To determine the prevalence and correlates of 4 types of elder abuse and neglect in a geriatric psychiatry service.
We conducted a cross-sectional retrospective chart review of new in- and outpatients seen by the Montreal General Hospital Division of Geriatric Psychiatry in one calendar year.
Abuse or neglect was suspected or confirmed in 20 (16%) of 126 patients, comprising financial abuse in 16 (13%), neglect in 7 (6%), emotional abuse in 5 (4%), physical abuse in 3 (2%), and multiple abuse in 7 (6%). On bivariate analysis, patients living with nonspouse family, friends, or other persons were significantly more likely to have suffered abuse than were those living with their spouse or in a supervised setting (OR 10.5; 95%CI, 2.3 to 47.8); widowed, divorced, or separated patients were significantly more likely to have suffered abuse than were married patients (OR 4.7; 95%CI, 1.02 to 22.0). Nonsignificant trends included female sex (OR 4.1; 95%CI, 0.89 to 18.6); alcohol abuse (OR 2.1; 95%CI, 0.71 to 6.2); behaviour problems (OR 1.9; 95%CI, 0.71 to 5.2); and chronic cognitive impairment (OR 1.4; 95%CI, 0.55 to 3.8). Although living situation with nonspouse family, friends, or others and marital status of widowed, divorced, or separated were significantly associated with abuse when examined in separate logistic regression models, both were nonsignificant when examined together, suggesting collinearity. Both were retained in the model because they probably represent different aspects of vulnerability. The final model included living situation with nonspouse family, friends, or others (OR 6.1; 95%CI, 0.75 to 49.5) and widowed, divorced, or separated marital status (OR 2.4; 95%CI, 0.21 to 26.8). Nonsignificant trends included female sex (OR 2.6; 95%CI, 0.45 to 14.4); alcohol abuse (OR 2.2; 95%CI, 0.59 to 7.9); and lowest quartile on the Global Assessment of Functioning (GAF) scale (GAF or = 2.0) included female sex, alcohol abuse, and lowest quartile of functional status. Study limitations include a cross-sectional retrospective chart review design, a clinically derived sample, a small sample size, and a lack of structured instruments for several variables.