1. Introduction

For some time, scholars and health care advocates have been interested in the determinants, rationale and strategies for assuring adequate national health care spending. Regardless of how price-conscious, consumer-driven or technologically sophisticated a nation may be, health care expenditures will consume significant portions of its economy, often in place of other public obligations. How health care expenditures are accrued (i.e., what services are provided) and paid for (i.e., the balance of private and public sector appropriations) can spark lively discourse. Recent US debate about the appropriate role of government in defining and funding health services reflects not only diverging economic, but also socio-political and cultural perspectives that may inhibit a nation’s capacity to forge effective health care policies [1,2]. Comparative health care studies can inform the understanding of how to address a nation’s health care needs in ways that will maximize effectiveness (i.e., achieving measurable reductions in morbidity and mortality), efficiencies (i.e., doing so with minimal commitment of human, material and financial resources) and/or equity (i.e., distributing costs and benefits of care fairly and impartially). As Anderson and Poullier note, “looking at data from many countries allows policymakers to compare where individual countries are doing well, doing poorly, or simply making different choices [3]”.

Cross-disciplinary research has documented sizable differences in expenditures within and across nations [4]. Overall, health care expenditures account for roughly 9% of the gross world product, although the proportion of nations that commit less than 4% of their Gross Domestic Product to health care approximates the proportion that spend 10% or more [5].

What accounts for such differences? Research on the associations among health care spending, supply, prices, utilization and health impact have been plentiful in the context of developed, industrialized democracies [6–9], but far less attention has been given to the interplay of such factors in the developing world. Access to economic resources and, to a considerably lesser extent, demographic factors has been recognized as predictors of a nation’s health care expenditures. Wealthy nations have a disproportionately greater economic capacity to allocate to health care [10]. Newhouse, among the first to examine these relationships, found per capita income to account for a substantial portion of variation in per capita health expenditures among 13 developed nations [11]. Other research has highlighted the development of medical technologies, expansion in the number of insured lives, increased earnings expectations among health care professionals and general inflation of health care costs as being responsible for much of the observed increases [12]. Some have considered the impact of recent “demographic transitions” on health service demands and cost escalation [13,14], while others assert population aging has had a very modest impact on health care costs [15]. Fuchs and others have emphasized how the availability of health care personnel and associated technology fuels health care expenditures [16–18]. Recent analyses found inflated personnel and service costs, readily accessible technologies and behavioral risks (i.e., obesity) as primary sources for elevated health care spending [8,9].

Negligible attention, by comparison, has focused on the nature of political institutions and their bearing on health care expenditures. Global increases in health care spending are understood to parallel economic development [3], but spending as related to the second and third waves of democratization has yet to garner attention [19].

The financing and delivery of health services, by their natures, are political. In principle, they reflect beliefs about rights to well-being of organized people, while in practice they readily reveal fundamental socio-political inequalities operating within a society [20]. Moreover, in as much as health services delivery to varying degrees is funded through public dollars, it is modifiable through public policy and its character reflects political exigencies as much as technological or logistical capacities. In essence, a nation’s health care system is a consequence of economic, social and political power exerted through democratic or authoritarian national leadership, and, as such, the political dimensions of health care merit further study [21].

Understanding how, and to what extent, a nation expends resources on health care as a reflection of political ideology can provide insight about its capacities to address this and other public functions pertaining to education, public safety, physical infrastructure, communication, national security, etc. By doing so, one may better anticipate the health and social needs of at-risk populations and gain insight regarding the particular health burdens they may face. If, for example, spending on health care by democratic governments differs in either absolute or proportional terms from that of autocratic nations, it may be relevant to consider how worldwide movements toward democratization [22] may come to affect health care spending. This manuscript considers the relationship of a nation’s regime type (i.e., norms by which governing authorities function) to health care expenditure levels. This study hypothesized that democratically governed nations spend more, both proportionately and per capita, on health care services than those governed autocratically and that the association of regime type with expenditures will persist when accounting for demographic, economic, and health system characteristics. The implication of this analysis in addressing global health objectives is briefly considered.

2. Methods and materials

Data regarding the populations, health systems and political characteristics of 159 nations, with roughly 98% of the world’s population, were examined. The World Health Organization’s (WHO) Global Health Observatory Database is a repository rich with information on health care expenditures, health services and health status indicators [5]. The WHO health expenditure database, whose procedures for securing and validating information are well described [23], provides cross-national comparisons over a 10+ year period, utilizing national health account reports, governmental and public expenditure information.

3. Results

Sizable differences among the 159 nations regarding health care expenditures, demographic, health system and socio-economic conditions are evident in Table 1. One half of the 159 nations expended less than 6% of their GDPs on health. Predictably, the US expended more of its GDP than other nations on health during 2008 (15.2%), with the smallest percentage (1.7%) noted for Equatorial Guinea and Turkmenistan. With a range of 0.7–26.1%, the median expenditure on health by governments in 2008 was 10.9% of total government expenditures. Per capita total expenditures varied from 18 Int$ for Eritrea to 7164 Int$ for US, while per capita governmental expenditures ranged from 7 Int$ for Sierra Leone to 4091 Int$ for Norway. Univariate associations between these four indicators and REGIME, reflected by the standardized regression coefficient (Beta), were positive and moderately strong; at successively more democratic classifications, expenditures on health increased. Considered in isolation, a one-unit change in REGIME type was associated with 0.40 to 0.49 standardized unit changes in health expenditures; overall variation among nations in REGIME type accounted for 24% to 16% of variation in expenditures across these 159 nations (p < 0.01).

Multivariate analyses (see Tables 2 and 3) markedly increased the proportions of explained variation in health expenditures, but did not eliminate the direct association observed between REGIME type and health care expenditures. Among 159 nations included in the analysis, health care expenditures as a proportion of a nation’s GDP was estimated to increase with increased democratization evident in a nation’s authority structure (e.g., an estimated 0.139% increment per a positive change in REGIME scores). As such, a 7-point difference in REGIME score between nations could be estimated to yield a 1% difference in the proportion of GDP expended on health care. This association was independent of those related to age composition, life expectancy, health care workforce and system effectiveness, as well as education and income variation across nations. Expenditures also were estimated to significantly increase with the percent of populations over 60 years of age (i.e., expenditures were estimated to increase 0.185% with every 1% increment in Over 60 values). Overall, the adjusted R² value revealed 38% of variation in the outcome variable that was associated with these seven predictors. Contrasting the difference in univariate and multivariate results for REGIME and TEH%GDP (Beta = 0.49 and 0.329 from Tables 1 and 2, respectively) suggested that one third of the initial association between measures could have been attributable to the association of REGIME with other predictor variables (0.161/0.49), while two thirds of the estimated association remained independent of those relationships.

Approximately one quarter of variability in the percent of general government expenditures attributable to health care was accounted for in the multivariate analysis. However, of the seven predictor variables, only REGIME yielded a significant association. The analysis estimated a 0.251% increment in the proportion of general government expenditures that were attributed to health expenditures for every unit increase in REGIME scores. For context, a difference in general government expenditures on health between REGIME scores for nations characterized as “fully institutionalized autocracies” (−10 to −5) and “fully institutionalized democracies” (+5 to +10) of 2.5% of general government expenditures could be anticipated. The standardized regression coefficient in this model for REGIME, Beta = 0.357, indicated that roughly three fourths of the association between outcome and predictor estimated by the univariate model (Beta = 0.48) was unique to REGIME and independent of other factors.

The multivariate model accounted for better than 80% of variation in total per capita spending on health. Three of the seven predictor variables were significantly associated with expenditures. A 34.3 Int$ per capita increment was estimated for every unit increment in REGIME score. The standardized coefficient (Beta = 0.171), though significant, was decreased approximately 60% from the univariate statistic (Beta = 0.41), suggesting the majority of REGIME’s initial effect was associated with its relationship to other factors included in the multivariate model. The model also estimated an increase in per capita expenditures which increased 52.4 Int$ for every 1% increment in a nation’s population over age 60, along with per capita increments of 3.36 Int$ with every additional doctor, nurse and/or midwife per 10,000 persons in the population.

According to the final equation of Table 2, REGIME, along with Over60, LifeExpect, DrRN/10K and PcGNI exhibited significant associations with per capita general government expenditures on health. Per capita expenditure increased 42.2 Int$ for every percent increase in population over age 60, 2.94 Int$ for every additional doctor, nurse or midwife per 10,000 persons, 37.6 Int$ for every additional 1000 Int$ in a nation’s gross national income and 22.4 Int$ for every increment in REGIME score. Together, the model accounts for 82.4% of variation in PcGGEH.

Data re-analyzed for a subset of nations (N = 84) classified in the Human Development Index [34] as having achieved “medium” to “low” development are reported in Table 3. In general, the associations, albeit weaker, were consistent with those based on the full data set. REGIME continued to be positively associated with expenditures. Among these nations not yet considered to have achieved high development, every unit increment in REGIME score increased the percentage of GDP developed to total health by 0.11%, while the increase in the percentage of general government expenditures spent on health was 0.20%. Per capita total expenditures were estimated to increase by 4.42 Int$, with per capita government expenditures estimated to increase by 3.57 Int$.

4. Discussion

The tenet of the WHO Constitution that “governments have a responsibility for the health of their people [42]” like Rudolf Virchow’s observation that “politics is nothing but medicine at a larger scale [43]” underscores the relevance of central government authority in allocating rights, privileges and responsibilities as a root determinant of public’s health. Whether and how governance affects the way health care is organized, delivered and/or financed is an important, but largely unexamined, subject of study. As global initiatives toward democratization proceed, important questions arise about health thresholds that governments must achieve in securing legitimacy from the populous. Equally consequential is the re-thinking of governance as “risk” given its enduring effects on economic productivity, educational achievement, socio-cultural practices, environmental stewardship, and the like.

The association of a nation’s relative wealth with health care expenditures has been studied, and it is reasonable to attribute significant differences in health care expenditures to economic differences across nations [10,11]. This study diverged from previous studies by considering demographics, political, economic and health system features of nations concurrently with characteristics of their systems of governance. It was found that the authority structure of central governments has had a significant association with levels of health care expenditures. More democratic nations spent more, both proportionately and per capita, on health care than less democratic nations and that relationship persisted while controlling relevant economic, demographic and health system characteristics. For every unit increment in a nation’s REGIME score toward a more democratic approach to governance, this study estimated significant (p < 0.05) increments in the percent of GDP expended on health care (+0.14%), the percent of general government expenditures targeted to health care (+0.25%), the total per capita expenditures on health (+34.4 Int$) and the per capita general government expenditures (+22.4 Int$).

Despite the consistency of findings with the expectations of this study, there are many reasons to withhold judgments about causation. For one thing, analyses of secondary data may be challenged for incomplete, inaccurate and/or inappropriate information. As such, the applicability of findings may be questioned. Differences in health care expenditures across geographic locations, for example, may account for many of the findings presented here. Such an “explanation” is largely useless, however, as it offers neither opportunity to modify nor to understand the mechanisms of the observed associations. In order to meaningfully advance socio-political initiatives to drive health services and health status improvements for populations at-risk, ongoing examinations of the costs and benefits associated with various approaches to governance are needed. This research attempted to quantify expenditure differences among “peer nations” that exhibited comparable levels of development. Relationships reported for the full dataset of 159 nations persisted when data were limited to 84 nations designated to be “medium” and “low” on the Human Development Index [36]. These findings are consistent at the least, with a plausible association between a nation’s type of governance and its commitment to human health, even if they do not establish the view.

Some critics may consider “arbitrary” selection of specific indicators (e.g., Polity IV project scores rather than Freedom House scores, Global Democracy Rankings, etc.), while others may suspect the quality of information collected from poor nations. The use of U.N., WHO and similar public data sources consistent with expectations about how and why such information was gathered (i.e., to advance understanding of factors contributing to human development) offers some reassurance about data validity and the analytic framework, but cannot completely resolve such concerns.

Another related concern pertains to whether these results reflect a genuine causal relation or a mere co-variation between indicators. In a separate analysis, it was found that the regression model that included any of the four indicators of health care expenditures as predictors was significant in accounting for variation in REGIME scores among nations. Only further study using different sources, tools and time frames for what the researchers believe is a relevant, fruitful line of inquiry can resolve such questions.

How health care spending affected health outcomes was beyond this paper’s scope, but the findings of greater expenditures as characteristic of democratic nations portends better health outcomes for those populations. Complementary studies of authority structures and health outcomes have been reported [44–47], but additional research is needed.

What implications do these findings suggest? Health care expenditures frequently occur at the expense of other investments in a society’s infrastructure and institutions. Understanding health care expenditures to be a consequence of underlying democratic structures, in contrast or complementary, to the impacts of larger economies and/or demographic transitions, would suggest the direction and intensity of future expenditures. Decisions by nations and governments about whether and to what degree to fund health care for citizens are important to the physical and socio-political health of the public-at-large. Good health is essential to personal well-being and civic participation. Democratically elected leaders may place greater emphasis on health care than leaders of autocratic institutions in order to appeal to an electorate. A more compelling and plausible argument is that health and well-being may more likely be viewed as a basic human right in democracies and as a privilege in autocracies.

Nations transitioning from an autocracy to a democracy may be particularly stressed to commit greater public resources in support of population health and well-being, the failure of which may limit or destabilize societal institutions during political transition. Cross-national analyses also may inform policy makers about whether a nation’s expenditures are in line with peers or may be effectively modified without substantial detriment to population health status. Such inquiry reflects WHO goals of improving health, addressing legitimate expectations of populations and the fair distribution of government financial burdens.

Abbreviations: