That income “causes” health is a well document phenomenon. People with higher means have the potential for longer longevity and do in fact live longer. Recent research published in the Journal of American Medical Association (JAMA) supports this trend. Using the most comprehensive U.S. data to-date (1.4 billion records) the authors not only conclude that income and overall standard of living are associated with life expectancy, but that “life expectancy increases continuously with income.” In fact, the gap in life expectancy between the top 1% percent and the bottom 1% of the income distribution stood at 15 years for men and 10 years for women.
Before we all sound the “inequality” alarm, it is important to note that the authors did not find any association between life expectancy for those at the lower end of the income scale to income inequality measure.
The study found a negative correlation between average life expectancy and Gini coefficient – lower life expectancy was associated with higher income inequality – “across commuting zones when pooling all income groups.” The results were similar at the state-level as well, suggesting that the latter is the result of the underlying concave relationship between income and health, where inequality resulting from larger share of individuals at the lower spectrum of the income scale leads to lower life expectancy overall. Correspondingly, life expectancy was more negatively associated with measures of income inequality for individuals at the higher end of the income scale. This is counterintuitive since income is said to have a predominant adverse effect on health at lower income levels not higher.
The importance of this study both in terms of scale and the potential for policy prescriptions cannot be overstated, but it lacks in various aspects. First, measures of life expectancy as were used, are always just that –inherently forecasting in nature and depend on measures used and the type of “adjustment” implemented. The risk is that once one adjusts for a variable there might be a confounding effect that invariably influence the measure adjusted. Such is the case in this research where the authors adjusted for race where the economic and social consequences of race largely influence the trajectory of life expectancy. Furthermore, predicting life expectancy at 40 years of age assumes “mortality rates remain unchanged as each individual ages. Yet in the United States today, midlife mortality rates among whites – especially poorly educated whites – are increasing, while mortality rates for older persons continue to decline” claims Angus Deaton in his written opinion to the paper. Deaton is referring to his recent publication.
Second, the authors’ conclusion regarding lack of association between longevity and rates of uninsured, is again adjusted for the mortality rates of minorities where they are precisely the ones who are most likely to be uninsured themselves. Lastly, the authors’ reliance on social capital as a measure used to challenge the association between health and social cohesion is less than optimal. If health is predicated upon the nature of social relationships in the civil space, how might social capital influence health? We can view social capital and/or social cohesion as having more of a social construct, while income inequality, alternatively, might be considered as a specific indicator of a more general social inequality. Despite large and vast research looking into the implication of lack of social capital as a proxy for income inequality and measures of health outcomes – the results are often contradictory. Many conclude that states with low social capital also had higher proportions of residents who reported their health as being only fair or poor. Similar results were found for those living in states characterized as having low and medium group membership and low and medium reciprocity. Yet some contradicted these results on the grounds that social cohesion/capital is a poor predicator of health.
What I find the most fascinating and of tremendous importance is the study’s inquiry into the potential effect of area characteristics on life expectancy. The paper, however, only examined trends for areas with more than 590,000 people, which is a somewhat limiting and excludes most rural areas. The implication of this selection bias could be that results may overstate gains in life expectancy for lower income population – and therefore understates the resulting gap – since people who live in the periphery often do not benefit from the same access to medical care, are often engaged in health worsening behavior such as smoking and drinking, and are less educated (pardon the generalization).
The authors found strong association between life expectancy and measures of smoking, obesity and exercise “suggesting that any theory for differences in life expectancy across areas must explain differences in health behaviors.” While this is intuitively true, in my own research, I had found little correlation between smoking, obesity rates, excessive drinking, or lack of exercise and household income at the county level (minimum county size examined was 65,000 persons), suggesting that the relations between county context and behavioral risks are considerably more nuanced. Alcohol-related behavior, for instance, may be associated with social norms through mechanisms that are unique to certain geographical areas or social groups.
The study’s immense progress towards understanding the importance of area effect and the understanding that individual traits alone cannot explain all the inequalities in health is of tremendous value. The need to account for the effect of area-based socioeconomic disparities that cause residential segregation and may result in spatial inequalities in resources is crucial.
The paper concludes that the top 1% will do well and be healthier no matter where they live. But for others, particularly the poor, location matters. In New York City where income inequality rates are one of the highest, life expectancy for the poor population is one of the highest as well, which serves to indicate that local-area health policies intervention have desired effects. In other parts of the country, particularly the South, adults with the lowest incomes have expected longevity average similar to people in much poorer countries such as Rwanda. Their life spans are also said to be getting shorter. Consequently, much like politics is local, we might want to look at health as being local. Instead of asking why American have poor health outcomes, we should ask why citizens of Tulsa and Las Vegas have such dismal results. Instead of contemplating interventions at the state level, we should perhaps examine what are the factors and policies that might improves the health in a community. Lessons can be shared, success and failures can be learned from and improved health can be enjoyed by all.
Categories: Means of Reproduction