Risco de insustentabilidade financeira dos beneficiários de uma operadora de planos de saúde: uma comparação de modelos de classificação

Abstract

This study aimed to present an analytical study on the financial sustainability of the beneficiaries of a health insurance carrier portfolio. The sample investigated in this study is a provider of health plan linked to the Banco do Nordeste do Brazil SA (BNB) and consisted of 38,875 users, active between the years 2011 and 2013. Specifically, in this study we sought to identify the main determinants of financial sustainability of beneficiaries of the portfolio of carrier. Statistical techniques for supervised classification were employed logistic regression, classification trees and nearest neighbor classifier. In addition, the ROC curve was used to compare the performances of the techniques used, with an area under the curve (AUC), the main measure observed. The results showed that the majority of the sample is composed of organic recipients. The logistic regression model obtained accuracy of 68.43% with AUC of 0.7501, the trees obtained 67.76% and AUC of 0.6855, while the nearest neighbor classifier had an accuracy of 67.22% and AUC of 0.7258. The variables considered most important by the first two models, considering a joint analysis are the age and the type of plan, among those that define the user profile, and the IRS, and Dental Consultation, those that define the historical use of user.