Family context as a predictor of individual medicine use

This analysis examines individual use of prescribed and nonprescribed medicines by placing the individual in his family context and shows that the set of family context variables is a better predictor of individual medicine use than theSet of individual characteristics.

Abstract

The family unit is being increasingly recognized as a key determinant of individuals' health and illness behaviors. Medicine taking is the most frequent of all preventive and illness-related behaviors. To date, empirical research on medicine use has focused on determining the socio-demographic and morbidity characteristics associated with individuals' medicine use and simple descriptive statistics have usually been used to characterize such behaviors. This analysis examines individual use of prescribed and nonprescribed medicines by placing the individual in his family context. Data on more than 2700 individuals living in almost 800 families were analyzed using hierarchical and step-wise multivariate regression. Variables measured at the family level included: family size, family income, per cent of females and per cent of children in the family, average family morbidity, and average others' nonprescribed and prescribed medicine use. Variables measured at the individual level included:, age, sex, race and morbidity. The hierarchical regressions showed that the set of family context variables is a better predictor of individual medicine use than the set of individual characteristics; other family members' medicine use behaviors are strong predictors of individuals' medicine use, prescribed being a better predictor of nonprescribed use and vice versa; the relative importance of the independent variables depends on whether prescribed or nonprescribed medicine use is being predicted; in the step-wise regressions some family context variables (most notably others' medicine use) were more powerful predictors than individual variables; the model is a better predictor of prescribed medicine use than of nonprescribed medicine use; some variables, such as family size, which have been shown previously to be significantly related to medicine use are not important in the presence of other variables in a multivariate analysis.