Which colorectal cancer screening modality is most effective in improving clinical outcomes?

In this statistical model colonoscopy was more effective than other screening approaches, including no screening, in increasing life expectancy, decreasing colorectal cancer incidence and mortality, and decreasing costs. This model did not address the potential harms of testing. (LOE = 2b)

Barzi A, Lenz HJ, Quinn DI, Sadeghi S. Comparative effectiveness of screening strategies for colorectal cancer. Cancer 2017;123(9):1516-1527.  [PMID:28117881]

Decision analysis

Government

Not applicable

These authors developed a Markov model to simulate the natural history of colorectal cancer and then applied data from various screening trials and an additional screening simulation model. They compared 13 different screening strategies—various combinations of colonoscopy, sigmoidoscopy, computed tomographic colonography, fecal immunochemical test (FIT), fecal occult blood test (FOBT), and stool DNA testing—with no screening. They ran the model against a simulated sample of the US population aged 50 years to 75 years who were at average risk of developing colorectal cancer and followed up for up to 35 years or until death. Given the long time horizon, the authors appropriately discounted years of life gained, costs, cancers prevented, and cost-effectiveness ratios. They used costs of screening tests and cancer treatments from Medicare and Medicaid and also included an analysis that used out-of-pocket costs. The authors were quite explicit as to the sources of data used to drive their model. Since these kinds of models are highly dependent on the validity of the underlying assumptions, the authors performed limited sensitivity analyses on test compliance and on resetting the sensitivity and specificity of DNA testing to 100%. Finally, the authors validated their findings against another widely accepted model, the Microsimulation Screening Analysis (MISCAN), which reported a 26% relative reduction in the incidence of colorectal cancer and a 21% relative reduction in colorectal cancer mortality based on 83% adherence to screening. In the Markov model in this study, colonoscopy was the most effective strategy across multiple clinical scenarios. The strategies, in order of effectiveness, are colonoscopy, computed tomographic colonography, flexible sigmoidoscopy, DNA testing, FOBT, and FIT (DNA testing was more effective than FOBT and FIT by a small margin, but keep reading). Colonoscopy, compared with no screening, increased life expectancy modestly by 0.022 discounted life years gained (1.2 weeks). Screening was also associated with a 5% to 23% relative-risk reduction and a 12% to 34% cancer-specific mortality risk reduction compared with no screening. Additionally, colonoscopy was associated with the lowest costs (of screening, as well as cancer treatment). Colonoscopy was robust, even if the sensitivity and specificity of stool testing increased to 100%. The effect of the harms of testing was absent from this analysis. Finally, in this analysis, fecal DNA testing was not cost-effective and the researchers could identify no set of assumptions under which it would ever become cost-effective.