Evaluating Patient-Centered Surgical Care Quality Using Patient-Reported Measures of Shared Decision-Making.

Abstract

Background: Despite the importance of shared decision-making (SDM) in surgical care, evaluating it from the patient's perspective is not a performance measure (PM). We aimed (1) to determine whether levels of SDM could be distinguished among hospitals and surgeons and (2) to understand implementation feasibility by estimating minimum response numbers per hospital and surgeon to maintain acceptable levels of statistical reliability.

Study design: Two patient-reported SDM measures were administered during the American College of Surgeons NSQIP Patient-Reported Outcome Measures demonstration project between 2020 and 2023: the CollaboRATE measure and the 9-item Shared Decision Making Questionnaire (SDM-Q-9). Twenty-four candidate performance measures were constructed to evaluate hospital- and surgeon-level performance. Mixed models estimated statistical reliability for each PM and projected minimum sample sizes required to achieve reliabilities of at least 0.70 and 0.40. Effects of risk adjustment, nonresponse bias, and ambulatory procedure classification were explored.

Results: Among hospitals, the highest median reliability was achieved with CollaboRATE after correcting for nonresponse bias and among the subgroup of ambulatory procedures at 0.81 (interquartile range 0.72 to 0.85) and 0.81 (0.73 to 0.85), respectively. Projected minimum sample sizes required to achieve acceptable reliability were the lowest at 235 and 192 responses per hospital, respectively. Among surgeons, the highest median reliability was also achieved with CollaboRATE after correcting for nonresponse bias and among the subgroup of ambulatory procedures at 0.49 (0.41 to 0.61) and 0.47 (0.39 to 0.59), respectively. Projected minimum sizes were lowest at 46 and 46 responses per surgeon, respectively.

Conclusions: Although both CollaboRATE and SDM-Q-9 could distinguish patient-centered care quality, CollaboRATE identified performance differences with higher statistical reliability than the SDM-Q-9. With few responses required per surgeon, implementing CollaboRATE as a surgeon-specific PM is realistic and feasible to distinguish patient-centered SDM practices.