The case-control method has been used frequently to evaluate the impact of screening behavior on cancer mortality, most prominently with regard to the PAP test for cervix cancer and mammography for breast cancer. Although such observational studies are unlikely to be definitive they can provide useful exploratory evidence for the preventive impact of the intervention under study. Motivated by our recent experience in analyzing a study of skin self-examination and its role in preventing mortality from melanoma, we have developed a methodology for separately estimating the impact of the intervention on primary and secondary prevention, and we believe that the method adds both insight and validity to the existing methods for analyzing case-control studies. The general goal of this proposal is to further develop this new methodology and to examine its statistical properties, with a view to developing general guidelines for the design, conduct and analysis of studies examining the joint effects of primary and secondary prevention. Specifically, we will examine the impact of violations of key assumptions of the new method, namely stationarity of incidence rates and screening behavior. We will develop methods for determining how much follow-up time is required for a valid analysis. Finally, we will develop guidelines for study design, including sample size calculations.
| Begg, C B; Cramer, L D; Venkatraman, E S et al. (2000) Comparing tumour staging and grading systems: a case study and a review of the issues, using thymoma as a model. Stat Med 19:1997-2014 |
| Begg, C B; Berwick, M (1997) A note on the estimation of relative risks of rare genetic susceptibility markers. Cancer Epidemiol Biomarkers Prev 6:99-103 |
