A spatio-temporal auto-regressive model for generating savings calls to a health line

Abstract

Urgency admission is one of the most important factors regarding hospital costs, which can possibly be mitigated by the use of national health lines such as the Portuguese Saúde24 line (S24). Aiming future development of decision support indicators in a hospital savings context, based on the economic impact of the use of S24 rather than hospital urgency services, this study investigates spatio-temporal dependencies of the number of S24 calls generating savings in each Portuguese municipality, over the period 2010–2016, under an autoregressive approach. An econometric analysis of the savings obtained by the use of S24 is also carried out considering a savings index. Combining insights from classical spatial econometrics and from the analysis of spatio-temporal data, novel Bayesian Poisson spatio-temporal lag models are presented and applied in this paper. This extends to time the ideas of a Bayesian Poisson spatial lag model, considering both a parametric and a non-parametic structure for time and space-time effects. The results obtained for the savings index reveal that, over the last seven years, there has been a more comprehensive spatial effectiveness of the S24 line in solving the non-urgent emergency situations, that could be handled by primary health care services or in a self care basis.