Preferences for colorectal cancer screening of physicians and individuals at average risk in the United States: A discrete choice experiment.

Abstract

60 Background: Several colorectal cancer (CRC) screening options are considered in guidelines for individuals at average-risk (IAR). These options differ in aspects such as invasiveness, recommended frequency, and precision that need to be compared and weighed. This study elicited and compared the relative importance that physicians and IAR place on these screening aspects. Methods: Primary care physicians [PCPs] and gastroenterologists [GIs] who recommended/performed ≥1 screening one month prior to study and adult IAR completed a discrete choice experiment (DCE). Participants repeatedly chose between screening tests described by type of test, frequency, true-positive (TP), true-negative (TN), and adenoma TP (physicians only). The instrument was tested in qualitative (physicians: n=6; IAR: n=6) and quantitative pilots (physicians: n=100; IAR: n=202). A mixed logit model was used to estimate relative attribute importance (RAI) and predicted choice probabilities for colonoscopy, multi-target stool DNA (mt-sDNA), fecal immunochemical test (FIT), and methylated septin 9 (mSEPT9) blood test. Generalizability to the population was confirmed. Results: 1,249 IAR and 400 physicians participated. IAR were 46% male and the mean age was 58.9. Physicians were 79% male and their mean age was 53.4. Preferences were most affected by TP rates (IAR RAI=58%; physicians RAI=42%). Physicians also placed high importance (RAI=41%) on adenoma TP rates. TN rates (IAR RAI=33%; physician RAI=9%), frequency (IAR RAI=6%; physician RAI=2%) and type (IAR RAI=4%; physician RAI=6%) were less important. Despite both IAR and physicians placing most importance on precision, preferences for screening modalities differed. On average, physicians preferred colonoscopy, while IAR preferred mt-sDNA over colonoscopy (p<0.001). Both preferred mt-sDNA and colonoscopy over FIT (p<0.001), with a mSEPT9 blood test being least preferred (p<0.001). Preferences of IAR were heterogeneous with individuals who underwent colonoscopy or sigmoidoscopy screening preferring colonoscopy and the rest preferring mt-sDNA (p<0.001). Conclusions: While both GI and PCPs overwhelmingly preferred colonoscopy, preferences of IAR were heterogenous, with mt-sDNA being preferred on average other modalities. Offering choices in addition to colonoscopy could improve screening uptake. [Table: see text]