Abstract
Introduction
The application of artificial intelligence (A.I) using patient reported outcomes (PROs) to predict benefits, risks, benefits and likelihood of improvement following surgery presents a new frontier in shared decision-making. The purpose of this study was to assess the impact of an A.I-enabled decision aid versus patient education alone on decision quality in patients with knee OA considering total knee replacement (TKR). Secondarily we assess impact on shared decision-making, patient satisfaction, functional outcomes, consultation time, TKR rates and treatment concordance.
Methods
We performed a randomized controlled trial involving 130 new adult patients with OA-related knee pain. Patients were randomized to receive the decision aid (intervention group, n=65) or educational material only (control group, n=65) along with usual care. Both cohorts completed patient surveys including PROs at baseline and between 6–12 weeks following initial evaluation or TKR. Statistical analysis included linear mixed effect models, Mann-Whitney U tests to assess for differences between groups and Fisher's exact test to evaluate variations in surgical rates and treatment concordance.
Results
The intervention group showed greater decision quality (K-DQI, Mean difference = 20%, p<0.0001), collaboration in decision-making (CollaboRATE, 12% (intervention group), 47% (control group) below median, p<0.0001), satisfaction with consultations (NRS-C, 14% (intervention group), 33% (control group) below median, p=0.008), improvement in functional outcomes from baseline up to 12 week follow-up (KOOSJR, 4.9 pts higher (intervention group), p=0.029) without significantly impacting consultation time. No differences were observed in TKR rates or treatment concordance.
Conclusion
A.I-enabled decision aids incorporating PROs in predictive algorithms can improve decision quality, level of shared decision-making, satisfaction with patient-provider consultations, and functional outcomes, without extending consultation times. The combination of advanced predictive technologies and patient reported data to forecast surgical outcomes presents a paradigm shift in shared decision making and the delivery of high value care for patients with knee OA.
