University of Sheffield

Tony O'Hagan - Academic pages - Abstracts


On Estimators of Medical Costs with Censored Data

Anthony O'Hagan and John W. Stevens

Statistical Services Unit, University of Sheffield, Sheffield, England and Astra Charnwood, Loughborough, England

Publication details: Paper submitted to Journal of Health Economics, 23, 615-625, 2004.


In the assessment of cost-effectiveness of alternative medical technologies, it is necessary to estimate the mean total cost per patient over the relevant patient population. Where information about costs comes from a clinical trial, a common problem is censoring. In the presence of censoring, care is needed to estimate mean total costs. Two apparently quite different approaches have been proposed by Lin et al (1997) and by Bang and Tsiatis (2000). We examine these estimators with a view to finding theoretical connections between them.

We conclude that if patient costs are only available at the time of death or censoring, then the approaches are equivalent because the Bang and Tsiatis complete-case estimator can be identified with a previously unreported limiting form of the first Lin et al estimator. Using this connection, we clarify the Bang and Tsiatis partitioned estimator and compare it with the second Lin estimator.

Finally, we discuss the robustness of the methodology underlying both approaches.

Return to my publications page.
Updated: 27 April 2004
Maintained by: Tony O'Hagan