Reappraisal of the options for colorectal cancer screening report
The "Reappraisal of the options for colorectal cancer screening report" is a substantial update to the "Option appraisal of population-based colorectal cancer screening programmes in England" (Tappenden et al. 2007). It includes a significant development in modelling methodology, and incorporation of data from (1)the first two rounds of the England gFOBT BCSP, (2) a large randomized UK trial of FS for ages 55 to 64 years, and (3) further data on the immunochemical FOBT (iFOBT)
S Whyte, J Chilcott, K Cooper, M Essat, J Stevens, R Wong, N Kalita
Colorectal cancer (CRC) is the third most common form of cancer in the UK; 36,600 new cases were diagnosed in 2007 and there were 16,259 CRC-related deaths in 2008.(1) The aim of population-based screening for CRC is to reduce mortality through both prevention (by the removal of adenomas) and earlier diagnosis of CRC.
In 2004, Tappenden et al. produced a report to the English Bowel Cancer Screening Working Group which appraised the options for colorectal cancer screening evaluating cost-effectiveness, cost-utility and resource impact. (2, 3) This study used a mathematical model to compare screening options using the guaiac faecal occult blood test (gFOBT) or flexible sigmoidoscopy (FS) for different age groups. The report concluded that screening using FOBT and/or FS is potentially a cost-effective strategy for the early detection of colorectal cancer. This report informed the Department of Health’s policy on bowel cancer screening in England. The Bowel Cancer Screening Programme (BCSP) commenced rollout in England in 2006 offering biennial screening with gFOBT to persons aged 60 – 69 years, and in 2009, rollout to include the 70-74 age group commenced.
Since the original options appraisal, the ScHARR CRC screening model has been updated considerably. The model now uses a Bayesian approach with the Metropolis Hastings algorithm to jointly estimate the CRC natural history state transition parameters and gFOBT test characteristics.(4) This approach generates parameter estimates using the ScHARR CRC natural history and screening model, together with several data sources including CRC incidence in the absence of screening and data from the first round of screening.