Analysis of screening data: Colorectal cancer

Background. This paper illustrates how data gathered from an existing screening programme against colorectal cancer can be used to produce new information on the natural history of colorectal cancer as well as the characteristics of the unhydrated Hemoccult II screening test. Methods. A mathematical model is used, which on the basis of prevalence and interval incidence data from a randomized screening project initiated in Funen County, Denmark, estimates the sensitivity of the screening test and the sojourn time of the disease. Results. The sensitivity of the Hemoccult is estimated at 62.1% and the mean sojourn time is estimated to be 2.1 years. Conclusions. The short sojourn time indicates that overall effectiveness of a Hemoccult II screening programme can be improved significantly by more frequent screening.