Voting, health and interventions in healthcare settings: a scoping review

Background In democracies, voting is an important action through which citizens engage in the political process. Although elections are only one aspect of political engagement, voting sends a signal of support or dissent for policies that ultimately shape the social determinants of health. Social determinants subsequently influence who votes and who does not. Our objective is to examine the existing research on voting and health and on interventions to increase voter participation through healthcare organizations. Methods We conducted a scoping review to examine the existing research on voting, health, and interventions to increase voter participation through healthcare organizations. We carried out a search of the indexed, peer-reviewed literature using Ovid MEDLINE (1946–present), PsychINFO (1806–present), Ebsco CINAHL, Embase (1947–present), Web of Science, ProQuest Sociological Abstracts, and Worldwide Political Science Abstracts. We limited our search to articles published in English. Titles and abstracts were reviewed, followed by a full-text review of eligible articles and data extraction. Articles were required to focus on the connection between voting and health, or report on interventions that occurred within healthcare organizations that aimed to improve voter engagement. Results Our search identified 2041 citations, of which 40 articles met our inclusion criteria. Selected articles dated from 1991–2018 and were conducted primarily in Europe, the USA, and Canada. We identified four interrelated areas explored in the literature: (1) there is a consistency in the association between voting and health; (2) differences in voter participation are associated with health conditions; (3) gaps in voter participation may be associated with electoral outcomes; and (4) interventions in healthcare organizations can increase voter participation. Conclusion Voting and health are associated, namely people with worse health tend to be less likely to engage in voting. Differences in voter participation due to social, economic, and health inequities have been shown to have large effects on electoral outcomes. Research gaps were identified in the following areas: long-term effects of voting on health, the effects of other forms of democratic engagement on health, and the broader impact that health providers and organizations can have on voting through interventions in their communities.


Background
The idea that health is strongly determined by social factors and processes-what we now call the social determinants of health-has long been a central idea within public health [1]. All social determinants of health are shaped by the distribution of power and resources within societies and at a global level [2,3]. A number of processes influence this distribution of power and resources, including constitutions that define the rights and responsibilities of citizens and governments, policies that determine the minimum wage, work conditions, and social assistance, as well as the budgetary decisions that direct resources toward (or away from) education, child development, housing, and social services.
In democracies, citizens can play a variety of roles in the processes that shape the social determinants. Voting is one key aspect of democratic engagement, defined as "a multi-faceted phenomenon that embraces citizens' involvement with electoral politics, their participation in 'conventional' extra-parliamentary political activity, their satisfaction with democracy and trust in state institutions, and their rejection of the use of violence for political ends" [4]. More simply, democratic engagement is "the state of being engaged in advancing democracy through political institutions, organizations, and activities" [5]. Democratic engagement can include electoral participation (voting, campaign displays, volunteer, campaign contributions), expressing a political voice (protest, boycott, contacting officials), having political knowledge/awareness (following government affairs, watching/reading/listening to news, talking about politics), and holding certain attitudes (promoting common good, affirming common humanity) [5].
The effect of voting on the social determinants of health is multi-factorial and complex. In a simple conceptualization, when larger numbers of people from certain communities and groups participate in voting, it translates into greater influence over determining who holds political power. Those in power in turn put forward and support policies that respond to the needs and demands of their constituents that shape the social determinants of their health. Not only does voting partially decide who forms government in democracies, and subsequently what policies shape social determinants, but the relationship may work in the opposite direction as well, in that the social determinants of health affect voting patterns. For example, socioeconomic status is associated with the likelihood of voting. Across many contexts, having a low income and a lower level of education is associated with lower rates of voting during elections [6][7][8]. Numerous theories explain this association including decreased social trust, diminished social capital, fewer chances to vote, and weakened educational opportunities about the policy process [7,8].
Public health scholars have been called upon to better understand the functioning of politics at national and sub-national levels-and the mechanisms that connect politics to public health [9]. Our objective in this scoping review was to examine the existing research on voting and health, and on interventions to increase voter participation through healthcare organizations. We sought to understand the following questions: What is the relationship between voting and individual health? What healthcare-based interventions exist to support voting, and what have been their outcomes?

Methods
We conducted a scoping review with the central objective of identifying the existing peer-reviewed research on the association between voting and health, and on interventions that aim to increase voter participation through healthcare organizations [10,11]. Voting was used as a proxy for democratic engagement in this scoping review as it is easily identifiable, measurable, and is an essential and defining characteristic of healthy democracies. We chose to focus on healthcare-based interventions, to explore the role that the health sector-which has frequent contact with large numbers of individuals from communities with relatively lower rates of voting-can play in supporting voting. We searched Ovid Medline (1946-present), PsychINFO (1806-present), Ebsco CINAHL, Embase (1947-present), Web of Science, Proquest Sociological Abstracts, and Proquest Worldwide Political Science Abstracts in March 2018. We used a broad search expression (Additional file 1, Additional file 2) in order to include as many articles as possible. Our search timeframes were chosen to include the full scope of articles available on each research platform. We limited our search to the peer-reviewed, indexed literature in English. The titles and abstracts of citations identified were reviewed independently against our inclusion and exclusion criteria by two authors (CB, DR), followed by review of the full-text articles. In this scoping review, we did not perform backward reference tracking.
We included peer-reviewed articles where the main focus was the relationship or association between voting and individual health, or focused on interventions in healthcare settings aimed at increasing voter participation. We excluded articles solely focused on the links between health and other forms of democratic engagement (ex. activism, protest) to focus more narrowly on the link between the act of voting and health.
After our initial review of full-text articles to ensure they met our inclusion and exclusion criteria, two authors (CB, DR) completed data extraction. We extracted information on the geographic location and context, area of focus, the effect size, measures used, and confounding variables. We prepared summaries for each article on the key themes and findings in one shared document, and then the entire study team reviewed these summaries and identified common overarching themes relevant to our review objectives.

Results
Our initial search identified 2041 citations (Fig. 1), and after reviewing titles and abstracts, 49 articles met our inclusion criteria. Following full-text review, 40 articles were included in the final analysis ( Table 1). As we put in place a broad search strategy, many articles were not relevant to the research questions. Most of those were articles that focused on subjects like healthcare policy, democratic engagement (activism, civic engagement), voting patterns, political engagement, and health equity more broadly without actually discussing the link between voting and health or describing healthcare interventions in the voting process. The included articles were published in a diversity of research disciplines (classified according to journal and study design): health science (geriatrics, pediatrics, psychiatry), public health and epidemiology, political science, and social science. Most of the research was done in high-income countries, with a focus on Europe, the USA, and Canada. Although most of the research has been more recent, with 27 articles being written from 2010-present, the articles included were published between 1991 and 2018. Study designs included cross-sectional studies, cohort studies, case studies, qualitative studies, literature reviews, and critical commentaries.
Four common themes emerged: (1) there are consistent patterns in the association between voting and health; (2) differences in voter participation are associated with health conditions; (3) gaps in voter participation may be related to electoral outcomes; and (4) healthcare interventions exist to increase voting and democratic engagement. We chose these four overarching themes after reviewing summarized notes of the key findings and details of each included article (see Methods). Although there is partial overlap, we believe that articles included under each theme deliver four distinct messages that inform our main research questions in unique ways.

Consistent patterns in the association between voting and health
Seventeen studies examined the association between voting and health in numerous jurisdictions and levels of government (municipal, state or province, and federal elections), and in numerous locations across North America and Europe. Lower voting rates are consistently associated with poor self-rated health. In most studies, health was measured by surveys that included questions about self-reported health [6, 13, 16-19, 22-24, 26]. Other measures included health risk behaviors [12,14], mortality [21,27], chronic health conditions [14], health indices [14,15,18,20], and hospitalization data [25]. This health data was then linked to data on voting, measured in various ways including self-reported voting registration and national statistics. Blakely, Kennedy, and Kawachi analyzed the data of 280,000 respondents of an American Current Population           Survey and found that voting is positively associated with self-rated health, independent of income inequality [6]. Similar patterns were found by Burden et al. in an older Wisconsin population [15]. Globally, similar correlations between voting and health have been found in Ireland [18], Russia [27], Sweden [20], Canada [16,19], Europe [26], and the OECD more broadly [24]. Both Couture and Breux [16] and Habibov and Weaver [19] looked at large sample sizes from Canada's General Social Survey and found a correlation between self-rated health and voting. Habibov and Weaver connected this association between voting and health to the importance of social capital [19], as did many other articles in our review [6, 21-23, 26, 29, 30, 38]. Most studies were cross-sectional, with only a few longitudinal studies finding an association between voting and health and socioeconomic benefits. Adjusting for confounders like sex, education, geography, and chronic illness, Arath showed that voting abstention was associated with 1.3 times higher odds of reporting poor health two years later [13]. Ballard, Hoyt, and Pachucki looked at longitudinal data that followed adolescents into adulthood and found that voting was positively associated with better mental health and health behaviors over time, along with improved income and education level [14].

Differences in voter participation are associated with health conditions
Although the connection between voting and health was researched in the above articles, the next overarching theme further analyzes this connection by discussing voting patterns in distinct sub-populations. People with physical, intellectual, and psychological disabilities have lower rates of voting. Agran, MacLean, and Kitchen found lower voting rates in communities of people with intellectual disabilities [28]. Matsubayashi and Ueda [33], Mattila and Papageorgiou [34], and Shields, Schriner, and Schriner [37] discovered low voter turnout rates among people with disabilities, with barriers to  [35], and Ojeda found that depression reduced voting participation [36]. In a qualitative study, Bergstresser, Brown, and Colesante interviewed 52 consumers of mental health services who described political participation as contributing to their recovery by increasing social inclusion [30]. There are differences in voter participation by race, gender, age, and disease type. Ard et al. found a positive association between engagement in politics and self-rated health in connection to racial health disparities in the USA [29]. Disparities in health and voting in African American communities were found in two studies by Bazargan, Kang, and Bazargan [7] and Bazargan, Barbre, and Torres-Gil [8], which saw a voting gap in elderly black communities in the USA. Being elderly can lead to certain vulnerabilities, such as social isolation and physical impairment, which can then lead to lower voting rates [26,43]. Higher political participation (which includes voting participation and registration to vote) in American women is also strongly correlated with lower mortality [32]. Interestingly, the type of disease an individual has can affect their voting behavior. Acute illnesses like influenza can affect voter turnout [39]. Focusing primarily on chronic diseases, Gollust and Rahn found that those with heart disease and disability were less likely to vote in the 2008 US election, whereas those with cancer were more likely to vote [31]. One hypothesis was that strong social support networks in the cancer community, and less stigma compared to other diseases, led to higher voting rates among people with cancer. Sund et al. saw similar results: those with cancer and COPD often voted more, whereas those with neurodegenerative brain disease, addiction, and mental health disorders voted less [38].

Gaps in voter participation may be related to electoral outcomes
Although only three articles were included under this theme, we nonetheless created a distinct category due to the unique and important findings of these articles, namely, differences in health status and subsequent differences in voting patterns can impact electoral outcomes. In two population health studies, Rodriguez [40] and Rodriguez et al. [41] analyzed the association between poor health and voting and the broader impact these inequities can have on our political systems. They hypothesize that "through the early disappearance (i.e., death) of the poor, continuing socio-political participation of high-SES survivors helps to perpetuate inequality in the status quo" [40]. The citizens most expected to vote in line with redistributive health policies are the same citizens that have higher mortality rates during the time when they are most likely to vote-middle age. Previous to this study, Rodriguez et al. looked at how racial inequality in the USA leads to excess mortality and therefore a loss of votes. In introducing the subject of racism and voting, Rodriguez et al. point out current US voter suppression practices aimed at marginalizing minority populations, from felony disenfranchisement laws, to redrawing of electoral boundaries, to shortened polling hours. This article focuses on the effects of health inequity as another threat to minority voting power. They found that from 1970-2004, there were 2.7 million excess black deaths due to racial inequality, which led to 1 million lost black votes in the 2004 election [41].
This study concluded that many close state-level elections in the US over this period of time would likely have had different electoral outcomes if not for these excess mortality rates.
Using a multivariate analysis and controlling for sociodemographic characteristics, Ziegenfuss, Davern, and Blewett [42] found that individuals with healthcare access problems were significantly more likely to vote for Democratic candidates in the 2004 election. They connected this to the Democratic Party comprehensive approach to healthcare reform in the 2004 election. If inequities in access to healthcare services and in health outcomes can change who wins elections, a vicious cycle can emerge: worse health leads to lower voting rates, leading to policy that does not prioritize addressing inequities, leading to worsening health inequities.

Healthcare interventions exist to increase voting and democratic engagement
Healthcare interventions aimed at increasing voting rates have emerged within nursing, social work, and medicine. Regan, Hudson, and McRory conducted a literature review that looked at the role of nurses in ensuring patients' right to vote, issuing a call to action for nurses to help ensure this right through policy guidelines and increased support for patients [46]. Anderson and Dabelko-Schoeny argued that civic engagement can lead to better health in nursing home residents and called for social workers to develop and implement interventions that increase engagement [43]. White and Wyrko wrote that healthcare professionals should make every effort to ensure hospital patients can vote in the UK [48]. They suggest an approach focused on increased awareness and discussion among healthcare practitioners, promotion of voting access, and the consideration of emergency proxy voting.
Within the healthcare setting, Wass et al. found that proxy voting as a voter facilitator instrument can increase voter turnout for those suffering from ill health or disability [47]. Hassell and Settle ran an interventional study that induced life stressors on patients and found that increasing stress decreased likelihood to vote for typical nonvoters [44]. Liggett et al. conducted an evaluation of clinician-led, nonpartisan voter registration drives over 12 weeks within two university-affiliated health centers in the Bronx, New York [45]. The project was successful in registering 89% of eligible voters, demonstrating the importance of health centers as, "powerful vehicles for bringing a voice to civically disenfranchised communities".

Discussion
Our review found an association between voting and health. Poor health is often associated with lower rates of voting. This was consistent across diverse health outcomes, jurisdictions and governments. A few studies provided weak evidence that voting may lead to better health and well-being [13,14], although there have not been enough studies in this area to strongly confirm this association. Individuals living with disability, mental and physical illness, minorities, and older individuals, tend to vote at lower rates in general. Votes lost to morbidity and mortality in marginalized populations may potentially impact electoral and policy outcomes, including public health policy. Among some of the included studies, the causal relationship between voting and health was seen as bidirectional: voting affects health as it shapes who is in power and what policy is made, and individual health can affect voting. Taken together, a cycle can develop of poor health and political disempowerment, although further research is required to fully characterize this process. Despite the importance of this relationship, the association between voting and health has not received significant attention in the public health literature to date [49]. This review provides some conceptual clarity to this developing research area.
Many articles included calls to action for healthcare practitioners to engage in and advocate for democratic engagement in their patient communities through policy change, accessibility, support, and even intervention to help increase voter participation. Healthcare organizations are well suited to engage directly with marginalized populations and can be involved in improving democratic engagement through education and interventions similar to Liggett et al., who undertook a clinician-led voter registration [45]. Other possible interventions could include reducing barriers to voting (proxy voting at hospitals), organizing nonpartisan townhalls, or compiling and sharing information for communities on the voting process [50,51].
Many authors proposed theories to explain why poor health and lower voting turnout were associated. These included that people with poor health had lower cognitive resources, worse sense of efficacy, unmet accessibility needs (especially for those with disabilities), and limitations in time, social/emotional, and financial resources due to health burden [25,29,31]. Several authors cited social capital and social connectedness as part of the causal link between voting and health. Voting could be seen as a form of social capital as it entails social trust and civic engagement, but even more than that, having social networks who vote and talk about voting can reinforce voting patterns within a community. Social connectedness can improve mental and physical health, lead to less risky health behaviors, and increase access to community networks, institutions, and resources to improve health [16,[20][21][22]. Gollust and Rahn explored the role social capital played in voting and health by discussing one of the only populations where voting rates increase with a chronic health condition: people living with cancer [31]. They hypothesized that people living with cancer are much more likely to join social and advocacy cancer support groups than people with other diseases. For example, people with breast cancer form more than forty times more support groups than people with heart disease. These social and advocacy groups not only then support the act of voting, but also equip members with skills that help them better understand the political process, which then leads to higher voter participation. Overall, many authors linked voting to health through social capital. This is an important area of future research for the field of public health, as social capital is a key social determinant of health in itself [3]. This links back to an important consideration in our scoping review: how voting is connected to the social determinants of health.
Our review had limitations. Voting was chosen as a proxy for democratic engagement, but there are numerous other forms of democratic engagement: activism, protest, donations to political groups, political education, and more. Also, democracy comes in many forms, between countries and within countries at different times. We recognize this would influence how voting occurs in different contexts, the meaning it would have to citizens, and the subsequent relationship between voting and health. Voting is also deeply connected with other social determinants of health-namely income and education-which may confound some of the research presented. Most of the articles addressed confounders within their statistical analysis, including sex, age, marital status, race, education, employment, income, geography, and more. Addressing these confounders was imperative in claiming an association between voting and health, but it is important to note that the articles often measured these factors differently and used a differing combination of factors. Future work should synthesize this literature to develop a more holistic picture of the connection between other forms of democratic engagement and health. Future research should also examine the long-term effects of voting on health, as well as the impact of health organizations actively intervening in and advocating for democratic engagement in their communities.

Conclusion
This review has supported the association between voting and health. Communities marginalized by disability, mental and physical health, race, and age tend to be the most affected by the positive association between health and voting. Differences in voter participation related to health inequities can have some effect on overall electoral outcomes, shaping overall policy and possibly deepening healthcare inequities. Future research should study the long-term effects of voting on health, the effects of other forms of democratic engagement on health, and the impact healthcare practitioners can have on voting activity in their community through intervention and advocacy.
Additional file 1: MEDLINE search strategy.
Additional file 2: Articles identified by database.