National Cancer Institute Home at the National Institutes of Health |
Please wait while this form is being loaded....
The Applied Research Program Web site is no longer maintained. ARP's former staff have moved to the new Healthcare Delivery Research Program, the Behavioral Research Program, or the Epidemiology & Genomics Research Program, and the content from this Web site is being moved to one of those sites as appropriate. Please update your links and bookmarks!

Developing Innovative Methods to Estimate Costs of Cancer Care

Principal Investigator: Anirban Basu, PhD
Associate Professor, Department of Health Services, School of Public Health
Department of Pharmacy
University of Washington, Seattle

What's the problem?

Estimating the total cost of cancer care, from diagnosis to death, is an important goal for researchers and policy makers. These costs generally follow a U-shaped curve, in which initial costs around diagnosis and primary treatment are very high. Costs then diminish during survivorship. If the cancer recurs, costs escalate again because of renewed treatment and end-of-life care.

Deriving a complete and accurate estimate of the total costs of care is difficult, however, because data are often incomplete. For example, a researcher may be working with Medicare claims data, which begin at age 65, but if diagnosis for members of the study sample occurred before age 65, the costs of those earlier years of care are not captured. In addition, many patients observed in the data may still be alive and it may not be clear if they have or have not entered the late phase of high cost care. These missing data are referred to as "censored data". Censored data are well-known for their idiosyncrasies and the difficulties they pose for applied health services researchers. This problem is further compounded by the fact that researchers may not necessarily take into account all the variables that affect treatment costs, such as factors that influence a patient to choose one treatment over another. Finally, considerable heterogeneity may exist in the cost trends for different types of patients.

[Return to top]

How has this research addressed the problem?

Building on previous work on treatment costs conducted by ARP researchers, Dr. Basu is developing innovative statistical methods that account for these problems. He is focusing on prostate cancer in this work.

Dr. Basu's methods bridge the literatures on modeling skewed cost data to account for right censoring (data missing at the end of life), and on the use of instrumental variables (IV), a statistical method used to estimate or compare the impact of treatment or policy when all variables may not have been accounted for. It also can be used to account for heterogeneity in outcomes.

[Return to top]

Significance of the research & results

In developing these statistical methods and applying them to prostate cancer, Dr. Basu's imaginative scientific thinking will help investigators effectively model health expenditure data for this common cancer. His methods can be used to study the effect of a treatment or policy on total costs and also the composition of that effect that may consist of effects generated due to changes in survival and/or changes in intensity of medical care use.

He has focused in particular on methods that can be used in studies with observational data, as opposed to controlled clinical trial data. This is particularly relevant to prostate cancer, which has a paucity of evidence on comparative effectiveness available from controlled clinical trials. Moreover, his work will provide valuable information about comparative cost estimates for this cancer, which will have important implications for cost-effectiveness and other policy analyses.

[Return to top]

Recent related publications of interest

Basu A. Estimating Decision-Relevant Comparative Effects Using Instrumental Variables. Stat Biosci 2011 Sep;3(1):6-27. [View Abstract]

Basu A, Manning WG. Estimating lifetime or episode-of-illness costs under censoring. Health Econ 2010 Sep;19(9):1010-28. Erratum in: Health Econ. 2011 Jan;20(1):125-6. [View Abstract]

Basu A, Heckman JJ, Navarro-Lozano S, Urzua S. Use of instrumental variables in the presence of heterogeneity and self-selection: an application to treatments of breast cancer patients. Health Econ 2007 Nov;16(11):1133-57. [View Abstract]

[Return to top]

Last Modified: 03 Sep 2013