Re-revisiting the behavioral model of health care utilization by Andersen: A review on theoretical advances and perspectives

Abstract

The Behavioral Model of Health Services Use by Ronald M. Andersen and colleagues is the most widely adopted theoretical framework for analyzing and predicting health care utilization. Among other things, it is employed in the German Federal Health Reporting since 2001. It differentiates need factors, predisposing factors and enabling factors both on the contextual level and the individual level as determinants of individual health services use. From the viewpoint of social epidemiology, one of the key strengths of the Behavioral Model is its capability to systematize and empiricize equity and inequity in the access to health services by specifying need vs. predisposing and enabling factors. This strength could be even promoted by including direct effects on utilization of psychological factors (besides social factors) as contributing to inequity. Another strength of the Behavioral Model since its fifth version is that it conceptualizes need factors, predisposing factors and enabling factors both on the contextual level and the individual level in a structurally equivalent manner. Thus, not only are theoretically consistent multilevel models possible on the predictor side, but general theories of action and behavior from sociology and psychology are more easily applied on the behavior of professionals working in health policy and services. On the side of health-related behaviors (as mediating factors) and the outcomes of the model (including, since its sixth version, quality of life), the question is why these are represented as individual entities only, thus excluding relevant prevalences and incidences only from the scope of the model. Here, the Behavioral Model could be further developed by integrating assumptions of the Basic Behavioral Epidemology Model by Thomas von Lengerke and colleagues which - following the micro-micro-model of sociological explanation - allows the description and explanation of collective outcomes. Finally, regarding the empirical and statistical application of the Behavioral Model, improvements are possible by use of the methodological differentiation between mediation and moderation. For instance, it should be clarified whether only enabling factors, which according to classical social ecology tend to moderate associations between other variables (in the present case need factors and utilization), may be conceptualized and modelled as effect modifiers, or predisposing factors as well. In the context of data analyses oriented by the Behavioral Model, this would have direct implications for specifying hierarchical models and relevant interaction terms. © Springer Science+Business Media New York 2014.