How should individuals with a false-positive fecal occult blood test for colorectal cancer be managed? A decision analysis.

Several industrialized nations recommend fecal occult blood testing (FOBT) to screen for colorectal cancer (CRC), but corresponding screening guidelines do not specify how individuals with a prior false-positive FOBT result (fpFOBT) should be managed in terms of subsequent CRC screening. Accordingly, we conducted a decision analysis to compare different strategies for managing such individuals. We used a previously developed CRC microsimulation model, SimCRC, to calculate life-years and the lifetime number of colonoscopies (as a measure of required resources) for a cohort of 50-year-olds to whom FOBT-based CRC screening is offered annually from 50 to 75 years. We compared three management strategies for individuals with a prior fpFOBT: (i) resume screening in 10 years with 10-yearly colonoscopy (SwitchCol_long); (ii) resume screening in 1 year with annual FOBT (ContinueFOBT_Short) and (iii) resume screening in 10 years (i.e., the recommended interval following a negative colonoscopy) with annual FOBT (ContinueFOBT_long). We performed sensitivity analyses on various parameters and assumptions. When using different management strategies for individuals with a prior fpFOBT, the variation in the number of life-years gained relative to no screening was <2%, whereas the variation in the lifetime number of colonoscopies was 23% (percentages are calculated as the maximum difference across strategies divided by the lowest number across strategies). The ContinueFOBT_long strategy showed the lowest lifetime number of colonoscopies per life-year gained even when key assumptions were varied. In conclusion, the ContinueFOBT_long strategy was advantageous regarding both clinical benefit and required resources. Specifying an appropriate management strategy for individuals with a prior fpFOBT may substantially reduce required resources within a FOBT-based CRC screening program without limiting its effectiveness.