Multivariate Modeling of Anxiety and Depression in Community-Dwelling Elderly Persons

Abstract

A sample of community-dwelling geriatric primary care patients were given survey measuring which demographic, medical, and environmental factors predict higher levels of state anxiety, trait anxiety, depression, and comorbid anxiety-depression. The authors used multivariate linear regression methods to predict scores for the outcome measures. Younger age, more stressful life events, increasing medical comorbidity, and fewer years of formal education predicted higher levels of state anxiety. Lower perceived health status, and comparative health status, and benzodiazepine use predicted higher trait anxiety scores. Older age, male gender, benzodiazepine use, more stressful life events, worse comparative health status, and less social support predicted higher depression scores. More stressful life events, less social support, poorer perceived general health status, and benzodiazepine use predicted higher comorbid anxiety-depression scores. The authors discuss the utility of developing clinical models to predict anxiety and depression in geriatric patients. Anxiety and depressive symptoms are distinguished by different psychosocial, clinical, and environmental factors, even in a sample of older subjects, who, by all measures, have good physical and psychological health.