Why do Americans have shorter life expectancy and worse health than people in other high-income countries?

Abstract

Americans lead shorter and less healthy lives than people in other high-income countries. We review the evidence and explanations for these variations in longevity and health. Our overview suggests that the US health disadvantage applies to multiple mortality and morbidity outcomes. The American health disadvantage begins at birth and extends across the life-course, and it is particularly marked for American women and for regions in the US South and Midwest. Proposed explanations include differences in health care, individual behaviors, socioeconomic inequalities, and the physical and built environment. While these factors may contribute to poorer health in America, a focus on proximal causes fails to adequately account for the ubiquity of the US health disadvantage across the life-course. We discuss the role of specific public policies and conclude that while multiple causes are implicated, crucial differences in social policy might underlie an important part of the US health disadvantage.

Introduction

During the last several decades, life expectancy gains in the US have not kept pace with gains in other high-income countries. In 2012, life expectancy in the US ranked 32 worldwide, below most other industrialized nations. Recent reports(5, 10, 12, 60, 64–66, 94) suggest that Americans also experience higher rates of disease, injury and health-damaging behaviors than men and women in other high-income countries. Initial reports noted a US health disadvantage for ages 50 and above(5, 10, 12, 94), but recent reports suggest that American men and women from all ages up to 75 have worse health and higher mortality compared to their counterparts in 13 other wealthy nations in Western Europe, Japan, Australia and Canada(38, 60, 66). Life expectancy among European countries has also diverged and converged at several points, partly coinciding with major wars and economic hardship episodes in European history(57). In contrast, the US health disadvantage emerged during the second half of the 20^th century and has steadily grown, which is remarkable given that this coincided with a period of unprecedented economic growth and stability in the US. This raises questions about specific aspects of post-war America which may be responsible for the US health disadvantage.

In this paper, we review current evidence and theories for the US lag in health and life expectancy. After characterizing the US health disadvantage, we critically discuss common explanations in light of recent studies. Proposed theories so far provide a partial account, falling short of explaining why the US health disadvantage is pervasive across the life-course. We discuss alternative hypotheses and propose a programme of future research on the role of public policies.

The US Health disadvantage

Earlier reports have summarized differences in health and life expectancy between the US and other high-income countries (13, 64–66). This section draws on this literature to illustrate four key features of the US health disadvantage: First, Americans have both higher mortality and morbidity than men and women in other high-income countries. Second, the US health disadvantage begins at birth and extends across the life-course. Third, the lag in US life expectancy is particularly large for American women. Finally, the US health disadvantage is most pronounced for the Midwest and Southeast regions of the US.

Mortality and Life expectancy

Figure 1 shows that the last 50 years have witness remarkable gains in life expectancy in the US and 16 other country members of the Organization for Economic Cooperation and Development (OECD). However, improvements have occurred at different pace across nations(39, 64–66, 70). Between 1960 and 2008, total gains in life expectancy at birth ranged from 15.9 years in Japan to only 6.6 years in Denmark among women; and from 15.1 years in Portugal to 6.1 years in Denmark among men. US gains in life expectancy (7.5 years for women and 9 years for men) have been substantial but only about half of those in the best performing country. Next to the US, Denmark, the Netherlands and Norway have had comparatively modest gains in life expectancy, while women in Japan and Southern Europe (Portugal, Spain and Italy) have enjoyed the largest gains. As a result, in 2008, the US had the shortest life expectancy for both women (80.6) and men (75.6), while life expectancy was longest for Japan (86.1) among women and for Switzerland (79.8) among men.

Figure 1.

Life expectancy at birth in the US and 16 other OECD countries, 1960–2010

Source: OECD Health Data, OECD Health Statistics (Database)(70)

Cross-national variations in life expectancy at ages 40, 60 and 80 years are smaller than differences in life expectancy at birth (Supplemental Figure 1, see the Supplemental Material link in the online version of this article or at http://www.annualreviews.org/), suggesting that excess deaths before age 40 substantially contribute to life expectancy variations between the US and other countries. Nevertheless, with the exception of Denmark, female life expectancy at ages 40 and 60 is lowest in the US compared to any other high-income country. Among men, life expectancy at ages 40 and 60 is similar or better in the US than in Portugal, Denmark and Finland, but shorter than in other high-income countries. Beyond age 80, life expectancy in the US is around average or better than in other high-income countries(39, 58).

The fact that mortality at relatively young ages accounts for much of the US life expectancy disadvantage was highlighted in a recent analysis examining mortality under age 50 across countries(38, 66). Results from this study indicate that mortality differences below age 50 account for two thirds of the gap in life expectancy at birth between men in the US and an average of 17 other OECD countries, and 40% of this difference among women. These findings underscore the point that the US life expectancy disadvantage originates at early age and extends across the life-course.

Causes of death

Age-standardized rates of mortality from selected causes occurring disproportionately at young and middle-age are presented in Figure 2 for women and in Supplemental Figure 2 (see the Supplemental Material link in the online version of this article or at http://www.annualreviews.org/) for men. Mortality rates from infectious diseases; complications of pregnancy, childbirth and the puerperium; and conditions originating in the perinatal period are higher in the US than in nearly all other OECD countries. Differences in some causes emerged around 1980; for example, the gap in transport accidents and accidental poisoning became stark in recent decades due to larger declines in other countries paired with increasing or stagnant trends in the US. In contrast, homicide mortality has consistently been higher in the US for several decades, which is consistent with prior evidence of substantially higher US rates of fire-arm related deaths (46). Recent evidence indicates that the major causes of death contributing to years of life lost below age 50 between the US and an average of 17 other OECD countries among women were non-communicable diseases, perinatal conditions, transport injuries and non-transport injuries(38, 66). Among men, homicide mortality was the largest contributor, followed by transport injuries, non-transport injuries and perinatal conditions(38, 66).

Figure 2.

Mortality from external causes, maternal conditions and infections in the US and other high-income countries, 1960–2010, women

Source: OECD Health Data, OECD Health Statistics (Database)(70)

The contrast between the US and other high-income countries is less stark for mortality from non-communicable diseases that disproportionately affect older populations (Figure 3 for women and Supplemental Figure 3 for men, see the Supplemental Material link in the online version of this article or at http://www.annualreviews.org/). However, mortality from ischemic heart disease (IHD), diabetes, nervous system diseases, and respiratory diseases (women only) is higher in the US than it is most other high-income countries(30, 38). Stroke mortality is lower in the US, although larger declines in other countries have led to a smaller US advantage in recent years(30). While mortality from these causes is driven by mortality at older ages, a recent study concluded that mortality from non-communicable diseases also contribute to excess premature mortality, explaining 29% of years of life lost below age 50 in the US compared to other OECD countries among women, and 18% among men(38). US cancer mortality is relatively low for males and around average for females.

Figure 3.

Mortality from non-communicable diseases in the US and other high income countries, 1960–2010, women

Source: OECD Health Data, OECD Health Statistics (Database)(70)

Differences in health and morbidity

The US health disadvantage is not limited to mortality but extends to many other non-fatal health outcomes beginning at birth and extending across youth, mid-life and old age(5, 7, 10, 32, 65, 66). Supplemental Figure 4 (see the Supplemental Material link in the online version of this article or at http://www.annualreviews.org/) provides an example for selected morbidity outcomes. Compared to most other countries, Americans have higher prevalence of low birth weight, traffic injuries and HIV incidence. Paradoxically, Americans are more likely to rate their own health as good than men and women in other high-income countries, but this pattern appears to be driven by cross-national differences in the style of reporting(12, 19, 43). A recent review shows that Americans have also higher prevalence of preterm births and poor maternal health; adolescent pregnancy and sexually transmitted infections; and overweight, obesity and diabetes during childhood and mid-age(66).

Figure 4 shows that older Americans report a higher prevalence of heart disease, stroke, hypertension, diabetes, obesity, lung disease and limitations with basic instrumental activities of daily living (IADL) than their European counterparts at ages 50 and above. Similar patterns have been reported for ages 50–74(5, 10, 12, 19, 94). While US adults are also more likely to report a cancer diagnosis, this is likely to reflect more aggressive screening and possibly better cancer survival rates in the US (5, 10, 18, 19, 29, 78, 81). Differences between the US and Europe are also evident for biologically assessed outcomes such as blood pressure, blood cholesterol, fasting glucose levels and C-reactive protein(10, 19, 60).

Figure 4.

Prevalence of chronic disease and disability among men and women aged 50 years and older in the United States, England, and Europe: HRS, United States, 2004; ELSA, England, 2004; and SHARE, Europe, 2004

AT is Austria, DE is Germany, SE is Sweden, NL is the Netherlands, ES is Spain, IT is Italy, FR is France, DK is Denmark, GR is Greece, CH is Switzerland, US is United States, and EN is England

Some Americans are at greater health disadvantage

In a series of studies, Murray and colleagues sub-divided the US into eight race-county combinations, referred to as the ‘Eight Americas’, and found large differences in life expectancy between these groups(62, 63). For example, life expectancy for black males living in high-risk urban environments is 21 years lower than life expectancy for female Asian Americans. For young and middle-aged males and females, mortality in the disadvantaged Americas is up to two times worse than that in the worst OECD country(62, 63). Disparities across US regions have grown since the 1980’s, a factor that has contributed to the overall US lag in life expectancy(47, 99).

Supplemental Figure 5 (see the Supplemental Material link in the online version of this article or at http://www.annualreviews.org/) shows that American women in the five census divisions located in the south and Midwest regions have higher mortality than women in most other OECD countries, but even the best US divisions, the US Pacific and New England, have higher mortality than 11 other OECD countries. Earlier reports indicate that even in the healthiest US regions, female life expectancy lags behind that in the least healthy regions of countries such as Japan and France(99). Among men, most divisions in the south and Midwest perform poorly compared to most other OECD countries, while the Pacific and New England divisions have relatively low mortality. Both men and women in the East South Central US divisions have the highest rates.

Explaining the US Health Disadvantage

Table 1 presents an overview of proposed explanations for the US health disadvantage, some of which have been empirically examined(64–66). The evidence reviewed here and elsewhere(13, 64–66) suggests that multiple factors are likely to be responsible for poorer health in the US compared to other high-income countries. In this section, we critically discuss the rationale and evidence for each of these explanations.

Table 1.

Overview of explanations for differences in life expectancy and health between the US and other high-income countries

Broad mechanism	Specific factors
Medical care and public health	Access to health care insurance
	Quality of medical care
	Quality of public health system
Individual behaviors	Tobacco use
	Obesity
	Diet
	Physical inactivity
	Alcohol and other substance use
	Sexual practices
	Violence (especially firearm suicide & homicide)
	Automobile reliance
Social/demographic factors	Socioeconomic inequality and poverty
	Racial disparities and residential segregation
	Social integration and social interactions
Physical Environmental Factors	Built environment (urban design, transport infrastructure, land use mix, urban planning and design)
	Food environment

Medical care and public health systems

The US spends more on healthcare than any other OECD country(70), yet medical care is often proposed as an explanation for the US health disadvantage (13, 64–66). A recent report reviewed evidence of differences in medical care and public health systems; the quality of health care; access to health care services and medical care; timing of care; and the prevalence of medical errors, among others(66). These and other comparisons provide a mixed picture and do not systematically point towards worst quality of care in the US compared to other OECD countries(66, 78, 101). Although whether insurance coverage is causally linked to health status has been debated(83), the lack of universal coverage may be an exacerbating factor (66). Nonetheless, both insured and uninsured Americans experience poorer health than their European counterparts, suggesting that health insurance might not be the only explanation(10, 66).

Overall, health care provides at best a partial explanation. For example, excess deaths from violent causes (homicides, suicides, accidents) are hardly due to lack of health care; indeed, if it were not for advances in emergency medical care, it is estimated that thousands of more homicides would be recorded in the United States each year(34). In addition, US survival rates for several chronic conditions contributing to the US health disadvantage, such as heart disease, ischemic stroke and cancer, might be better in the US than in other high-income countries, suggesting that care for these conditions might not be worse in the US than in other OECD countries(18, 29, 66, 78). Macinko, Starfield and Shi(52) have linked the weaker primary health care system in the US to higher premature mortality(52). Nevertheless, regardless of cross-national differences in access to quality medical care, the fact remains that the overwhelming contributors to the incidence of disease (e.g. poor health behaviors) operate largely outside the influence of medical care.

Individual behaviors

Differences in tobacco use, diet, physical inactivity, obesity, alcohol and other drug use, sexual practices and harmful behavior have been proposed as potential explanations of the US health disadvantage (Table 1). A recent report released by the National Academy of Sciences (NAS) concluded that smoking was likely the most important factor explaining the lag in US life expectancy at older age, particularly among women(64, 65). While the US enjoys currently lower smoking prevalence than most other high-income countries, the smoking epidemic started earlier and reached a higher pick in the US than in other countries, particularly among women(20, 64, 65, 77). Due to the long lag between smoking and lung cancer, current mortality reflects smoking trends two to three decades earlier. A recent study concluded that smoking explained two fifths of the difference in male life expectancy between the US and other high-income countries, and over three quarters of the difference in female life expectancy(65, 79).

Assessing the role of other individual behaviors has proved challenging given limited comparable data on risk factors across decades and countries. Yet, data suggest that the US has one of the highest total caloric intake and the highest sugar intake among all OECD countries(66, 71). The US also ranks high in total fat intake and total protein intake, while vegetable and fruit consumption in the US is similar to that in several other OECD countries(71). A poor diet, in combination with relatively low levels of physical activity(33, 91), may explain the high US obesity rates. Recent estimates based on macro-level data suggest that obesity might explain as much as two thirds of the US shortfall in male life expectancy and two fifths of the US female life expectancy disadvantage(80). However, this contrasts with another report showing that increasing trends in obesity are not specific to the US and might not explain current differences in life expectancy(2). In support of this view, cohort studies suggest that even after adjusting for obesity and other risk factors, differences in morbidity across countries remain(4, 5, 10, 60).

While it is likely that smoking and other unhealthy behaviors contribute to the poorer health of Americans, smoking does not explain why Americans have poorer health and worse trends in mortality below age 50(6, 38, 66). It is unlikely that parental smoking alone could account for the higher rate of infant mortality, poorer childbirth outcomes, injuries and homicide in the US compared to other high-income countries. Since mortality below age 50 from these and other causes explains two thirds of the difference in life expectancy at birth between the US and other countries among men, and two fifths among women(38, 66), smoking is at best only one among several factors explaining the US health disadvantage.

While understanding the contribution of individual behaviors is crucial, an approach that focuses solely on behavioral differences is impoverished by its focus on “proximal” individual choices. The earlier adoption of smoking among US females, for example, may reflect features of the US environment that encouraged American women to smoke more than women in other countries. The fact that Americans behave poorly only raises the follow-up question of why Americans more often than adults in other countries make behavioral choices that are detrimental to their health.

Social and demographic explanations

The US is characterized by pronounced racial, ethnic and socioeconomic disparities in health, which may contribute to the overall US health disadvantage. For example, although life expectancy for the United States as a whole improved during the past three decades, Ezzati et al.(24) documented declining or stagnant life expectancy between 1983–1999 for women in 963 out of 2,068 counties, and 59 counties for men. The history of the United States also diverges from that of other OECD countries in terms of its legacy of three centuries of slavery followed by post-abolition Jim Crow laws (1876–1965) which shaped racial segregation, whose cumulative influences are still felt to this day. For example, Williams & Collins(98) argue that the persistent residential segregation of African Americans shapes their educational opportunities and labor market success, and contributes to their unequal exposures to environmental pollutants, violence, and other health threats.

While it is true that poor and black Americans are at increased health disadvantage, studies suggest that also white, middle class Americans have poorer health than their European counterparts(5, 7, 10). For example, in a widely cited cross-national comparison of the health of American and English people, Banks et al.(10) found that Americans in the top third of the income distribution (97% of whom already have access to health insurance) had rates of hypertension and diabetes comparable to those in the bottom third of income earners in England. The comparison was all the more striking because it was restricted to whites in both countries.

Nevertheless, the largest share of the American health disadvantage is likely to be borne by the poor and least educated, who have much higher rates of disease and death than their counterparts in Europe(5, 7, 10, 59). The role of socioeconomic status may be particularly salient for mortality under age 50. For example, US mortality from homicide is nine times higher among young men in the bottom decile of socioeconomic deprivation compared to young men in the affluent top decile(88, 90). Strikingly, US girls in the bottom decile are fourteen times more likely to die from HIV/AIDS than their counterparts in the top affluent decile(88, 90). Similar differences by socioeconomic deprivation exist in childhood mortality(89, 90).

Turning to the role of social integration, a recent paper concluded that social participation and integration did not explain the US health disadvantage relative to other European countries(8). Limited evidence on the extent of variations across countries makes it difficult to assess whether they contribute to the US lag in life expectancy(32, 64, 65).

The built physical environment

A separate line of explanations argues that aspects of the built physical environment, such as access to recreational facilities, land use mix, transportation infrastructure, urban planning and design, as well as access to fast food outlets and fresh fruit and vegetable stores, might underlie cross-national variations in healthy behavior and associated health outcomes(66). For example, the built physical environment in most of the US provides limited opportunities for physical activity with few alternatives other than driving. The reliance of Americans on automobiles as their primary mode of transport is well documented(45). Interestingly, the fatality rate per 100 million vehicle kilometers travelled is similar in the US and a set of other 15 high-income countries, but the annual number of kilometers driven in the US far exceeds that in other countries (38, 96). In other words, Americans die more from car crashes because they drive more.

There is an extensive literature on the relationship between the built environment and healthrelated behavior(26), but there are no systematic investigations of the contribution of the built environment to the US health disadvantage. Although it is difficult to draw firm conclusions, explanations based on the physical environment beg the follow-up question of why the US has less health-promoting built environments compared to other countries.

Upstream policies and the US health disadvantage

Proposed explanations do not explain why Americans of all ages behave poorly; live in a physical environment that is less conductive to health; and why their health suffers more from socioeconomic deprivation than the health of populations in other high-income countries. This raises the question of whether upstream policies might underlie some of these health variations(66). Table 2 summarizes specific areas of social policy that differ dramatically across the US and other high-income countries, and for which there is some evidence that they may influence health and mortality. This section speculates on how these policies might offer promising avenues for future research on the upstream causes of the US health disadvantage.

Table 2.

Public policies that may contribute to differences in health and life expectancy between the US and other high-income countries

Public Policy domain	Specific programmes
Childcare and early childhood education policies	Policies determining the availability, cost, and quality of childcare and early childhood education programs
Education policies	The share of public vs. private education systems
	Compulsory schooling laws
	Spending and distribution of resources for education
	Access to higher education
Labor and Employment	Labor laws that affect job security, work conditions, working hours, worker's benefits and work flexibility
Protection Policies
	Parental leave
	Minimum wage laws
	Trade union membership laws
	Work incentives and worker's compensation
	Retirement policies
	Unemployment insurance policy
	Active labor market programmes
Income support and Family and children support policies	Child poverty alleviation and income tax credits
	Family allowance programmes
	Child support maintenance systems
	Child-related leave
Housing policies	Incentives for homeownership
	Access to public housing
	Policies to improve housing conditions
Income inequality	Tax and redistribution policies

Childcare and early childhood education policies

Early childhood education in the US is less well established than in Europe, where formal and subsidized pre-primary education is often the norm. The typical starting age for early childhood education in the US is four years, compared to three years or younger in 21 other OECD countries(68). While regulations in most of Europe require that a qualified teacher delivers a formal curriculum, this is less well regulated in the US(72). The overall enrolment rate in early education programmes is 69% in the US, compared to rates above the OECD average of 80% in most European countries. While 84% of children in the OECD attend public or Government-funded private institutions, only 55% of early childhood pupils in the US attend public schools(68). In addition, as a percentage of GDP, the US spends far less on childcare support for families than almost any other OECD country(72).

Early childhood interventions appear to bring important health benefits, especially among disadvantaged children (37, 44). Studies indicate that early education programmes do not only improve educational outcomes but also lead to higher immunizations and height-for-age, and reduce child mortality at ages five to nine(37, 44). More comprehensive childcare and early education programmes for children in Europe may thus partly contribute to their better health compared to that of American children.

Education policies

There are important differences between the US and other high-income countries in education policy. While the US spends more on public school education than most other OECD countries(69), American students perform around or below the OECD average(69). There are moreover substantial disparities in the quality of public schooling (e.g. reflected by student-teacher ratios) across communities in the United States, which are partly driven by residential segregation and the financing of the public school system by local property taxes(69). While educational attainment is relatively high in the US(74), inequalities in spending may lead to substantial disinvestment among socially disadvantaged groups most at risk of poor health.

Evidence from across the US (31, 50, 51) and Europe(11, 97) suggests that education policies such as compulsory schooling laws have had long-run effects on health and mortality. Other policies, such as education grant aid programmes, have increased schooling completion and college attendance(22). The health benefits of these and other policies expanding access to and improving quality of education remain poorly understood.

Labor and employment protection policies

The US stands out for its weaker employment protection laws (euphemistically referred to as “labor flexibility”) compared to other OECD countries. US workers face comparatively high risks of job displacement, as employers bear relatively low costs associated with collective dismissals or contract termination(67, 76). In addition, social policies to protect workers who become ill or displaced, as well as maternity leave policies, are modest in the US in comparison to most European countries(67). For example, net wage replacement rates for long-term unemployment insurance for a single-earner married couple with two children in 2010 were 45% in the US, compared to 66% in France, 86% in Sweden and 90% in Japan(74). Programmes to support working parents are also substantially less comprehensive in the US. In 2011–2012, the duration of fully-paid maternity and parental leave was 45 weeks in France, 46 weeks in Sweden, and 21 weeks in the Netherlands, compared to none in the United States(72). A potential hypothesis is that as a result of these policies, Americans work longer hours(82); spend less time cooking and eating meals at home(15); drive more as opposed to investing time in healthy transportation alternatives(38, 96); and in general spend less time in non-market activities that might be conductive of health.

Some employment protection policies have been shown to improve health. For example, extending weeks of job-protected paid maternity leave significantly decreases infant mortality rates and improves child health(84), with the large effects on post-neonatal mortality(85, 93). Longer maternity leave may also improve maternal mental health around the post-partum weeks(16) and increase mothers’ labor market attachment(14, 86), leading to lung-run benefits for mothers and children(86). Statutory retirement age laws may also influence health and mortality(17). Less is known about the health impact of unemployment insurance and other employment protection laws. However, the negative effects of unemployment on workers' subsequent earnings are mitigated through generous unemployment benefit systems or strict labor market regulation(27, 28). More evidence is needed to assess whether differences in these policies contribute to the US health disadvantage.

Income support policies

After taxes and transfers, poverty rates are considerably higher in the US than in other OECD countries, particularly among children. In 2010, 21% of children in the United States lived in poverty, compared with 11% in France, 10% in the United Kingdom, and 8% in Sweden(75). In 2006, educational deprivation –a measure of whether children have the necessary items for school - was 5% in the United States, as opposed to 1% in France and 2% in Sweden and the United Kingdom. These differences partly reflect the fact that cash minimum-income benefits are considerably lower in the US than in most other OECD countries(73).

There is some evidence that income transfer programmes have important health effects on low-income mothers and their children. Expansions of the Earned Income Tax Credit may have led to increased birth weight and reduced maternal smoking(92). Similarly, pregnancies exposed to the Food Stamps program had better birth outcomes than pregnancies unexposed, particularly among African American mothers(3). Income transfer programmes may also improve the health of older Americans. For example, an increase in state maximum Supplemental Security Income benefits was shown to reduce disability among older Americans(36). Comparable programmes in Europe are far more comprehensive than those in the US(1, 67), which may contribute to the poorer health of Americans(5, 7).

Housing Policies

Larger differences exist in policies to promote access to public housing and homeownership across the OECD. Many European countries offer generous cash housing benefits for rental accommodation for families in need(67). By contrast, there is no US Federal programme for housing assistance, with only some states delivering programmes targeted to very low income households(73). Life-cycle housing wealth accumulation patterns also differ across the US and Europe. Compared to US adults, British adults move into homeownership at younger ages, and a larger fraction of their wealth is concentrated in housing(9), while American hold a larger fraction in financial assets. Policies promoting homeownership in some European countries may have contributed to these differences. For example, the ‘right-to-buy’ scheme, introduced in 1980, granted UK households living in government housing for a minimum duration the right to buy their home with large discounts, which may have contributed to their larger housing wealth compared to US households(9).

The health impact of housing policies is poorly understood, but evidence suggests that housing itself is important to health. Results from the Moving to Opportunity (MTO) project, a randomized experiment in which families in poor neighborhoods were offered vouchers and assistance to move to ‘low poverty’ neighborhoods, showed improvements in mental health, behavior and educational achievements for young girls(49) and adults(87) (albeit it may have led to poorer outcomes among boys). A recent study in Chicago showed similar benefits of a programme randomly offering housing vouchers on female child mortality(42). Recent trials and policy evaluations also suggest that improvements in housing conditions, such as insulation and ventilation, effectively reduce hospitalizations and improve child health outcomes(40, 41). Homeowner occupiers have better health(21, 23, 35, 53–55, 61, 100) and lower mortality(25, 48) than renters, although whether this is due to selection or actual health benefits of homeownership is yet unknown. While more research is needed, differences in housing policy may contribute to differences in child and adult health between the US and other OECD countries.

Tax policies, redistribution and income inequality

European countries have more progressive tax systems and are designed to protect the poor to a larger extent than the US. As a result, not only are social policy programmes more comprehensive in Europe(1, 73), but income and wealth inequalities are also smaller than in the US(102, 103). A potential hypothesis is that Americans have poorer health because they have larger income and wealth inequalities. However, evidence on the causal impact of income inequality on population health across high-income countries is as yet inconclusive(32, 56). On the other hand, there is some suggestion that income inequality may have a causal effect on causes of death contributing to excess US mortality below age 50. A recent study using panel data from 21 developed countries found that income inequality increases mortality up to age 15 for females, and up to age 50 for males(95). Research on the role of income inequality in explaining differences in years of life lost below age 50 between the US and other OECD countries offers a potential avenue for future research.

Future studies: Do public policies contribute to the US health disadvantage?

The evidence discussed above suggests that public policies on early childhood, education, employment, income support, housing and income redistribution might influence health and mortality. A separate line of evidence suggests that these programmes are less comprehensive in the US than in most European countries. Somewhat surprisingly, however, there is as yet no literature linking these two phenomena to estimate to what extent public policies are causally linked to the US mortality disadvantage. One potential reason for this is the fact that most policy evaluation studies focus on local programmes targeted to sup-populations within a specific country, while less is known about the impact of national public policies on population health and mortality.

To advance our understanding of how public policies might be linked to the US life expectancy disadvantage, we propose a line of future research following two streams. We first call upon quasi-experimental studies that exploit the large variation in public policies across OECD countries to estimate their causal impact on national mortality. While cross-national studies are not without limitations, several successful examples suggest that they can provide useful insights on the impact of policies on population health. For example, Ruhm(85) and Tanaka(93) used data on parental leave policies across OECD countries over the last decades to estimate their impact on infant mortality. These studies exploit variations over time across countries in the timing and generosity of paternity leave policies in country fixed effect models. A second line of research should examine to what extent public policies shown to influence health and mortality explain differences in mortality trends between the US and other high-income countries. An approach to address this question is to use simulation models that combine estimates from quasi-experimental studies on the causal effect of national policies with data on the timing and comprehensiveness of public policies across countries.

To illustrate, in 2009, the infant mortality rate was 6.4 per 1,000 live births in the United States, compared to 3.9 in France(70). Infant mortality was thus 64% higher in the US than France. Following up on the example of maternity leave policies, results by Tanaka(93) suggest that a 10-week extension of the number of weeks of fully paid job-protected parental paid leave is associated with a reduction of 2.5% in infant mortality, and as much as 4.5% reduction in post-neonatal mortality. US Federal law requires no paid parental leave, while France, for example, requires 16 weeks of fully paid leave. What would infant mortality be in the United States if mothers in the US had similar parental leave benefits as mothers in France? Combining information on the causal effect of the policy with data on parental leave benefit generosity across countries over the last decades in a simulation model would provide us with an estimate of the extent to which parental leave policies account for the larger infant mortality in the US than France. A similar approach can be applied to other public policies to examine their contribution to the US mortality disadvantage.

Conclusion

Our review leaves little doubt that the US has poorer health and shorter life expectancy than other high-income countries. We find that the US health disadvantage begins at birth, extends across the life-course, and is more pervasive for Americans living in the South and Midwest of the US. Differences in health care, individual behavior, socioeconomic inequalities and the physical environment are all likely to contribute to the explanation, yet they offer only a partial account for the pervasiveness of the US health disadvantage across the life-course and for many different outcomes. We hypothesize that much of the US health disadvantage is due to variations in non-medical determinants of health, some of which result from dramatic differences in public policies across the US and other OECD countries.

There is ample evidence that social policies and programmes potentially affecting health across the life course are less comprehensive in the US than in other OECD countries. This includes policies affecting outcomes since early childhood –through less comprehensive early education and childcare programmes; at early adulthood and middle age –through more unequal access to high-quality education and less comprehensive employment protection and support programmes; and at older ages –through less comprehensive housing and income transfer programmes affecting older individuals. While the impact of many of these policies on social outcomes is well documented, the extent to which they influence health and contribute to differences in longevity among high-income countries is yet to be established.

We propose a future line of research that combines estimates from quasi-experimental studies on the impact of policies from cross-national studies, with data on the timing and comprehensiveness of these policies across the US and other OECD countries to simulate a counterfactual: What mortality and life expectancy would be in the US if it had the public policies shown to influence health and mortality available in other OECD countries. While not a trivial task, disentangling the role of public policies is crucial to unravel why the most prosperous economy during the second half of the 20^th Century continues to lag behind other high-income countries in life expectancy.

Acronyms and definition list

OECD

Organization for Economic Cooperation and Development

US

United States

IHD

Ischemic Heart disease

IADL

Instrumental Activities of Daily Living

NAS

National Academy of Sciences

SHARE

Survey of Health, Ageing and Retirement in Europe

HRS

Health and Retirement Study

ELSA

English Longitudinal Study of Ageing