An NEJM randomized clinical trial finds a 50% reduction in colorectal cancer deaths after colonoscopy screening but concludes only minimal benefits David E. Scheim, Ph.D., a Colleen Aldous, Ph.D., MASSAf b (This article as a PDF.) a US Public Health Service, Commissioned Corps, Inactive Reserve. email: dscheim@alum.mit.edu b College of Health Sciences, University of KwaZulu-Natal, Durban, South Africa; fellow of the Academy of Science, South Africa Abstract A randomized clinical trial published in October 2022 in the NEJM examined the efficacy of colonoscopy screenings in preventing incidence and mortality of colorectal cancer. Only 42% of the intervention group of this study who were invited to undergo a colonoscopy actually underwent that procedure, and the study report concluded, using intention-to-treat (ITT) analysis, that this procedure yielded only a marginal benefit in reducing colorectal cancer deaths. An accompanying editorial in the NEJM likewise concluded that study’s “relatively small reduction in the risk of colorectal cancer” was “both surprising and disappointing.” However, the use of ITT analysis in drawing this conclusion was patently flawed; based on per-protocol analysis, which was appropriate here, this study found a 50% reduction in colorectal cancer deaths from colonoscopy screenings. It is noteworthy in considering the unfounded conclusion reported by this study and accompanying editorial that its coauthors, collectively, had disclosed conflicts of interest to nine manufacturers of advanced colon cancer screening equipment. Introduction On October 9, the New England Journal of Medicine (NEJM) published a randomized clinical trial tracking 85,000 subjects over ten years which concluded that colonoscopies provided less protection against colon cancer than previously assumed, with only a 10% reduction in deaths. 1 An accompanying NEJM editorial found the study’s “relatively small reduction in the risk of colorectal cancer” to be “both surprising and disappointing.” 2 News articles echoed these conclusions, for example, “Screening Procedure Fails to Prevent Colon Cancer Deaths in Large Study,” as Bloomberg reported. 3 CNN summarized, “the benefits of colonoscopies for cancer screening may be overestimated.” 4 Only on its sixth page, however, does this NEJM study report that its colonoscopy arm consisted of subjects invited to receive a colonoscopy. Only 42% of those actually had a colonoscopy and in those who had them, colon cancer deaths were reduced by half vs. controls—a dramatic 50% benefit. The NEJM editorial mentions this point as well, but only tangentially in its fourth paragraph, as one of “several potential explanations for these discouraging results.” Inappropriate Application of Intention-to-Treat Analysis The upshot of this study was that a disregarded invitation to have a colonoscopy gave no preventative benefit against colon cancer. That would be like concluding that nail clippers were only 50% effective because in a study that distributed them, only half the subjects cut their nails. This was the reasoning behind the study’s and editorial’s patently absurd main conclusion, which derived from an inappropriate application of intention-to-treat (ITT) analysis. If, for example, a randomized trial compared results from drug treatment and placebo groups, subjects initially doing well taking the drug might be more inclined to follow the protocol to completion (“per-protocol”).