Objectives: To determine the detection rate, false positive and false negative rates associated with a policy of retesting with dietary restriction after an initial positive Haemoccult screening test for colorectal cancer, in order to compare the cost effectiveness of such a policy with the alternative, in which all subjects with a positive test would proceed directly to diagnostic colonoscopy.
Methods: Over four years in a large randomised control trial in Nottingham 35,260 subjects had a mean of 1.5 screening rounds each at two-yearly intervals, and were followed up for a minimum of 27 months. During this period subjects with positive screening tests were asked to repeat the test with dietary restrictions. Estimates of costs of the initial screening and of diagnostic colonoscopy were used to estimate the cost for each cancer detected by the different policies.
Results: 1209 subjects had a positive initial screening test and 1033 (85.4%) completed the retests. Four hundred and ninety nine subjects were investigated and 89 cancers detected. In the 710 subjects with negative retests six interval cancers were diagnosed in the two years after screening. If these had been detected by screening under a policy of immediate colonoscopy, test sensitivity would have been improved from 53.6% to 57.2% (P = 0.02), but the cost for each cancer detected would have increased from pound 773 to pound 1509.
Conclusion: Retesting with dietary restrictions reduces costs and maximises the benefit of limited colonoscopy resources, but results in a small but significant reduction in test sensitivity compared with a policy for immediate colonoscopy.