Cognitive aging and rate of hospitalization in an urban population of older people

Abstract

Background: The association of age-related cognitive change with hospitalization is not well understood.

Methods: At 3-year intervals for a mean of 8.7 years, 2,273 older residents of a geographically defined urban community underwent cognitive testing from which a global measure was derived. Hospitalization data were obtained from Part A Medicare beneficiary records. The association of level of cognitive function and rate of cognitive decline in each 3-year interval with subsequent rate of hospitalization was assessed using mixed-effects count regression models.

Results: There were 9,091 hospitalizations involving 1,810 of the 2,273 individuals in the cohort (79.6%). Rate of hospitalization increased by 9.7% (95% confidence interval [CI]: 7.2, 12.3) with each additional study year; by 32.7% (95% CI: 26.8, 38.0) for each 1 point lower on the global cognitive measure at the beginning of an observation interval; and by 24.3% (95% CI: 16.6, 32.6) for each 1-point decrease in the global cognitive measure during the previous observation period. These associations persisted after adjustment for comorbidities and exclusion of those with a Mini-Mental State Examination score less than 26.

Conclusion: Individual differences in trajectories of cognitive aging are associated with subsequent risk of hospitalization.

Keywords: Cognitive aging; Epidemiology; Hospital related; Public health..

Figure 1.

Predicted rates of hospitalization in participants with moderately low (25th percentile, dotted line) or high (75th percentile, solid line) prior cognitive level (left side) and moderately fast (25th percentile, dotted line) or slow (75th percentile, solid line) prior cognitive decline (right side), based on a mixed-effects count regression model adjusted for age at baseline, sex, race, and years of education.

Figure 2.

Predicted rates of hospitalization in participants with moderately low (25th percentile, dotted line) or high (75th percentile, solid line) prior cognitive level (left side) and moderately fast (25th percentile, dotted line) or slow (75th percentile, solid line) prior cognitive decline, based on three separate mixed-effects count regression models adjusted for age at baseline, sex, race, and years of education, using an executive function measure (upper panel), an episodic memory measure (middle panel), and the Mini-Mental State Examination (lower panel).