Risk Adjustment Revisited Using Machine Learning Techniques

23 Pages Posted: 26 May 2017

See all articles by Alvaro Riascos

Alvaro Riascos

Universidad de los Andes, Colombia - Department of Economics

Mauricio Romero

ITAM, Centro de Investigación Económica

Natalia Serna

Quantil | Matemáticas Aplicadas

Date Written: March 10, 2017

Abstract

Risk adjustment is vital in health policy design. Risk adjustment defines the annual capitation payments to health insurers and is a key determinant of insolvency risk for health insurers. In this study we compare the current risk adjustment formula used by Colombia’s Ministry of Health and Social Protection against alternative specifications that adjust for additional factors. We show that the current risk adjustment formula, which conditions on demographic factors and their interactions, can only predict 30% of total health expenditures in the upper quintile of the expenditure distribution. We also show the government’s formula can improve significantly by conditioning ex ante on measures indicators of 29 long-term diseases. We contribute to the risk adjustment literature by estimating machine learning based models and showing non-parametric methodologies (e.g., boosted trees models) outperform linear regressions even when fitted in a smaller set of regressors.

Keywords: Risk Adjustment, Diagnostic Related Groups, Risk Selection, Machine Learning

JEL Classification: I11, I13, I18, C45, C55

Suggested Citation

Riascos, Alvaro and Romero, Mauricio and Serna, Natalia, Risk Adjustment Revisited Using Machine Learning Techniques (March 10, 2017). Documento CEDE No. 2017-27, Available at SSRN: https://ssrn.com/abstract=2973514 or http://dx.doi.org/10.2139/ssrn.2973514

Alvaro Riascos (Contact Author)

Universidad de los Andes, Colombia - Department of Economics ( email )

Carrera 1a No. 18A-10
Santafe de Bogota, AA4976
Colombia

Mauricio Romero

ITAM, Centro de Investigación Económica ( email )

Camino a Santa Teresa No. 930
Col. Héroes de Padierna
Ciudad de México
Mexico

Natalia Serna

Quantil | Matemáticas Aplicadas ( email )

Bogotá
Colombia

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
130
Abstract Views
847
Rank
395,220
PlumX Metrics