Determinação dos melhores indicadores contábeis e técnicas de análise multivariada para prever insolvência de operadoras de saúde médico-hospitalares

Abstract

The supplementary health system became essential in the life of modern society. Since the public health system has several deficiencies ranging from overcrowding of the hospital units to the lack of materials and medicines to be provided to patients. Given the importance of the service offered by medical-hospital health care providers, observing their performance in relation to solvency is paramount, because the ruin of these companies cause significant losses to their beneficiaries. On this, this research seeks, through multivariate analysis techniques, to identify which accounting indicators are most relate to the solvency status of a health care provider and which of the techniques plays a better role predicting insolvency. For the experiment, 177 health care providers were randomly selected, 59 of which were insolvent and 118 were solvents. The logistic regression techniques, classification and regression trees, artificial neural networks and random forest were applied to the accounting indicators. The results of the study indicate that the indicator of General Liquidity are the one that most relate to the solvency of the operators and that the artificial neural networks present better performance in the forecast of this situation.