A study of the population burden of cognitive impairment in the US

Abstract

This doctoral thesis investigates novel facets of cognitive health dynamics in the older US population. Prior research on partnership status and cognitive impairment has mainly focused on their association. No study has examined the burden of cognitive impairment without a partner, with partners being an essential source of caregiving, social support, and successful aging. Cognitive impairment often coexists with disability in activities of daily living. However, their cooccurrence (termed coimpairment), a highly disadvantageous health condition, has received less attention in population health research. Furthermore, the absence of a cure for cognitive impairment underscores the importance of identifying its modifiable risk factors. Using advanced multistate models and the longitudinal Health and Retirement Study, this thesis first shows that Blacks, Latinx, and lower-educated have longer expectancies in one of the most expensive health conditions, cognitive impairment, while being unpartnered. Second, these subgroups experience greater lifetime risk of, earlier onset of, and longer expectancy in coimpairment. Counterfactual exercises reveal that educational inequalities play an important role in these disparities. Third, the study shows that loneliness is an important modifiable risk factor for cognitive function and that the association is partially mediated by depression. This PhD thesis significantly advances our understanding of the socio-demographic inequalities in the burden of cognitive impairment in the US.