OBJECTIVES: To examine the longitudinal association between decline in cognitive function and risk of elder self‐neglect in a community‐dwelling population.

DESIGN: Prospective population‐based study.

SETTING: Geographically defined community in Chicago.

PARTICIPANTS: Community‐dwelling subjects reported to the social services agency from 1993 to 2005 for self‐neglect who also participated in the Chicago Health Aging Project (CHAP). Of the 5,519 participants in CHAP, 1,017 were reported to social services agency for suspected elder self‐neglect from 1993 to 2005.

MEASUREMENTS: Social services agency identified reported elder self‐neglect. The primary predictor was decline in cognitive function assessed using the Mini‐Mental State Examination (MMSE), the Symbol Digit Modalities Test (Executive Function), and immediate and delayed recall of the East Boston Memory Test (Episodic Memory). An index of global cognitive function scores was derived by averaging z‐scores of all tests. Outcome of interest was elder self‐neglect. Logistic and linear regression models were used to assess these longitudinal associations.

RESULTS: After adjusting for potential confounding factors, decline in global cognitive function, MMSE score, and episodic memory were not independently associated with greater risk of reported and confirmed elder self‐neglect. Decline in executive function was associated with greater risk of reported and confirmed elder self‐neglect. Decline in global cognitive function was associated with greater risk of greater self‐neglect severity (parameter estimate=0.76, standard error=0.31, P=.01).

CONCLUSION: Decline in executive function was associated with risk of reported and confirmed elder self‐neglect. Decline in global cognitive function was associated with risk of greater self‐neglect severity.