Applied Analytics: Obesity, Heart Disease, and In$urance
“The rich are different from you and me,” Hemingway famously said, “They have more money.” They also have access to what can literally be the difference between life and death: health insurance. In this short analysis, I explore the relationships among obesity, heart disease, and income in the United States. The data suggest not only that low income influences obesity and heart disease, but also that it exacerbates the link between the two.
Based on data collected from 188 countries from 1980–2013, researchers estimate that a third of the world’s population is either obese or overweight. Of this, the highest proportion, almost 15%, live in the United States.(1)
Overweight is defined here as having a weight-to-height ratio of between 25 and 30, and obesity as having a ratio greater than 30. Obesity was formerly considered a problem of the socioeconomic elite in high-income countries, but it is now established that obesity’s burden is shifting towards lower-income groups.(2)
In 2015, the Centers for Disease Control (CDC) launched the 500 Cities Project to estimate health-related data for the 500 largest American cities, with at least one city from each state. These data are organized at the geographic scale at which the U.S. Census Bureau collects and maintains its survey information, the census tract. I have merged CDC and census data to investigate how income influences the link between obesity and heart disease, which according to CDC, is the leading cause of death for both men and women and for most ethnicities.
Figure 1 features a simple regression model of obesity versus heart disease. The numbers represent the percentage of respondents in a census tract who were measured as being obese versus those reporting they had been diagnosed by a health professional as having angina or coronary heart disease. To ensure the numbers conform to modeling assumptions regarding probability distributions, I transformed the original data by taking their logarithms.
The relationship is clearly and strongly positive. Generally, across all census tracts, the higher the percentage of obesity, the higher the percentage of heart disease. As straightforward as this statistical relationship appears, the actual biological pathway leading from obesity to heart disease is indirect. Obesity increases the risk of health factors such as high blood pressure, diabetes, inflammation, and insulin resistance. It is these other health factors that link directly to heart disease.(3)
Figure 2 breaks down the previous regression model by income levels. At the lowest income level, heart disease increases with increasing obesity at a constant rate. At mid- and high-income levels, heart disease also increases with increasing obesity, but at a diminishing rate, with the high-income group showing the greatest decreasing rate.
There may be many reasons why obesity’s influence on heart disease is exacerbated by low income.(4) I explore one possible reason in Figure 3. Figure 3(a) suggests that after an initial lag, after median household income reaches about $26K (the antilog of 10.17 on the graph, where the red, dashed line intersects the x-axis), access to insurance coverage increases steadily. And Figure 3(b) indicates that census tracts with a higher percentage of people with access to insurance tend to have a lower percentage of heart disease.
The converse is both true and tragic. People in households with low income generally have limited access to insurance coverage, and consequently, tend to bear a higher risk of heart disease.
Finally, Figure 4 shows the results of a cluster analysis, which groups the data into bins based on similarities in income, insurance coverage, obesity and heart disease. These attributes were mapped on to the two dimensions in the graph. The 500 cities in the CDC project break into four groups, where the cities in group 1 have the lowest income and insurance coverage, and the highest proportion of obesity and heart disease. Group 4 has the highest income and insurance coverage, and the lowest incidence of obesity and heart disease. The other two groups fall between these two extremes. Of the cities in group 1, the five with the highest proportion of heart disease are listed in the graphic.
The foregoing is a preliminary and cursory analysis of the data from CDC’s 500 Cities Project. In future analyses, I will focus on the health impacts of different types of insurance coverage. If you want to play around with your own analysis, email me for a copy of the raw data in csv format.
REFERENCES:
1. M. Ng et al., Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013. The lancet 384, 766–781 (2014).
2. C. A. Monteiro, E. C. Moura, W. L. Conde, B. M. Popkin, Socioeconomic status and obesity in adult populations of developing countries: a review. Bulletin of the World Health Organization 82, 940–946 (2004).
3. M. Bastien, P. Poirier, I. Lemieux, J.-P. Després, Overview of epidemiology and contribution of obesity to cardiovascular disease. Progress in cardiovascular diseases 56, 369–381 (2014).
4. G. K. Singh, M. Siahpush, R. E. Azuine, S. D. Williams, Increasing Area Deprivation and Socioeconomic Inequalities in Heart Disease, Stroke, and Cardiovascular Disease Mortality Among Working Age Populations, United States, 1969–2011. International Journal 3, 119–133 (2015).
