The dynamic interdependence in the demand of primary and emergency secondary care: A hidden Markov approach

Mauro Laudicella,Paolo Li Donni

doi:10.1002/jae.2882

Abstract

SummaryThis paper develops an extension of the class of finite mixture models for longitudinal count data to the bivariate case by using a hidden Markov chain approach. The model allows for disentangling unobservable time‐varying heterogeneity from the dynamic effect of utilisation of primary and secondary care and measuring their potential substitution effect. Three points of supports adequately describe the distribution of the latent states suggesting the existence of three profiles of low, medium and high users who shows persistency in their behaviour, but not permanence as some switch to their neighbour's profile.

Highlights

The demand for emergency secondary care (ESC) is growing fast in many publicly funded health care systems absorbing a large share of their resources and jeopardising their financial sustainability (Berchet, 2015)
The model extends the class of finite mixture models for longitudinal count data (Bago d’Uva, 2005; Deb and Trivedi, 1997) to the bivariate case, in which the number of primary care (PC) and ESC visits are jointly modelled and depend on a latent process following a first-order Markov chain
The first model consists of two separate Poisson equations with random effects (REPo) that allow for individual time-fixed heterogeneity by using a dynamic panel specification with normally distributed random effects, but model PC and ESC utilisation as two independent processes

Summary

Introduction

The demand for emergency secondary care (ESC) is growing fast in many publicly funded health care systems absorbing a large share of their resources and jeopardising their financial sustainability (Berchet, 2015). Unobserved heterogeneity in the individual health status and preferences may affect the demand of both PC and ESC producing an additional source of endogeneity To deal with these issues, existing empirical applications make use of the exogenous variation in the supply of PC induced by a policy reform over time and across geographical areas. The model extends the class of finite mixture models for longitudinal count data (Bago d’Uva, 2005; Deb and Trivedi, 1997) to the bivariate case, in which the number of PC and ESC visits are jointly modelled and depend on a latent process following a first-order Markov chain This feature allows the researcher to disentangle the contribution of past utilisation and time-varying unobserved heterogeneity on present utilisation.

Preliminaries

Proposed model

Results

Conclusion