skip to main content
Centre for Health Economics and Policy Analysis


Method to improve analysis of cost-of-illness data to be discussed at seminar

27 Mar 2009

A method to provide more accurate health care cost data, when the observations are censored for analysing either lifetime or per-episode-of-illness costs, will be discussed at the CHEPA monthly seminar on April 8.

Willard G. Manning, professor in the Harris School of Public Policy Studies and in the Department of Health Studies in the Medical School at the University of Chicago, will share the results of a research paper that identifies a major bias when estimating mean cumulative costs in health care when the data collected are censored, as often occurs in data from trials and from some disease registries used in cost-effectiveness and other health economics studies. The authors propose an alternative to more fully incorporate information from censored periods.

The paper builds on the issue of skewness that results from using health care costs based on fixed periods of time to analyse either lifetime or per-episode-of-illness costs. Such data is not well suited to cost-effectiveness studies, which may use data with varying periods of observation and right censoring of cases before death or the end of the episode of illness. While methods have been developed to adjust for this issue, there is some concern about those methods in terms of robustness and bias.

Manning’s paper, Estimating Lifetime or Episode-of-illness Costs, prepared with University of Chicago colleague Anirban Basu, supports the theory that estimators based on inverse probability weighting may yield biased estimates of accumulated costs in situations with substantial censoring. The alternative presented in this paper is consistent and more efficient for these designs.

Manning is a renowned health economist who earned his PhD at Stanford University. His primary area of interest is the effects of health insurance, and he has also examined statistical, measurement and economic issues in modelling the use of health services and health care expenditures.

The seminar will take place from 12:30 to 1:30 p.m. in Room HSC-1J8. All are welcome to attend.

Please send questions, comments or suggestions to the Webmaster ( ).
Please review our Copyright Disclaimer policy

Interactive Marketing | Cubicle Fugitive