Three Essays on Diabetes and Mortality in the United States

Diabetes is on the rise. The world faces an emerging deadly threat on people’s health and nations’ public-health systems with a large economic burden. According to the World Health Organization (WHO), the number of people living with diabetes has risen from 108 million in 1980 to 422 million in 2014 (WHO, 2016). In the United States, more than 80 million people are living with diabetes or prediabetes and this number is projected to double by 2050 (CDC, 2017). Diabetes is a serious and life-threatening condition that occurs when the body cannot regulate blood sugar. When diabetes is not appropriately controlled, it causes dire consequences for health and well-being of individuals. Furthermore, compelling empirical evidence has shown a strong association between diabetes and premature death (Gu, Cowie and Harris, 1998). In 2016, the number of deaths in the world due to diabetes was 1.6 million and this number is expected to increase (WHO, 2016). Thus, reducing excess deaths from diabetes has been one of the main goals of global efforts, such as 2030 Agenda for Sustainable Development developed by the WHO (UN General Assembly, 2015). Accordingly, public health policy makers and health providers have called for better diabetes prevention interventions and disease management programs designed to reduce the risk of mortality due to diabetes. This dissertation aims to contribute to these efforts by filling critical gaps in the previous literature on diabetes and mortality. The gaps this dissertation focuses on are threefold. First, the fraction of total deaths attributable to diabetes is seriously underestimated when an underlying cause of death on death certificates is used to assess the contribution of diabetes to overall mortality, which is a commonly used method to assess the contribution of diabetes to a national mortality profile. Second, despite a body of literature on the association between diabetes and mortality, few prospective studies have examined this association by using information on self-reported diagnosed diabetes and biomarkers of diabetes measured at the start of the mortality follow-up (or baseline) simultaneously. Third, prior findings on nativity disparities in diabetes in the United States have been inconsistent and most studies have relied solely on self-reported diabetes, which may be biased due to immigrants’ limited access to healthcare. Each of these three gaps is addressed in the following chapters. In Chapter 1, I examine the contribution of diabetes to US life expectancy using a Population Attributable Fraction (PAF) approach. I found that life expectancy at age 30 increases by 0.89 years when diabetes is eliminated based on the PAF, and this estimate is much higher than a gain of 0.34 years based on underlying cause of death data. In Chapter 2, I estimate the risk of death associated with diabetes using both self-reported diagnosed diabetes and a level of HbA1c measured at baseline. Results show that self-reported diagnosed diabetes had a stronger association with mortality than HbA1c at baseline, and individuals with self-reported diagnosed diabetes were at a higher risk of death compared to those whose diabetes had not been previously diagnosed, regardless of their diabetes status identified by baseline HbA1c level. In Chapter 3, I examine nativity differentials in diabetes using self-reported diagnosed diabetes, HbA1c and undiagnosed diabetes. I find that the foreign-born have a higher prevalence of diabetes than the US-born individuals, but that this excess was explained by controls for sociodemographic characteristics and Body Mass Index (BMI). In addition, I observe that among the foreign-born, those who had arrived in the United States in the last 15 years were less likely to have self-reported diabetes than those who had arrived earlier, suggesting that the immigrant health advantage deteriorates with time in the United States. When diabetes was diagnosed based on baseline HbA1c level, however, there was no such advantage for the short-term immigrants over the long-term immigrants. These findings suggest avenues for future research. First, it is important to consider the various ways in which information on diabetes is collected in surveys and on death certificates in assessing its impact on mortality. The results from Chapter 1 show that a focus on diabetes as an underlying cause of death seriously underestimates its impact on life expectancy. Furthermore, Chapter 2 demonstrates that self-reported diagnosis of diabetes provides important information about the risk of death due to diabetes beyond biomarkers measured at the start of the mortality follow-up. Thus, future studies need to consider multiple ways in which diabetes is measured. Second, in Chapters 2 and 3, the issue of measurement is a key focus and results from these chapters suggest that diverse measurements of diabetes should be considered at the same time, not only because they reflect different aspects of diabetes, but also using them together provides more comprehensive information on how diabetes is associated with mortality. Third and most importantly, future research should seek to understand the mechanisms through which diabetes contributes to mortality. Because all three chapters of this dissertation are based on cross-sectional data, the current dissertation cannot directly examine the connections that link the onset of diabetes to subsequent mortality. Patients who are diagnosed with diabetes are likely to experience several changes in their life, such as modifications in health behaviors, using healthcare services and taking medicine to control diabetes, and change the perception of their own health, which are likely to affect the risk of death. Therefore, future research that accounts for these modifications will shed light on the mechanisms associated with the onset of diabetes and subsequent mortality