A Learning Curve for Shared Decision Making: The Impact of Physician Experience on Decision Aid Efficacy in Severe Aortic Stenosis

Coylewright M, O'Neill E, Sherman A, Gerling MJ, Adam K, Xu K, Grande SW, Dauerman HL, Elwyn G



Background: The adoption of shared decision making (SDM) in cardiology practice is difficult. This study investigates the impact of physicians’ experience using AS Choice, a decision aid (DA) to help patients with severe aortic stenosis explore options between transcatheter aortic valve replacement and medical therapy, on patient-centered outcomes.

Methods: Surveys were conducted at 3 time points, with unique patients at each use: T0 (usual care without DA), T1 (physicians used DA for the first time) and T2 (fifth use). Patient knowledge was obtained via pre- and post-visit surveys. Patient satisfaction and SDM were measured by CAHPS and CollaboRATE, respectively. Comparisons were made by paired t-test and Fisher exact test.

Results: Four physicians and one physician pair (visits held together) participated in the study at two medical centers. Thirty-four patients were included with no patient using DA twice: 25 at T0 (5 per physician), 4 at T1 and 5 at T2. Patients had a mean age of 86.8 years and over half were female. Patients at T2 were significantly older than those at T0 (p=0.03). Patient knowledge increased stepwise from baseline to first use to fifth use, with a significant difference between baseline and multiple uses (p=0.04). Patient satisfaction and SDM also improved with multiple uses (p=0.01 for both). (Figure)

Conclusion: This study suggests a step-wise improvement in patient-centered outcomes with physicians’ greater experience using a DA, and may inform future implementation efforts.

a learning curve Figure. Patient outcomes by physician. There were 25 patients at baseline, 5 for each physician or physician pair with an average score calculated for each physician or physician pair (one missing patient at T1). (A) A four-item measure was used to assess patient knowledge about treatment choices via pre- and post-visit surveys. (B) CAHPS is a six-item survey that evaluates patient satisfaction. © CollaboRATE is a three-item measure of shared decision making. CollaboRATE score was calculated by using the sum of three items divided by the perfect score of 27. (D) Paired t-test was used to compute mean of paired differences with 95% CI. Patient knowledge T0 v. T2: mean differences=18% (CI: 1.2-34.8%), p=0.04; Patient satisfaction T0 v. T2: mean differences=6.7% (CI: 2.5-10.8%), p=0.01; Shared decision making T0 v. T2: mean differences=3.9% (CI: 1.6-6.1%). CI= Confidence interval, CAHPS= Consumer Assessment of Healthcare Providers and Systems

Github repo

Github repo