Generalized modeling approaches to risk adjustment of skewed outcomes data

Abstract

There are two broad classes of models used to address the econometric problems caused by skewness in data commonly encountered in health care applications: (1) transformation to deal with skewness (e.g., ordinary least square (OLS) on ln(y)); and (2) alternative weighting approaches based on exponential conditional models (ECM) and generalized linear model (GLM) approaches. In this paper, we encompass these two classes of models using the three parameter generalized Gamma (GGM) distribution, which includes several of the standard alternatives as special cases-OLS with a normal error, OLS for the log-normal, the standard Gamma and exponential with a log link, and the Weibull. Using simulation methods, we find the tests of identifying distributions to be robust. The GGM also provides a potentially more robust alternative estimator to the standard alternatives. An example using inpatient expenditures is also analyzed.

Economic evaluation of a task-shifting intervention for common mental disorders in India

“…To assess the differences between the treatment and control groups and to allow for the skewed nature of the data on costs and outcomes, generalized linear models were used, with gamma distributions and log transformations. [34][35][36] Estimates were adjusted for baseline depression scores and are presented with confidence intervals (CIs). The data analyses used in earlier studies on the MANAS trial were based on cluster-level summaries 37 but there were too few clusters to use this approach in the present study.…”

Effect of Chronic Kidney Disease and Comorbid Conditions on Health Care Costs: A 10-Year Observational Study in a General Population

“…The first part of the two-part model used logistic regression to predict the probability of positive costs. A log-gamma generalized linear model (GLM) was used in the second stage to analyze positive expenditures [24]. Logistic and discrete-time hazard regressions [25] were used to model the probability of any hospitalization in cross-sectional and longitudinal models, respectively.…”

Impacto económico de la resistencia a meticilina en pacientes con bacteriemia por Staphylococcus aureus en hospitales de Bogotá

“…Se ajustó un modelo lineal generalizado con distribución gamma y enlace de logaritmo acorde con lo descrito en la literatura científica para variables económicas que no cumplen con los supuestos de la regresión lineal (28). La variable de respuesta fue el costo directo de la atención.…”