Cost-Effectiveness of Colorectal Cancer Screening Strategies-A Systematic Review

Background & aims: Widespread screening for colorectal cancer (CRC) has reduced its incidence and mortality. Previous studies investigated the economic effects of CRC screening. We performed a systematic review to provide up-to-date evidence of the cost effectiveness of CRC screening strategies by answering 3 research questions.

Methods: We searched PubMed, National Institute for Health Research Economic Evaluation Database, Social Sciences Citation Index (via the Web of Science), EconLit (American Economic Association) and 3 supplemental databases for original articles published in English from January 2010 through December 2017. All monetary values were converted to US dollars (year 2016). For all research questions, we extracted, or calculated (if necessary), per-person costs and life years (LYs) and/or quality-adjusted LYs, as well as the incremental costs per LY gained or quality-adjusted LY gained compared with the baseline strategy. A cost-saving strategy was defined as one that was less costly and equally or more effective than the baseline strategy. The net monetary benefit approach was used to answer research question 2.

Results: Our review comprised 33 studies (17 from Europe, 11 from North America, 4 from Asia, and 1 from Australia). Annual and biennial guaiac-based fecal occult blood tests, annual and biennial fecal immunochemical tests, colonoscopy every 10 years, and flexible sigmoidoscopy every 5 years were cost effective (even cost saving in most US models) compared to no screening. In addition, colonoscopy every 10 years was less costly and/or more effective than other common strategies in the United States. Newer strategies such as computed tomographic colonography, every 5 or 10 years, was cost effective compared with no screening.

Conclusions: In an updated review, we found that common CRC screening strategies and computed tomographic colonography continued to be cost effective compared to no screening. There were discrepancies among studies from different regions, which could be associated with the model types or model assumptions.