Eliciting patient-important outcomes through group brainstorming: when is saturation reached?

Abstract

Purpose: Group brainstorming is a technique for the elicitation of patient input that has many potential uses, however no data demonstrate concept saturation. In this study we explore concept saturation in group brainstorming performed in a single session as compared to two or three sessions. Methods: Fifty-two predominately African American adults patients with moderately to poorly controlled Diabetes Mellitus participated in three separate group brainstorming sessions as part of a PCORI-funded group concept mapping study examining comparing methods for the elicitation of patient important outcomes (PIOs). Brainstorming was unstructured, in response to a prompt designed to elicit PIOs in diabetes care. We combined similar brainstormed responses from all three sessions into a ‘master list’ of unique PIOs, and then compared the proportion obtained at each individual session, as well as those obtained in combinations of 2 sessions, to the master list. Results: Twenty-four participants generated 85 responses in session A, 14 participants generated 63 in session B, and 14 participants generated 47 in session C. Compared to the master list, the individual sessions contributed 87%, 76%, and 63% of PIOs. Session B added 3 unique PIOs not present in session A, and session C added 2 PIOs not present in either A or B. No single session achieved >90% saturation of the master list, but all 3 combinations of 2 sessions achieved > 90%. Conclusions: Single sessions elicited only 63-87% of the patient-important outcomes obtained across all three sessions, however all combinations of two sessions elicited over 90% of the master list, suggesting that 2 sessions are sufficient for concept saturation. Trial registration: NCT02792777. Registered 2 June 2016.