Determinants of access to physician services in Italy: a latent class seemingly unrelated probit approach

We examine access to general practitioners, public and private specialists in Italy. We develop a novel model using finite mixtures of probit models that provides a rich and flexible functional form. The mixed distribution is flexible and can accommodate non-normality of response probabilities. The empirical analysis shows that patient behavior can be clustered in two latent classes, and that it changes according to the kind of physician service demanded and the latent class to which the individual belongs. We find that income strongly influences the mix of services. Richer individuals are less likely to seek care from GPs and more likely to seek care from specialists, and especially private specialists. Health status and societal vulnerability are the most important indicators of class membership.