2016 journal article

A flexible model for correlated medical costs, with application to medical expenditure panel survey data

STATISTICS IN MEDICINE, 35(6), 883–894.

co-author countries: United States of America 🇺🇸
author keywords: model selection; generalized linear model; health econometrics; semiparametric regression; generalized estimating equation
MeSH headings : Aged; Aged, 80 and over; Bias; Computer Simulation; Data Interpretation, Statistical; Health Care Costs / statistics & numerical data; Health Care Costs / trends; Health Expenditures / statistics & numerical data; Health Expenditures / trends; Humans; Linear Models; Male; Models, Economic; United States
Source: Web Of Science
Added: August 6, 2018

We propose a flexible model for correlated medical cost data with several appealing features. First, the mean function is partially linear. Second, the distributional form for the response is not specified. Third, the covariance structure of correlated medical costs has a semiparametric form. We use extended generalized estimating equations to simultaneously estimate all parameters of interest. B‐splines are used to estimate unknown functions, and a modification to Akaike information criterion is proposed for selecting knots in spline bases. We apply the model to correlated medical costs in the Medical Expenditure Panel Survey dataset. Simulation studies are conducted to assess the performance of our method. Copyright © 2015 John Wiley & Sons, Ltd.