Abstract
Background:
Primary care physicians (PCPs) increasingly must identify which of their cognitively impaired patients who live alone are at greatest risk of harm due to self-neglect.
Objectives:
To determine whether brief patient self-report measures could accurately do this.
Methods:
Participants were ≥65 years, lived alone, and recruited from PCPs' practices, community agencies, a hospital emergency department, and acute care medical units. All had cognitive impairment (≤130 on the Dementia Rating Scale) and all had a PCP. Baseline self-report measures included: Geriatric Depression Scale (GDS), a social resources scale, a single item health rating scale, and the Quality of Life-Alzheimer's Disease Scale. We adjusted for baseline demographic, health, and mental status differences. We prospectively captured incidents of harm involving self-neglect or disorientation, resulting in physical injury, property loss, or damage, and requiring emergency services. These were obtained over a one-year longitudinal period, at 3-month intervals, from PCPs and caregivers. Emergency service records were obtained and reviewed for each incident. Proportional hazards modeling estimated how well the self-report measures predicted time to the first incident harm.
Results:
190 women and 34 men were followed. Based on the agreement of three medical raters, 23 participants (10%) experienced harmful outcomes. Being depressed on the GDS and rating one's health as fair or poor were the only two measures that significantly shortened time to first harmful outcome.
Conclusion:
GDS and self-rated health are simple measures to administer in the primary care setting and may be useful to PCPs in the earlier identification of those at greatest risk of harm in this vulnerable group of patients.