Population Health: Behavioral and Social Science Insights

By Samuel H. Preston

Abstract

Improvements in health and length of life for the average person have been among the greatest of human achievements. Global life expectancy at birth has risen from about 31 years in 1900 to 70 years today. In this chapter, I describe some important contributions to understanding the principal factors driving improvements in health since the mid-1800s. While this chapter does not represent a comprehensive account of the relevant literature, it does provide some examples and discussion of useful analytic studies that examine the progress that has been made over the past 150-plus years toward improving health and life expectancy, both globally and here in the United States.

Introduction

In the United States, life expectancy has increased from about 48 years in 1900 to 79 years today.^1,2 Nordhaus³ estimated that the imputed value of gains in life expectancy in the United States between 1900 and 1995 was approximately equal to the entire gain in per capita income over the period. If the value of improvements in morbidity were factored into the assessment, the value of health gains would have exceeded the value of increased consumption of all other goods and services combined (also Murphy and Topel⁴).

Social scientists have been the key contributors to understanding the broad social forces that have driven these improvements in health. Their central role in these discussions is a product of their concern with accurate measurement, their attention to issues of research design, and their emphasis on understanding population-level phenomena. While medical sciences have focused on molecules, micro-organisms, genes, and physiology, social scientists have attempted to assess what advances in the broad realms of personal income, medicine, and public health have meant for levels of population health. The results have helped to calibrate public investments in health and in other social sectors.

Most studies of these relations have used indicators of mortality. The main reason for using these measures is that they are often available from registries of death that usually span long periods of time and include nearly 100 percent population coverage. Furthermore, death is an unambiguous event whereas measures of self-assessed health and of disease incidence and prevalence are subject to many forms of error and bias. Costa⁵ demonstrates for the United States that secular improvements in longevity have been accompanied by huge reductions in the prevalence of major chronic conditions.

The subject of this review is massive, and what follows should be considered a set of illustrations of useful analytic studies rather than a comprehensive account of the relevant literature.

Selected Case Studies

England and Wales

The first serious effort to identify major factors responsible for improvements in longevity was made by a medical historian, Thomas McKeown. McKeown took advantage of the longest series of vital statistics on causes of death for a national population, that pertaining to England and Wales, which dates back to 1838. A series of journal articles culminated in a 1976 monograph.⁶ McKeown showed that the mortality decline was primarily attributable to declines in infectious and parasitic diseases and that, for disease after disease, the bulk of the decline had occurred before any medicine or therapy was available to combat the disease. Tuberculosis, the single most important disease, had declined by some 80 percent before effective medical therapy was available. McKeown argued that, if improvements in medical treatment were not responsible for the declines in infectious and parasitic diseases, then improvements in standards of living must be responsible. In particular, he attributed the bulk of the mortality decline to improved nutrition. McKeown presented no direct evidence about nutrition's role, and his process-of-elimination reasoning failed to eliminate some obvious alternatives, particularly improvements in preventive public and personal health practices.

Especially important elaborations of the part played by public health initiatives in England have been provided by Szreter⁷ and Woods.⁸ In a thorough and well-considered monograph, Mercer⁹ assembles a great deal of evidence about the importance of public and personal health practices in the mortality decline in England and Wales. McKeown did make one very important point that has for the most part stood the test of time: specific therapeutic medical treatments have played a minor role in reductions in infectious disease mortality in now-developed countries.⁶

United States

The role of diet appears to be modest at best in the 20th century mortality decline in the United States. The United States was already a very well-fed country by 1900, when life expectancy at birth was only 48 years; dietary reconstructions from direct inquiries suggest that the average daily caloric consumption per adult equivalent was about 3,700, higher than today.¹⁰ Demographers Ewbank and Preston¹¹ argue that the essential element in U.S. gains in mortality between 1900 and 1930 was an enormous scientific breakthrough—the germ theory of disease. This theory was empirically validated in the 1880s and was beginning to displace the misguided miasma theories by the turn of the century. The development and diffusion of the theory and its practical technologies is a plausible explanation of why the frontier of national life expectancy increased faster between 1880 and 1950 than in any equivalent period, as shown by demographers Vallin and Mesle.¹²

Although the new theory led to few practical drugs, as McKeown had argued,⁶ it did lead to an entirely new approach to preventive medicine, practiced both by departments of public health and by individuals. While public health departments had been concerned about the need to improve water supply and sewage disposal processes before the germ theory, the criteria for success were sight and taste and odor, rather than bacteria counts. All this changed rapidly. Enlightened public health officials were quick to recognize how the germ theory should guide their practice. in an important article, economists Cutler and Miller¹³ investigated the relation between changes in municipal mortality and the timing of water supply and sewage improvements in 13 American cities during the early decades of the 20th century. They concluded that 43 percent of the decline in the death rate in these cities between 1900 and 1936 was attributable to improvements in the cleanliness of water. Economist Richard Easterlin¹⁴ examined the mortality decline in the United States between 1880 and 1950 and concluded that the role of public policy was of critical importance. The market could not be counted on to provide pure water or milk, to control pests, or to disseminate knowledge about hygiene and infant care. Market failure in the face of the new bacteriology prompted a massive expansion of departments of public health.

By the time of the first White House Conference on Infant Mortality, held in 1909, public health officials realized that rapid advances in longevity required that they go beyond their normal domain of public works and attempt to change the personal health practices of individuals. The germ theory provided a number of powerful weapons for doing so. These included boiling bottles and milk, washing hands, protecting food from flies, isolating sick children, and ventilating rooms. Public health officials launched massive campaigns to encourage these practices.¹¹ Cutler, Deaton, and Lleras-Muney¹⁵ usefully draw the distinction between "macro public health" (filtering and chlorinating water supplies, building sanitation systems, draining swamps) and "micro public health," which comprises features of individuals' behavior that were often influenced by governmental efforts.

The Children's Bureau, created in 1912, adopted a primary focus on child health and producted a pamphlet, Infant Care, that became the largest selling publication in the history of the Government Printing Office, with some 12 million copies sold by 1940. By the 1920s, the Bureau was receiving and answering over 100,000 letters a year from parents seeking child care advice. Economic historian Joel Mokyr^{16, 17} documents the arduous changes that the germ theory introduced into household practices. The new hygienic practices appear to have been most vigorously adopted among professionals. Ewbank and Preston¹¹ show that, at the turn of the century, the children of physicians had mortality levels that were scarcely better than those of the average child, suggesting that physicians had few weapons at their disposal to advance survival. By 1924, the mortality of physicians' children was 35 percent below the national average. Children of teachers also advanced rapidly, and children of all professionals made great strides during the period.

Sri Lanka

The materials to study mortality decline in developing countries are much thinner than those in the developed world. Very few countries had vital registration data that could properly measure levels and changes in mortality during the period of rapid, often dramatic, improvements, following World War II. Sri Lanka (Ceylon) is an exception, thanks to a vital registration system established during British colonization.

Sri Lanka achieved a remarkable decrease in mortality after World War II; the crude death rate fell from 21.5 per 1,000 in 1945 to 12.4 in 1950. This period coincided with an insecticide-spraying anti-malarial campaign. The combination of data availability and dramatic change has drawn the attention of a score of social scientists. The most convincing study is by the economist Peter Newman,¹⁸ who examined the regional distribution of mortality declines in comparison to estimates of pre-campaign malarial endemicity. He concluded that the malaria eradication campaign in 1946 reduced Sri Lanka's crude death rate by 4.2/1,000 between 1936-1945 and 1946-1960, representing some 42 percent of the reduction in mortality during the period. Newman's study gained prominence because control of malaria appears to have been the single leading source of mortality decline in developing countries after World War II.¹⁹ In addition to antimalarial programs, immunization and antibiotics are thought to have played an important role in the rapid mortality declines in developing countries after World War II.²⁰ Unfortunately, the mortality data are simply too weak to draw a very precise quantitative conclusion about their importance.

International Studies

McKeown's claim that improved standards of living were responsible for the bulk of mortality improvements⁶ was put to a test by Preston.²¹ Using time series and cross-sectional evidence for nations with relatively good data during the 20th century, I demonstrated that a tight cross-sectional relationship existed between national levels of life expectancy and per capita income during the 1900s, the 1930s, and the 1960s. The correlation between the two series (with per capita income logged) was .88 in both the 1930s and the 1960s. Such a tight relationship implied that living standards, as captured in per capita income, were a dominant determinant of mortality.

The curve relating life expectancy to income shifted decisively during the century, however, as shown in Figure 1. If the contribution of gains in per capita income to improvements in life expectancy were effectively captured by the cross-sectional relations, then the gain in life expectancy attributable to gains in per capita income between the 1930s and the 1960s would have been about 1.3-2.5 years. In contrast, worldwide gains in life expectancy were about 12.2 years during the period. So, I concluded that factors exogenous to a country's current level of income probably account for about 85 percent of the growth in life expectancy for the world as a whole between the 1930s and the 1960s.²¹ The proportions were similar in industrialized and developing countries.

I have argued that the shift in the curve, and hence the bulk of advance in life expectancy, was primarily attributable to improvements in health technology and its application.²¹ The germ theory of disease was prominently featured in the account. However, no direct evidence was presented on the role of technology; it was a residual explanation. Life expectancy and per capita income were featured in the analysis because social scientists had developed good measures of each; no equivalent measure of the quality of health technology had or has yet been developed.

Figure 1. Scatter diagram of relations between life expectancy at birth and national income per head for nations in the 1900s, 1930s, and 1960s

Source: Preston.²¹ Used with permission.

The analytic apparatus in Figure 1—dubbed the Preston Curve by Angus Deaton²² —has been extensively employed. Maintenance of a strong cross-national relationship between mortality and income and continued shifts in the cross-sectional curve have been demonstrated by the World Bank,²³ Kenny,²⁴ Pritchett and Viarengo,²⁵ and Hum et al.²⁶ Pritchett and Viarengo²⁴ examine the relationship from 1902 to 2007. They conclude that the slope of the relation (elasticity) has remained quite constant and that "There is no indication the Preston curve is 'breaking down' and no indication from over 100 years of data that a very strong relationship between national income and life expectancy will not persist." They also find that the curve has continued to shift.

One question that has been raised about the curve is the direction of causality that it represents. While richer countries and richer people have the resources needed to extend life, it is also the case that healthier people are more productive. Adding complexity to the question is that lower mortality produces faster population growth, which can produce diminishing returns to labor.²⁷ A volume addressing the effects of improvements in health on economic productivity concluded that results were "inconclusive."²⁸,^a The most recent, but surely not the last, paper on this issue concludes that advances in health have not contributed to economic growth.²⁹ In the absence of compelling evidence to the contrary, it seems reasonable to treat the relation between life expectancy and income as reflecting primarily the effect of income on life expectancy. Confirmation of the causal impact of income on mortality is presented in an important article by economists Prichtett and Summers.³⁰ They used instrumental variables to model the effects of economic advances on infant and child mortality and concluded that country differences in income growth rates over the previous three decades explain roughly 40 percent of the cross-country differences in mortality improvement.

As can be seen in Figure 1, the relationship between income and life expectancy is curvilinear. The curvature suggests that a policy of redistributing income from rich to poor would improve average health outcomes, since the gains in health of those with low incomes would outweigh the losses of those with high incomes. There is little agreement about whether inequality has additional effects on health and mortality beyond those associated with this non-linearity (for opposing views, see Wilkinson and Pickett³¹ and Deaton³²).

Amendments and Extensions

Although the relationship between life expectancy and income has retained its basic shape since the 1960s, there have been several regional departures from the basic form. During the 1990s, the HIV/AIDS epidemic in Sub-Saharan Africa reduced life expectancy by as much as 20 years in the hardest-hit countries. This epidemic, now receding, had the effect of strengthening the relationship between income and life expectancy.^31,33 A second departure occurred in Eastern Europe and the former Soviet Union, where life expectancy has fallen below that implied by levels of income. The most persuasive explanation for the poor performance in this region is excessive consumption of alcohol, especially among males, whose life expectancy has fallen further behind the norm.³⁴ A third departure is the United States, a wealthy country with a level of life expectancy that has fallen further and further behind its affluent peers since 1980.

As these examples imply, one of the benefits of finding a close association between variables is the lessons that can be drawn from exceptions to the rule. Analyses of reasons for the poor performance of the United States relative to its high level of income have drawn a great deal of attention. It is clear that, above age 50, the history of exceptionally heavy smoking in the United States and its position as a frontrunner in the obesity epidemic are implicated in the country's poor performance.³⁵ But the poor performance by the United States also extends to ages below 50, including especially high rates of accidental and violent death among young men as well as high levels of infant mortality. In fact, a report demonstrating the pervasiveness of the U.S. disadvantage indicated that the United States had the highest age-specific death rates at all ages below 70 among 17 comparison countries.³⁶

The Woolf and Aron report³⁵ is one of the most heavily cited publications in the history of the National Research Council. It was the subject of nearly 100 newspaper or journal articles and has been downloaded more than 21,000 times. It also has been the subject of congressional hearings, congressional staff briefings, public commentary by members of Congress, and a response from the Department of Health and Human Services (data supplied by press office, National Research Council).

The examination of national exceptions has been extended to developing countries. John Caldwell, an Australian demographer, studied in great detail three populations in developing countries where life expectancy far surpassed what could be expected based on income in those countries: Sri Lanka, Costa Rica, and Kerala state in India.³⁷ Caldwell concluded that the principal reason for the exceptional performance in these regions was a history of women's empowerment, which manifested itself in women's high levels of schooling, greater authority within the family, and more intensive political participation. Greater political participation in turn helped to generate more egalitarian governance, including great emphasis on public health. Deaton³⁸ elaborates on the argument that a strong political voice for the poor is an important ingredient in supporting public health programs. Predominantly Muslim countries, with unusually low levels of female autonomy, help to confirm Caldwell's inference by having typical life expectancy levels far below those expected by their levels of income.

Caldwell's research³⁶ raised the question of whether educational attainment and/or literacy were more important drivers of mortality decline than income. Of course, educational attainment and income are highly correlated, so efforts to separate their effects are vulnerable to measurement error and other sources of imprecision. The addition of literacy or educational attainment to income in cross-national analyses of life expectancy has typically found significant coefficients on both.^19,24 Gakidou et al.³⁹ argue that increases in women's education over the period 1970-2009 in developing countries could account for 51 percent of the reductions in child mortality over the period. However, the estimated micro-level relation between women's education and child mortality that was used in the macro-level inference is not robust to selectivity biases. The contribution of changes in parental educational attainment to improvements in child mortality remains an important open question. Research designs that focus on exogenous changes in educational requirements, rather than on educational achievements per se, would help to resolve issues of selectivity (e.g., Klemptner et al.⁴⁰).

Developed Countries Since 1960

Aided by sulfa drugs since the 1930s and antibiotics since the 1940s, the bacteriologic revolution had essentially played itself out by 1960 in the developed world. If there were to be any additional gains in life expectancy—and contemporaries were skeptical about the possibilities—such gains would have to come from advances against the chronic diseases associated with older age, especially cardiovascular disease and cancer.

Between 1960 and 2000, age-standardized cardiovascular death rates declined by slightly more than 50 percent in the United States.¹⁵ Cardiovascular disease mortality reductions account for 70 percent of the 7-year increase in life expectancy between 1960 and 2000. Between 2000 and 2009, death rates from cardiovascular disease declined by an additional 31 percent.⁴¹ Other developed countries enjoyed comparable or even faster declines.⁴² Cutler⁴³ matches the results of clinical trials to actual mortality declines and attributes the bulk of the decline in cardiovascular disease mortality—as much as two-thirds of the reduction—to medical advances. Ford et al.⁴⁴ look specifically at the 50.2 percent decline in coronary heart disease mortality between 1980 and 2000. They conclude that approximately 47 percent of the decrease can be attributed to improved treatment, including secondary preventive therapies after myocardial infarction or revascularization (11 percent), initial treatments for acute myocardial infarction or unstable angina (10 percent), treatments for heart failure (9 percent), revascularization for chronic angina (5 percent), and other therapies (12 percent). Approximately 44 percent can be attributed to changes in risk factor distributions, especially reductions in smoking.

The technologies that were developed from the bacteriologic revolution were primarily embodied in public goods, at least during the major mortality declines in developed countries. In contrast, the new health technologies associated with improvements in cardiovascular disease were more like private goods, requiring both individual purchases and, often, skills in self-administration.⁵ One might expect them to be more widely diffused and more effective among richer countries and richer people within countries.

Mackenbach and Looman⁴⁵ reexamined the Preston curve in a group of European populations. They found that the shift in the curve has declined in magnitude since 1960, and the shift accounted for only one-fourth to one-half of the increase in life expectancy in Europe as a whole since 1960. Most of the gains over the period were attributable to increases in income. Countries with larger increases in income typically had larger gains in life expectancy, although the connection was not statistically significant. The slope of the relation between life expectancy and national income was quite constant between 1960 and 2008. However, the relation between changes in income and declines in cardiovascular mortality was significant in the expected direction.

Within countries, socioeconomic disparities in health and mortality are growing in most places. Kunst et al.⁴⁶ found a growing disparity in mortality by income quintile in most European countries, although not in Nordic countries. Glied and Lleras-Muney⁴⁷ found that education gradients in mortality in the United States became steeper for causes of death that had experienced greater technological progress. Goldman and Smith⁴⁸ found that the health benefits associated with additional schooling in the United States rose over time as measured by self-reported health status. This rise can be attributed to both a growing disparity by education in the probability of having major chronic diseases during middle age as well as better health outcomes for those with each disease.

The growing disparity in health outcomes for those with a disease may in fact be attributable to the greater access of better educated people to superior technology. But their lower prevalence of disease is also partly or largely attributable to better health habits.^49,50 That growing disparities by income class are not inevitable is suggested by the fact that major health improvements can be attained simultaneously with reductions in spending—that is, by smoking less, eating less, and walking instead of driving.

Issues of Timing

Table 1 summarizes a large literature on speed of mortality reduction and the causes responsible for the reduction. It includes studies of the period before 1880, which have not been the focus of this review.^{5,15,37,51,52} Since 1945, advances in the practice of medicine have played an important role in both developed and developing countries, although the claim is on shakier footing in the latter case because of poor data. The forces of improvement were sometimes fighting an uphill battle against unhealthy behaviors that left a clear mark on life expectancy: alcohol abuse in Eastern Europe, smoking and obesity trends in the United States, and practices associated with HIV/AIDS in sub-Saharan Africa.

Annual advances in life expectancy have become a routine expectation in developed countries, and the expectation is nearly always realized. Oeppen and Vaupel⁵³ examined the time series of the maximum life expectancy in a national population since 1800. They found that the gains have been steady at about 0.25 years per year. Vallin and Mesle¹² excluded weaker data from the calculation and found that the bacteriological era between 1880 and 1950 saw the fastest improvements at the life expectancy frontier. Even so, the frontier has advanced by 0.20 year per year since 1950.

Table 1. Speed of health improvements in different eras and main contributing factors

Note: Slow advance = gain of approximately 0.1 year of life expectancy per calendar year. Moderate advance = gain of approximately 0.2 year of life expectancy per year. Fast advance = gain of approximately 0.4-0.5 year of life expectancy per year.

Such steady improvements have given rise to a sense of inevitability in the advances in life expectancy. One possible mechanism producing such steady advances is a medical research establishment that continually reorients its aim towards the major health problems of the day. Such a process might be reflected in the "hand-off" from bacteriologic advances to improvements in cardiovascular disease treatment during the 1950-1970 period. But such a hand-off is not automatic. Cancer was a more promising target than cardiovascular disease in 1970, eliciting a national "war on cancer." But age-standardized death rates from cancer declined by only 12 percent between 1970 and 2008, in contrast to a decline of more than 50 percent for cardiovascular diseases.

This chapter has dealt only with contemporaneous relations between health outcomes and causal factors. Many causal factors operate with a lag. One clear example is smoking, which affects lung cancer and all-cause mortality with a substantial lag.⁵⁴ Some conditions in childhood also influence mortality throughout the lifetime of birth cohorts. These include infections, developmental problems, and nutrition in utero and during childhood. Studies of these relations are too voluminous to review here. There is no doubt that poor health in childhood is highly correlated with poor adult health.⁵⁵ Sweden and England both show clear cohort influences in their patterns of secular mortality decline, most likely a product of the persistent effects of tuberculosis infection on frailty throughout life.^56,57 However, cohort effects appear to be much less influential than period effects over time and space (Barbi and Vaupel⁵⁸ and references therein).

Effects on Policies and Programs

Studies by demographers and economists of the forces that have driven advances in health have created a broad landscape upon which health policies have been constituted. They have demonstrated what is feasible in the world of action and not simply in the rarified world of randomized control trials. The documented successes have created an air of optimism in discussions of both domestic and international health, especially when the achievements are artfully cumulated as by Deaton.³⁷ And, they have helped to focus discussions on specific health programs and their cost-effectiveness and discouraged the more passive stance that health levels advance primarily as a byproduct of economic development. This re-orientation away from a "development first" approach was explicit in the World Bank's watershed 1993 World Development Report, Investing in Health.^b This study, written under the supervision of economist Dean Jamison, presented a massive array of evidence on global health problems and cost-effective ways of addressing them.

One constituent of the 1993 report was the first study of the "Global Burden of Disease." This study generated estimates of the burden of premature mortality and disability for 107 diseases and 483 sequelae. This research clarified disease burdens for regions and nations and has been cited over 4,000 times. Supervised by an economist/demographer/physician, Christopher Murray, it has been updated and expanded twice, most recently for 2010.⁵⁹ Among many other accomplishments, this series of studies has brought into bold relief the burden that HIV/AIDS and malaria presented in Africa. One consequence of this and other demographic/epidemiologic efforts was a massive expansion of international aid to combat these diseases. Foreign aid from the United States for health rose from $1.7 billion in FY2001 to $8.9 billion in FY2012. The Congressional Research Service⁶⁰ estimates that, from 2006 through 2011, more than 59 million insecticide-treated nets and 11 million malaria treatments were procured with financing from the United States. Malarial deaths declined by roughly 33 percent over that period, from 985,000 in 2000 to 655,000 in 2010.

Private patterns of philanthropy were also affected by the report.²² Bill Gates said in 2003, "I remember reading the 1993 World Development Report. Every page screamed out that human life was not being as valued in the world at large as it should be. My wife Melinda and I were stunned to learn that 11 million children die every year from preventable causes. That is when we decided to make improving health the focus of our philanthropy" (cited in Jamison et al.⁶¹). According to its annual report for 2012, the Gates Foundation spent $893 million on global health that year, including principal support for the Institute for Health Metrics and Evaluation.⁶²

Domestically, the set of studies described in this review may have their largest policy impact within the Social Security Administration. The U.S. Social Security System is required by law to be in "close actuarial balance" over a 75-year period. In simulations performed by Social Security actuaries, the actuarial balance is more sensitive to the future of longevity than it is to any other index except real wages. So projections of longevity have important fiscal implications, and the relevance of these implications is ensured by legislation.

These projections are made through extrapolations of past changes, with allowance for potential changes in the factors responsible for mortality improvement. There is substantial disagreement over whether the rate of progress observed in the past, as represented in age-specific death rates, can be sustained in the future. Demographers have pointed out how steady such progress has been⁵²,⁶³ and argued that the decline in smoking prevalence will aid future advances in life expectancy.⁶⁴ Actuaries from the Social Security Administration have often taken a more conservative stance, arguing that the elements that produced past gains in longevity cannot be expected to contribute to future improvements. As in most of the research reviewed in this chapter, the focus of discussion in technical advisory committees is on personal behaviors such as smoking, on medical systems, and on advances in standards of living. Literally trillions of dollars are at stake in the actuarial balance of the Social Security Trust Fund. Social scientists' interpretations of the factors that drive mortality change are playing a central role in planning for the Nation's fiscal future.

Acknowledgments

This chapter was supported by a grant from the National Institute on Aging [R01AG040212]. The content is solely the responsibility of the author and does not necessarily represent the official views of the Agency for Healthcare Research and Quality, the National Institute on Aging, the National Institutes of Health, or the U.S. Department of Health and Human Services.

Author's Affiliation

Samuel H. Preston, PhD, is Professor of Sociology in the Population Studies Center, University of Pennsylvania.

Address correspondence to: Samuel Preston, Population Studies Center, 289 McNeil Building, 3718 Locust Walk. University of Pennsylvania, Philadelphia, PA 19104; email speston@sas.upenn.edu.

References

1. Preston S. Human mortality throughout history and prehistory. In: Simon J. (Ed) The state of humanity, p. 30-6. Blackwell: London; 1995.
2. World population prospects: the 2012 revision. New York: United Nations, Department of Economic and Social Affairs, Population Division; 2013.
3. Nordhaus W. The health of nations: the contribution of improved health to living standards. In: Murphy K, Topel R. (Eds) Measuring the gains from economic research: an economic approach, p. 9-40. Chicago: University of Chicago Press; 2003.
4. Murphy KM, Topel RH. The value of health and longevity. J Political Econ 2006;114(5):871-904.
5. Costa D. Health and the economy in the United States from 1750 to the present. NBER working paper 19685. Cambridge, MA: National Bureau of Economic Research; 2014.
6. McKeown T. The modern rise of population. New York: Academic Press; 1976.
7. Szreter S. The importance of social intervention in Britain's mortality decline 1859-1940. Soc Hist Med 1988;1(1):1–37.
8. Woods R. The demography of Victorian England and Wales. Cambridge: Cambridge University Press; 2000.
9. Mercer A. Micro-organisms, disease and mortality. Rochester, NY: University of Rochester Press; Forthcoming.
10. Preston S, Haines M. Fatal years: child mortality in late nineteenth century America. Princeton, NJ: Princeton University Press; 1991.
11. Ewbank DC, Preston SH. Personal health behavior and the decline of infant and child mortality: The United States, 1900-1930. In: Caldwell J, Findley S, Caldwell P, et al (Eds),What we know about health transition, pp. 116-48. Health Transition Series. Canberra: Australian National University; 1990.
12. Vallin J, Mesle F. The segmented trend line of highest life expectancies. Popul Dev Rev 2009;35(1):159-87.
13. Cutler D, Miller G. The role of public health improvements in health advances: the twentieth-century United States. Demography 2005;42(1): 1-22.
14. Easterlin RA. How beneficent is the market? A look at the modern history of mortality. In: Easterlin RA (Ed), The reluctant economist, pp. 101-38. Cambridge: Cambridge University Press; 2004.
15. Cutler DM, Deaton A, Lleras-Muney A. The determinants of mortality. J Econ Perspect 2006;20(3):97-120.
16. Mokyr J. Why was there more work for mother? Knowledge and household behavior, 1870-1945. J Econ Hist 2000;60(1):1-41.
17. Mokyr J, Stein R. Science, health, and household technology: the effect of the Pasteur revolution on consumer demand. In: Bresnahan TF, Gordon RJ (Eds), The economics of new goods, pp. 143-206. Chicago: University of Chicago Press; 1996.
18. Newman P. Malaria control and population growth. J Develop Stud 1970;6:133-58.
19. Preston SH. Causes and consequences of mortality declines in less developed countries during the 20th century. In: Easterlin RA (Ed), Population and economic change in developing countries. Cambridge, MA: National Bureau of Economic Research; 1980.
20. Caldwell J. Old and new factors in health transitions. Health Transit Rev 1992;2(Suppl):205-15.
21. Preston S. The changing relation between mortality and level of economic development. Popul Stud 1975;29(2): 231–48.
22. Deaton A. Health in an age of globalization. Brookings Trade Forum; pp. 83-130. Washington, DC: Brookings Institution Press; 2004.
23. World development report 1993: investing in health. Washington, DC: World Bank; 1993.
24. Kenny C. There's more to life than money: exploring the levels/growth paradox in income and health. J Int Dev 2009;24-41.
25. Pritchett L, Viarengo M. Explaining the cross-national time series variation in life expectancy: income, women's education, shifts, and what else? Human Development Research Paper 2010/31. London: United Nations Development Programme; 2010.
26. Hum RJ, Prabbat J, McGahan A, et al. Global divergence in critical income for adult and childhood survival: analyses of mortality using Michaelis-Menten. Elife 2012;1:e00051. DOI: 10.7554/eLife.00051 1 of 17.
27. Acemoglu D, Johnson S. Disease and development: the effect of life expectancy on economic growth. J Polit Econ 2007;115(6):925-85.
28. Spence M, Lewis M (Eds). Health and growth. Washington, DC: Commission on Growth and Development, International Bank for Reconstruction and Development; 2010.
29. Acemoglu D, Johnson S. Disease and development: a reply to Bloom, Canning, and Fink. Working Paper 20064. Cambridge, MA: National Bureau of Economic Research; 2014. Available at http://www.nber.org/papers/w20064. Accessed September 4, 2014.
30. Pritchett L, Summers LH. Wealthier is healthier. J Hum Resour 1996;31:4:841–68.
31. Wilkinson R, Pickett K. The spirit level: why more equal societies almost always do better. London: Allen Lane; 2009.
32. Deaton A. Health, inequality, and economic development. J Econ Lit 2003;41:113–58.
33. Bloom D, Canning D. The Preston curve thirty years on: still sparking fires. Int J Epidemiol 2007;36:498–9.
34. Leon DA. Trends in European life expectancy: a salutary view. Int J Epidemiol 2011;40:271– 7.
35. Crimmins E, Preston S, Cohen B. Explaining divergent levels of longevity in high income countries. National Research Council, Panel on Divergent Trends in Longevity. Washington, DC: National Academies Press; 2011.
36. Woolf S, Aron L. U.S. health in international perspective: shorter lives, poorer health. Washington, DC: National Research Council; 2013.
37. Caldwell JC. Routes to low mortality in poor countries. Popul Dev Rev 1986;12(2):171–220.
38. Deaton A. The great escape: health, wealth, and the origins of inequality. Princeton, NJ: Princeton University Press; 2013.
39. Gakidou E, Cowling K, Lozano R, et al. Increased educational attainment and its effect on child mortality in 175 countries between 1970 and 2009: a systematic analysis. Lancet 2010;376:959–74.
40. Klemptner D, Jurges H, Reinhold S. Changes in compulsory schooling and the causal effect of education on health: evidence from Germany. J Health Econ 2011; 30(2): 340-54.
41. American Heart Association. Heart disease and stroke statistics–2013 update. Circulation 2013;127: e6-e245.
42. Glei D, Mesle F, Vallin J. Diverging trends in life expectancy at age 50: a look at causes of death. In: Crimmins E, Preston S, Cohen B (Eds), International differences in mortality at older ages: dimensions and sources. Washington, DC: National Research Council; 2011.
43. Cutler DM. Your money or your life. Oxford: Oxford University Press; 2004.
44. Ford E, Ajani UA, Croft JB, et al. Explaining the decrease in U.S. deaths from coronary disease, 1980–2000. N Engl J Med 2007;356:2388-98.
45. Mackenbach J, Looman C. Life expectancy and national income in Europe, 1900-2008: an update of Preston's analysis. Int J Epidemiol 2013;42:1100-10.
46. Kunst AE, Bos V, Lahelma E, et al. Trends in socioeconomic inequalities in self-assessed health in 10 European countries. Int J Epidemiol 2005;34(2):295- 305.
47. Glied S, Lleras-Muney A. Technological innovation and inequality in health. Demography 2008;45(3):741-61.
48. Goldman DP, Smith JP. The increasing value of education to health. Soc Sci Med 2011;72(10):1728-37.
49. Jha P, Peto R, Zatononski W, et al. Social inequalities in male mortality, and in male mortality from smoking: indirect estimation from national death rates in England and Wales, Poland, and North America. Lancet 2006;368:367-70.
50. Martikainen P, Makela P, Peltonen R, et al. Income differences in life expectancy: the changing contribution of harmful consumption of alcohol and smoking. Epidemiology 2014;25:182-90.
51. Fogel RW. Economic growth, population theory, and physiology: the bearing of long-term processes on the making of economic policy. Am Econ Rev 1994;84(3):369 -95.
52. Fogel RW. The escape from hunger and premature death, 1700–2100. Cambridge: Cambridge University Press; 2004.
53. Oeppen J, Vaupel JW. Demography: broken limits to life expectancy. Science 2002;296(5570):1029–31.
54. Preston SH, Wang H. Changing sex differentials in mortality in the United States: the role of cohort smoking patterns. Demography 2006;43 (4):413-34.
55. Case A, Lubotsky D, Paxson C. Economic status and health in childhood: the origins of the gradient. Am Econ Rev 2002;92(5):1308-34.
56. Finch CE, Crimmins EM. Inflammatory exposure and historical changes in human life-spans. Science 2004;305:1736-9.
57. Hobcraft J, Menken J, Preston S. Age, period, and cohort effects in demography: a review. Population Index 1982;48(1):4-43.
58. Barbi E, Vaupel J. Comment on: Inflammatory exposure and historical changes in human life-spans. Science 2005;308:1743a-4a.
59. The global burden of disease: generating evidence, guiding policy. Seattle, WA: Institute for Health Metrics and Evaluation; 2013.
60. U.S. global health assistance: background and issues for the 113th Congress. Washington, DC: Congressional Research Service; June 21, 2013.
61. Jamison DT, Summers LH, Allleyne G, et al. Global health 2035: a world converging within a generation. Lancet 2013;382(9908):1898-955.
62. 2012 Annual report: Bill and Melinda Gates Foundation. Avalable at http://www.gatesfoundation.org/~/media/GFO/Documents/Annual%20Reports/2012_Gates_Foundation_Annual_Report.pdf
63. Lee RD, Carter L. Modeling and forecasting the time series of U.S. mortality. J Am Stat Assoc 1992;87:659-71.
64. Bongaarts J. Trends in causes of death in low-mortality countries: implications for mortality projections. Popul Dev Rev 2014;40(2):189–212.

^a. It is important to recognize that, in the analyses of the effects of health on economic productivity, the value of health and longevity for their own sake is not included in the value of economic product. If they were to be included, as noted earlier, the impact of health improvements on economic well-being would be enormous.

^b. "Wealthier is Healthier," a 1996 paper by a senior economist and former President of the World Bank, Lant Pritchett and Larry Summers, was widely viewed as advocating a "development first" approach to health. As such, it was seen as a refutation or at least an amendment of the 1993 report. However, the authors were more careful than many readers: "By estimating income effects, we are not attempting to compare the efficiency of investment in overall income growth with investments in child mortality judged solely on the basis of the improvement in child mortality. Improved child mortality is not the only benefit of economic growth, so obviously investments specific to child health improvements are expected to be more 'cost effective' in producing health gains than economic growth."²⁹

Samuel Preston, PhD, is Professor of Sociology at the University of Pennsylvania, where he also has served as Dean of the School of Arts and Sciences, Director of the Population Studies Center, and several other leadership roles. His major area of interest has been the health of populations, including methods for measuring and analyzing health and mortality. He has devoted special attention to cause-of-death patterns across space and time, the impact of cigarette smoking on aggregate mortality, and the factors that have driven significant improvements in mortality over the past century. Dr. Preston's current research is focused on reducing the substantial uncertainty about the mortality risks associated with obesity.