Surveillance colonoscopy in older adults with a history of colon polyps presents a complex decision, as procedural risks increase and potential benefits decrease with age. Despite this, no formal guidance exists to support decision-making. We aimed to develop, refine, and pilot-test a one-page conversation aid to facilitate shared decision-making (SDM) between older adults and primary care providers (PCPs) regarding surveillance colonoscopy. We conducted a multi-phase study incorporating both qualitative and quantitative methods to develop, refine, and pilot-test a conversation aid. In Phase 1, we iteratively refined a draft conversation aid through semi-structured interviews with older adults aged ≥ 70. In Phase 2, we surveyed older adults, PCPs, and gastroenterologists to assess acceptability and implementation preferences. In Phase 3, we pilot-tested the conversation aid during routine primary care visits. We assessed SDM using the Observer OPTION-5 tool and applied thematic analysis and descriptive statistics. Phase 1: Ten older adults participated in interviews and suggested minor improvements to the conversation aid, including the addition of illustrations. Phase 2: Surveys of 200 older adults, 25 PCPs, and 25 gastroenterologists showed that the conversation aid was clear, useful, and appropriately timed for discussions about repeat colonoscopy. Phase 3: The conversation aid was pilot-tested by five PCPs in clinical encounters with 17 patients. It was particularly helpful for patients who were undecided about continuing surveillance. The average OPTION-5 score reflected moderate levels of SDM, with variation across encounters. PCPs supported flexible use based on individual patient preferences and endorsed multiple modes of delivery. The conversation aid was feasible and acceptable for use in primary care. It shows promise for supporting individualized decisions about surveillance colonoscopy in older adults. Future research should evaluate its impact on decision quality and integration into clinical workflows.
Schubbe et al. (2026) studied this question.