Estimating the costs attributable to a disease with application to ovarian cancer

Abstract

This article is concerned with the methodological issues that arise when estimating the expected costs attributable to a disease. In particular, the article considers methods appropriate for handling incomplete or censored cost and survival data, incorporating discounting, and computing attributable costs. After motivating the need for an estimate of the average, present value of the attributable costs, we present the Kaplan-Meier sample-average (KMSA) estimator, which takes into account the censored nature of the data that are typically available. We investigate the statistical properties of the estimator and compare it to others employed in the literature, showing how certain methods for incorporating discounting can introduce bias. We demonstrate the utility of the estimator by applying it to estimation of the costs attributable to ovarian cancer, using data from a database linking Medicare claims with the Surveillance, Epidemiology, and End Results cancer registry. Our analysis suggests that the average, present value of the 15-year costs attributable to ovarian cancer is $21,285 for local-stage cases and $32,126 for distant-state cases in 1990 dollars.