Health is largely determined by factors situated outside the health care delivery system, said keynote speaker, David Williams, Florence and Laura Norman Professor of Public Health and professor of African and African American studies and of sociology at Harvard University. The health care system generally functions to provide care to those who have become sick. Yet it is where people live, learn, work, play, and worship that most influences their opportunities and chances for being healthy. Social policies can make it easier or harder for people to make healthy choices, Williams said.
It has long been recognized, Williams stated, that individual-level health interventions can be complemented by changes to the social, physical, and economic environments that determine health and risk factors for health. These “upstream interventions” can result in improved health without any conscious awareness or participation by individuals (Katz, 2009). Over the past century, dramatic improvements in health have been achieved through upstream interventions such as improved sanitation, improved working conditions and equipment safety, seatbelts in automobiles, road safety laws, initiatives to reduce drunk driving, elimination of lead in paint and gasoline, and the fluoridation of water.
Geographic location determines exposure to risk factors and resources that affect health, Williams said. Safety, cleanliness of the air and environ-
ment, parks and places for physical activity, the quality of housing, the upkeep of streets and homes, access to nutritious foods, access to high-quality medical care, and accessible safe modes of transportation all affect whether a community is a healthy or unhealthy place to live (RWJF, 2009). Williams described several historical and current examples of the health effects of where and how people live.
Residential segregation, the physical separation of the races by enforced residence in different areas, is a powerful example of the impact of social policy on health, Williams said. The implementation and influence of segregation laws spanned housing policies; the judicial system; neighborhood organizations; and major social, cultural, and economic institutions. Between 1860 and 1940, the extent of black-white segregation in both northern and southern cities increased dramatically and has remained relatively stable (Cell, 1982; Lieberson, 1980; Massey and Denton, 1993). Racial differences in socioeconomic status are heavily driven by segregation, as segregation determines access to educational and employment opportunities.
Segregation in the United States has created places that are markedly and distinctly unhealthy. Research suggests that the conditions linked to segregation constrain the practice of health behaviors, and promote unhealthy behaviors, and that segregation affects access to high quality medical care (Williams and Collins, 2001). Williams cited the work of Cutler and colleagues (1997) who concluded that the elimination of residential segregation would completely eliminate black-white differences in income, education, and unemployment and would reduce racial differences in single motherhood by two-thirds.
Williams described several examples of initiatives to address the conditions linked to place. In the 1990s, the U.S. Department of Housing and Urban Development (HUD) launched three major programs: Moving to Opportunity helped poor families move from high poverty public housing to better residential conditions; Jobs Plus sought to saturate public housing with high quality employment services and rent-based financial work incentives; and Bridges to Work was designed to link central city residents to suburban employment opportunities. An analysis of the HUD initiatives by the Urban Institute concluded that such interventions can increase income, improve safety and security, and improve physical and mental health (Turner and Rawlings, 2005). Moreover, although families
do respond to real opportunities, meaningful change requires sustained effort over time because people need help not only in finding jobs, but in keeping jobs (challenges include retention, advancement, commuting costs, and child care). Programs that work best are those that address multiple barriers (e.g., housing, safety, health, employment, education). Other scientific assessments of Moving to Opportunity found that after 3 years, parents and their sons1 who had moved from high-poverty to low-poverty New York City neighborhoods were doing better in terms of mental health (Leventhal and Brooks-Gunn, 2003), and 10 to 15 years later, there were reductions in the prevalence of severe obesity and diabetes in adults (Ludwig et al., 2011). Moving to Opportunity is a dramatic example, Williams said, of a non-health, place-based intervention that has had long-term dramatic and positive effects on health.
Another example highlighted by Williams is the Yonkers Housing Intervention, for which half of the public housing residents were randomly selected via a lottery to move to better housing in middle-class neighborhoods. After 2 years, those who had moved reported better overall health, less substance abuse, less neighborhood disorder and violence, lower use of welfare, and higher rates of employment than those who had not moved. They also reported greater satisfaction with public transportation, recreation facilities, and medical care (Fauth et al., 2004).
Although these examples show that simply changing the neighborhood environment, with no health intervention, can lead to improvement in health, this is not a good national model for how to solve the problem, Williams said, adding that one should not have to move to live in a better neighborhood. Additionally, Williams pointed out that there is nothing inherently wrong with living next to someone of your own race. The problem of segregation is not segregation per se, he continued, but what sociologist William Julius Wilson has described as the concentration effects of social ills that co-occur with segregation. There is a policy opportunity for a major infusion of economic capital to improve the social, physical, and economic infrastructure of the disadvantaged communities that could have huge spillover effects on health, Williams said.
Evidence suggests that social and economic policies intended to provide additional income to low-income individuals can also provide dramatic improvements to health, noted Williams. One study he cited, which
1 This outcome was found only in male children.
was conducted as part of the negative income tax experiments of the 1970s, showed that low-income mothers who received expanded income support had better birth outcomes (measured as infant birth weight) than women in the control group. It was thought that the additional income allowed for improved nutrition (Kehrer and Wolin, 1979). Other studies have associated income from the Earned Income Tax Credit with reduced rates of low birth weight and increased mean birth weight (Hoynes et al., 2012; Strully et al., 2010). Implementation of the Social Security program dramatically reduced poverty levels among America’s elderly and, research shows, led to mortality declines. Over the past several decades, when Social Security benefits have been increased, a commensurate decline in mortality among the elderly has been observed (Arno et al., 2011).
Another example shared by Williams was the Great Smokey Mountain Study in North Carolina, which assessed the impact of additional family income on the mental health over time of children aged 9 to 13 years at the start of the study. The opening of a casino on an Indian reservation provided additional income to American Indian residents of the community, but not to non-Indian residents. This natural experiment documented a decline in rates of deviant and aggressive behavior among adolescents whose families received additional income and showed that reduced rates of psychopathology persisted into adulthood when the adolescents had moved out of their parental home (Costello et al., 2003, 2010). Additional income for American Indian families who were poor at the time of the increase in income was also associated with higher levels of education, lower incidence of minor criminal offenses, and the elimination of the American Indian-white disparity on these outcomes. The authors postulated that the mechanism was improved parental behavior (Copeland and Costello, 2010).
The conditional cash transfer (CCT) programs are another source of evidence of the impact of increased income on health. These programs provide cash payments to low-income families contingent upon regular health care visits, school attendance, or participation in educational programs, Williams explained. Analysis of the first large scale CCT program in Mexico showed that increased income led to reduced rates of child illness and stunted growth, increased quality of prenatal care, and reduced rural infant mortality (Barber and Gertler, 2009; Barham, 2011; Rawlings and Rubio, 2005). Although the CCT programs included conditions for receiving the payments, Williams said that research shows that it was the additional income, not the conditions, that lead to the improvement in health. However, most experts believe that the conditions are important to ensure political support for these programs.
In the past 60 years, differences in health (measured as life expectancy
and mortality in national data) between African Americans and whites have narrowed and widened in tandem with racial differences in income, Williams said. The only time in recent history when the health of African Americans improved more rapidly than the health of whites was between the late 1960s and the late 1970s, when civil rights and anti-poverty policies narrowed the income gap between African Americans and whites (Cooper et al., 1981; Kaplan et al., 2008). Multiple studies demonstrate improvement in the health of African Americans in association with civil rights policy, for example, reduction in infant mortality and better birth outcomes associated with desegregation of southern hospitals (Almond and Chay, 2006). In the 1980s, the economic gap between African Americans and whites in the United States widened again, and health outcomes declined for African Americans. For example, Williams presented data from the National Center for Health Statistics that showed life expectancy at birth for African Americans declined from the 1984 level for 5 years in a row, while life expectancy for whites increased slightly during this same period.
Family structure can have significant impact on socioeconomic status and health. Compared to children raised by two parents, for example, children raised by a single parent (usually a female head of household) are more likely to grow up poor, drop out of high school, be unemployed in young adulthood, and not enroll in college. In addition, they have an elevated risk of juvenile delinquency and participation in violent crime, including homicide (McLanahan and Sandefur, 1994; Sampson, 1987). The sources of violent crime are rooted in social policy and related to structural differences in economic and family organization in communities. This harkens back to the role of segregation, Williams said, and he quoted Sampson and Wilson (1995, p. 41) who said that in the 171 largest cities in the United States, “the worst urban context in which whites reside is considerably better than the average context of black communities.”
Social policy can counter some of the negative effects of family structure on poverty and child health outcomes, Williams said. A 2000 report by the United Nations Children’s Fund (UNICEF) found, for example, that although the proportion of children in single-parent households in the United States and Sweden was comparable at 19 and 21 percent, respectively, the child poverty rate in single-parent households was 55 percent among Americans, but only 7 percent among Swedes. The overall child poverty rate based on income before taxes was comparable in the United States and Sweden, at 26.7 and 23.4 percent, respectively, however, based on income after taxes, the child poverty rate in Sweden
was 2.6 percent, while for the United States it was 22.4 percent. Williams also indicated that although there is a big emphasis in some policy circles on “cultural values” as drivers of group variations in family structure, there is impressive empirical evidence that indicates that marriage rates are associated with economic opportunities for males. An example is the research on the impact of military enlistment on the economic and family structure of households. Research shows that inactive duty military service promotes marriage over cohabitation, increased likelihood of first marriage, and greater stability of marriage (Teachman, 2007, 2009; Teachman and Tedrow, 2008). These patterns exist for both white males and African American males who enroll in the military, but the effects are stronger for African Americans. There is less economic discrimination in the military, Williams said, and more opportunities for advancement of minorities than in the civilian labor force, as well as generous military benefits to support a family. In this case, creating economic opportunity for males has a significant effect on family structure, which can have long-term positive effects on health, Williams said.
It has been suggested that reducing the black-white academic test score gap would have a significant impact on reducing racial inequality in earnings and, in turn, would help to reduce racial differences in health, family structure, and crime (Jencks and Phillips, 1998). Williams cited several examples of relatively simple psychosocial interventions that significantly reduced academic disparities. African American seventh graders who participated in a “self-affirmation intervention” (a writing exercise designed to affirm a sense of adequacy and self-worth) showed significantly improved grades across all of their academic work, compared to a control group. The intervention reduced the racial achievement gap by 40 percent after one administration (one class period), and the effect was still evident after 2 years (Cohen et al., 2006, 2009). A similar study of a “values affirmation exercise” with college women in the science, technology, engineering, and mathematics (or STEM) disciplines found similar effects, and reduced the male-female grade gap (Miyake et al., 2010). A “social belonging intervention” for minority college freshman sought to provide them with a sense of belonging and to help them to realize that all new students face difficulties in their college experience but that these challenges are temporary. Compared to a control group, students who participated in this intervention in their freshman year showed increased academic performance over their 4 years of college, and the black-white achievement gap was reduced by half. In addition, they had fewer doctor visits than the control group during their college years and improved self-
reported health (Walton and Cohen, 2011). Investments in early childhood education for low-income and minority children also have been shown to have decisive health benefits into adulthood (Reynolds et al., 2007).
Williams’ Closing Thoughts
Health care improvements alone will not solve America’s health problems, Williams stated. All policy that affects health is health policy, he said, and socioeconomic and racial or ethnic inequalities in health reflect the effective implementation of social policies that have reinforced such disparities. Eliminating disparities requires political will and a commitment to new strategies to improve living and working conditions, create opportunities to promote health for all, and remove barriers that make it nearly impossible for some Americans to make healthy choices. This will require collaboration across multiple sectors to build the science base to support the development of social policies that can produce systematic and comprehensive change. Williams concluded by highlighting the keys to long-term success as
- building the perspective of health into all policy making;
- including an explicit focus on health equity in policy making;
- convening, enabling, and supporting cross-sectoral collaborations;
- developing consensus-based standard data and methods for surveillance systems linking health, health equity, and the determinants of health; and
- investing in strengthening community capacity and potential for community advocacy.
During the brief discussion that followed the presentation, Williams responded to a question about the opportunity costs the interventions described. From a cost-benefit perspective, he said, it is in the long-term financial interest of society to implement these types of interventions. Americans have to realize that the status quo—gaps in health—is costing money. It is estimated that racial disparities in health cost the U.S. economy more than $300 million annually and socioeconomic inequalities cost the economy more than $1 trillion each year. There is also evidence that many of the interventions described may save money, but there needs to be a more systematic effort to build the business case for improving health, he said. Social policies affect health, but there is not yet a sufficient empirical base to indicate clearly which policies should be undertaken first and to inform the optimal timing and sequencing of policies.
In response to a question about engaging policy makers to effect systemic change, Williams said that the United States faces an educational and communication challenge. Most Americans do not think of the causes of poor health (i.e., the social determinants of health) as systemic. Our culture focuses more on the perceived contribution of individual values, choices, and behaviors in shaping individual health outcomes. Most leading policy makers have never considered the social determinants of health, he said. For them, health policy relates to coverage and perhaps quality and cultural sensitivity, rather than to factors outside the health care system.