A microeconometric analysis of Canadian health care utilization

Understanding health care utilization is important to design efficient and effective health systems. Toward this end, we develop a relatively simple and intuitively appealing microeconometric framework to analyse health care utilization and illustrate its use with recent Canadian microdata. We find that health care utilization consists of distinct stochastic processes requiring the use of two stochastic regression models. In particular, the latent class modelling framework is the superior statistical framework if the data permit modelling unobserved heterogeneity and overdispersion. In many instances, results differ for the classes of high and low users of health care services.