[Skip to Navigation]
Sign In
Comment & Response
September 16, 2021

Is Colorectal Cancer Screening Absolutely Beneficial for Older Adults?

Author Affiliations
  • 1Clinical and Translational Epidemiology Unit, Massachusetts General Hospital, Harvard Medical School, Boston
JAMA Oncol. 2021;7(11):1728-1729. doi:10.1001/jamaoncol.2021.4155

In Reply We appreciate the comments from Varadhan and Schoenborn on our cohort study in which we reported the benefit of screening lower endoscopy among adults after age 75 years, especially those without significant comorbidities.1 We concur that absolute risk reduction should be considered in the evaluation of the public health outcome of clinical interventions on a population level. Based on our results, Varadhan and Schoenborn calculated the number needed to screen (NNS) as 1350 for incident colorectal cancer (CRC) and 12 635 for CRC-related death for screening after age 75 years. However, these calculations were based on person-years or incidence rates. A more clinically understandable metric would be based on persons or risk. Thus, when we convert person-years to persons based on the mean follow-up time (6.8 years for CRC incidence and 8.1 years for CRC mortality for age >75 years), the risk of CRC was 489/(246 717/6.8) in older adults who were unscreened and 172/(138 592/6.8) in older adults who were screened. This translates to an NNS of 198. In other words, screening would be expected to prevent 1 incident case of CRC for every 198 screened persons in older adults. Similarly, the risk of CRC-related death was 212/(289 036/8.1) in unscreened and 111/(169 639/8.1) in screened older adults, translating into an NNS of 1579, or screening preventing 1 CRC-related death for every 1579 persons in older adults. These NNS estimates are consistent with results from randomized clinical trials in adults younger than 75 years. For example, the Prostate, Lung, Colorectal, and Ovarian (PLCO) trial reported that the number needed to invite for screening with sigmoidoscopy was 282 (95% CI, 210-427) to prevent 1 case of CRC, and 871 (95% CI, 567-1874) to prevent 1 case of CRC-related death.2 Thus, based on these estimates of NNS, we believe that the absolute benefits of screening in older adults are quite substantial. Nonetheless, we concur that recommendations for screening in older adults should weigh the potential complications of screening against absolute benefits.

Add or change institution