipal policy-makers themselves of the roles and responsibilities of municipalities in addressing health inequities. In another component of this study, we have investigated these questions within the metropolitan region of Metro Vancouver, Canada, which consists of nearly twenty autonomous municipal governments . Similar analyses are needed in other jurisdictions.
In addition to investigating the status of translation of the health inequities knowledge base by municipal policy-makers, regular efforts should also be made to review the knowledge base for emergent prescriptions for both governmental and non-governmental interventions. In August 2008, the WHO Commission released another report on the state of the SDOH and strategies for reducing population health inequities . The dynamic and evolving nature of this knowledge base suggests that the relevance of, and academic support for, strategies and interventions to reduce population health inequities can be short-lived.