Discretion and service use among older adults: the behavioral model revisited

Research applying the behavioral model to the use of services among older adults could be enriched by the classification of services along a discretionary dimension. Survey responses from 2,178 community-dwelling older adults were used to test the hypothesis that predisposing, enabling, and need characteristics are better predictors of discretionary than nondiscretionary service use. Logistic regression results predicting the use of a variety of community-based and medical services categorized as most discretionary, partially discretionary, and least discretionary generally support our hypothesis, underscoring the importance of need characteristics for least discretionary service use.