Date of Award

May 2019

Degree Type


Degree Name

Doctor of Philosophy



First Advisor

Changshan Wu

Committee Members

Kirsten Beyer, Alison Donnelly, Woonsup Choi, Kevin Matthews


dementia, disparities, mortality


Mainly due to the expansion of the geriatric population, the number of deaths attributable to dementia in the United States is likely to increase rapidly in the coming decades. In the hope of offering some valuable insights into discovering disease risk factors linked to geography, advocating place-based prevention and intervention strategies and supporting equitable access to end-of-life care, this study examines three aspects of socio-spatial disparities in dementia mortality in the U.S.: (i) small area spatial and temporal variations in dementia mortality risk; (ii) “place effects” on the differentials in individual dementia mortality risk; and (iii) disparities in place of death of decedents from dementia across populations and places.

Results from the study indicate that, first, there are substantial spatial and temporal variations in dementia mortality risk in the U.S. Specifically, regions including Pacific Northwest, Ohio River Valley and the Carolinas were found to be the most likely high-risk clusters; while counties in the Northeast and Florida were the most likely low-risk clusters. Temporal information of clusters suggested a reduction in the relative risk of Alzheimer’s disease and all-cause dementia mortality in most of the highly likely clusters. The results may provide etiologic clues linked to geography and time and propel public health agencies to evaluate the capacity of local health and social care to meet dementia patients’ needs before death, especially in those high-risk cluster areas.

Second, differences in individual dementia mortality risk were in part due to the “place effects.” Among the three environmental variables examined, area socioeconomic deprivation and PM2.5 concentration were significantly associated with dementia mortality risk; area social integration did not have a significant relationship in models adjusted for individual-level factors. Although the relationship between area socioeconomic deprivation was nonlinear, the association between PM2.5 concentration and individual dementia mortality risk revealed a dose-response relationship. The relationship between the three environmental factors and dementia mortality risk also differed by age group. The results suggested that environmental interventions, especially improving local air quality, might be an effective measure to reduce dementia mortality risk.

Third, results showed a persistent shift from deaths at institutional settings (hospitals and nursing home/long-term care facilities) to deaths at home and other places among decedents from dementia during 2000 and 2014. There were wide interstate variations in place of death of decedents from dementia. In addition to socio-demographic characteristics of the decedents, state-level factors including access to care facility resources and Medicare and Medicaid expenditure on long-term care might have contributed to the changing landscape.

Although biomedical approaches still dominate in our efforts to understand the diseases, this dissertation research examined the effects of factors “outside the body” on socio-spatial disparities in dementia mortality. The findings provide new insights to inform future epidemiological and health services research related to dementia.

Included in

Geography Commons