RANDOMIZED TRIAL TO EVALUATE SCREENING FOR COLON CANCER

明尼苏达结肠癌控制研究（CCCS），一项随机对照研究

The Minnesota Colon Cancer Control Study (CCCS), a randomized controlled trial, is designed to evaluate the effectiveness of a fecal occult blood test (Hemoccult) in reducing mortality from colorectal cancer. A total of 46,551 individuals age 50 to 80 were recruited between 1975 and 1977 and randomly assigned to an annually screened group, a biennially screened group or a control group. Those with a positive screening test were offered a colonoscopy examination at the University of Minnesota. The screening and diagnostic phases of the study concluded in February, 1992 and the initial results, published in the May 13, 1993 issue of the New England Journal of Medicine showed a significant 33 percent reduction in colorectal cancer mortality in the annual screened group. There was a nonsignificant seven percent mortality reduction in the biennial group. It was concluded that in view of the favorable stage distribution and survival rates, the cumulative colorectal cancer mortality rates for the biennial group should decline further relative to the control group, however, additional follow-up is necessary to determine the actual patterns. In addition, the preliminary results to date suggest age effects which can only be clarified with a larger number of events through extended follow-up and accrual of more events. The study has been supported by an NCI contract which expires December 17, 1994. This proposal is to continue to follow the cohort for five years to ascertain cases and deaths of colorectal cancer and perform a series of statistical analyses in order to address some important issues including the effectiveness of biennially screening and subgroup (age and sex) effects. Both of these issues have significant implications on the costs related to screening in the general population. Other important issues to address include the relative contribution of colonoscopy and Hemoccult to the observed mortality reduction in the annually screened group and the effect of polyp removal on the subsequent incidence of colorectal cancer. A number of other analyses are proposed which will provide important information on screening. These include: compliance, test performance, lead time and length bias. It is proposed to utilize the procedures that have been developed to ascertain events, collect supporting documentation for validation, track the cohort and review deaths. An extensive series of statistical analyses are proposed, including many which have already been developed and utilized to address specific questions relevant to a comprehensive assessment of screening efficacy and the characteristics of the screening test.