Assessing the social and physical determinants of circumpolar population health

Introduction Systematic reviews of the social and physical determinants of health provide metrics for evaluation of programs to mitigate health disparities. Previous meta-analyses of the population health literature have identified several proximate social and physical determinants of population health in the circumpolar north including addiction, environmental exposures, diet/nutrition and global climate change. Proximate health determinants are most amenable to early detection and modification or mitigation through disease prevention or health promotion interventions. Design There is a need for research to replicate these findings based on the latest science. This presentation describes a study applying Dahlgren and Whitehead's (1991) socio-ecological model of health determinants to identify the proximate social and physical determinants of health in the circumpolar north. Methods The study consisted of a systematic review of recent studies that link determinants of health with the leading causes of mortality and morbidity in Alaska. Our search strategy employed a keyword search using the Circumpolar Health Bibliographic Database (CHBD) and 4 databases within the Web of Knowledge (WoK) data gateway. Keywords included various terms for the arctic, all relevant nations and territories within the region, as well as leading health outcomes. Results Studies meeting the following inclusion criteria were reviewed: original research within a circumpolar population, published in English during 2011, and involving a rigorous demonstration of a link between a social determinant and selected health outcomes. Conclusions Study conclusions includes a list of determinants identified, their associated outcomes and the study designs implemented to assess that association.

S ocial and physical determinants of health are the social factors and physical conditions that shape whether individuals stay healthy or become ill (1). Furthermore, the social and physical determinants of health are a useful lens through which to view the health disparities that could lead to preventable health outcomes (2). In the period since the social and physical determinants of health gained recognition as a foci for health promotion programs, many important organizations have targeted specific determinants to set priorities for research and policy in both international (3) and national (4) initiatives. Some have argued, however, that these programs fail to include a circumpolar health perspective (5). Many circumpolar communities face health challenges that differ from those encountered by more southerly communities. Recent scholarship describes obstacles to accessing quality health care in geographically and socially isolated circumpolar settings (6), experiences with extraordinarily rapid economic development (7), and growing health disparities in culturally distinct and indigenous populations (8). In the field of environmental health, the Arctic has become a sentinel for issues such as the disruption of traditional ways of life due to climate change impacts (9). There is a need for further investigation into the determinants of health outcomes in the circumpolar north.
Faculty and staff at the Institute for Circumpolar Health Studies (ICHS) conducted a review to assess what determinants of circumpolar health were described in the published literature in 2011. This literature review was part of an ongoing process to inform the research portfolio at the ICHS, and contribute to the global health discourse on determinants by identifying contributions and gaps in the literature that may indicate advancements or lags in the conduct of research related to prospective determinants of circumpolar health. This review had 2 over-arching objectives: a. Identify studies that demonstrate associations, either positive or negative, between determinants of population health and circumpolar health outcomes, and b. Characterize key features of the research methodologies currently employed by researchers focusing on the circumpolar north.

Methods
Previous studies have described the limitations of electronic data archiving in searching for, and synthesizing, relevant publications during systematic reviews (10,11). We elected to employ the Web of Knowledge (WoK) because it is a journal index that includes a large body of health literature, and also contains a means of limiting the geographic search in a way that does not include an author's geographically based affiliation while still including geographic terms relevant to the reported study's location (12). Review criteria included original research taking place in the circumpolar north that includes a rigorous association between a social determinant and selected disease outcome. Publications were either in English or were English translations. We did not include review articles or those studies including populations outside the circumpolar north. We defined the circumpolar north as the geographic region north of 608 north latitude. Search terms for geographic keywords included English translations and English spellings of local names whenever possible. In the event of multiple translations, both versions were entered. Terms for search strings are shown in Table I.
Because this study was intended to inform the research portfolio at the ICHS, the priority health outcomes were the leading causes of mortality and morbidity in Alaska (13). We limited our health outcomes to the 7 leading causes of mortality in the state (Fig. 1), as these 7 categories account for more than 90% of all mortalities in Alaska and Canada. We also included 3 infectious diseases; influenza/pneumonia, sexually transmitted infections (STIs) including HIV, and tuberculosis, as these represent important categories of infectious disease morbidity in Alaska. We consulted the International Classification of Diseases (ICD)-10 (14) and extracted each unique disease term for each relevant set of codes for these 10 priority health outcomes.
The primary search involved entering the geographic search string in the WoK basic search box, and setting search criteria described above, including not limiting the search to the author's geographic affiliation while still including geographic terms relevant to the reported study's location. Then secondary searches were completed by individually entering a disease search string into the ''search within results'' textbox. This process was repeated 10 times, so that each disease category was individually searched within the complete set of geographical results (Table I). For each search, a single reviewer examined all titles that appeared in the search results. Any title that was not clearly excludable was downloaded to an Endnote database. Each abstract was read completely for exclusion. Abstracts that were not clearly excludable were retained. Complete articles were downloaded in the event of ambiguity over inclusion status. All articles were divided among a review panel composed of the 3 authors and distributed in a way that each article received review by 2 panellists. The panellists reviewed each article to assure inclusion criteria were met, and to assess associations, including study design, between any social determinants and the priority health outcomes. The panel then met and discussed discrepancies. All disputes were settled unanimously through discussion.

Results
The initial geographic search had 24,413 hits returned. Secondary searches included only 2,430 titles. Of those, only 573 abstracts remained for downloading. In the end, 30 (Table III) articles fit inclusion criteria. Figure 2 details the article identification process.
The 2011 review identified studies that explicitly assessed associations between 8 determinants and one or more of our 10 priority health outcomes: addiction, social connectedness, environmental exposures, diet/ nutrition/exercise, access to quality health care, access to clean water, sexual and reproductive health, and occupational health and safety. Six of these determinants were identified as the topics of health studies in a similar review of the literature in 2009, along with one other: climate change. The 2 additional determinants identified in 2011 from 2009 were sexual and reproductive health and occupational health and safety. One overlapping category, social isolation, was reclassified as social connectedness. While social isolation results in stress and contributes to disease, social connectedness provides a buffering effect leading to resiliency. Both ends of the continuum are included in the category of social connectedness (or lack thereof). Table II details   Several articles had overlapping determinants or leading cause of mortality or morbidity associations. The results included only one random controlled trial design. The single cohort study was associated with addiction. Six of the social determinant categories had 4 or more associated articles that included both cross-sectional and descriptive designs. The category of access to clean water and occupational health and safety each had 1 associated article. The research design for the articles included a cross-sectional and a descriptive, respectively. See Table III for other details.
Comparison with 2009 review results showed 38 points of connection between 11 disease categories and 9 social determinant categories (Fig. 3). The most common social determinant of the health outcomes assessed in this study is addiction. Addiction is associated with all 11 disease outcomes. Of the 38 points of association 29% are

Discussion
Among the most important questions raised in this series of reviews is the meaning of the observed gaps in publications. On the one hand, the gaps may indicate a lag in the research-publication process. On the other hand, they may indicate areas needing research; areas where additional social determinants may be discovered. It may be that gaps are representative of both issues. For example, social connectedness is a determinant with clear links to heart disease (45).
The above questions expose the limitations of this review. While it is our intention to sample the literature on an ongoing basis, it may be that the bi-annual review    process, which we are undertaking primarily due to limitations in personnel time, obfuscates the true range of topics researched. Additionally, the evolving methodology makes precise one-to-one comparisons between annual reviews impossible. This process monitors the pulse of research in the circumpolar north, but by no means provides a comprehensive treatment of the subject. The greatest use for this review is its utility as a planning tool. By following the current trends in research and publishing, those interested in social determinants of health can better allocate effort and resources. For example, ICHS, in partnership with the Center for Alaska Native Health Research, is planning a project that examines the relation between cultural values and cancer as a social and ecological stressor. Such a project is clearly tied to issues of social connectedness and has the potential to affect cancer incidence and outcomes. These data hold promise for a number of strategic planning processes, none of which is mutually exclusive. Without doubt, as the review process continues to inform our understanding of health and research, our strategies for review and application of findings will continue to change and mature.
Future plans include further incorporation of the 2009 and 2011 results into our strategic plan and research portfolio, as well as replication of the process for presenta-tion at the 2015 International Congress on Circumpolar Health.