Improving Patient-Centered Communication in Primary Care: Results from a Multisite Cluster Randomized Trial

Context: Despite various attempts to improve patient-clinician communication, there has been limited head-to-head comparison of these efforts.

Objective: to enhance patient-clinician communication, building on a previous pilot that created promising prototypes of interventions. Population

Studied: 4,852 patients and 114 primary care clinicians in 21 clinics in 3 health systems in 2 states in the U.S.

Intervention: A cluster randomized controlled trial with 3 arms: in-person communication coaching of clinicians with standardized patient instructor training (High-Touch), mobile application-based communication coaching of clinicians (High-Tech), and posters placed in exam rooms encouraging shared-decision making (ASK). In advance of visits, patients in the High-Touch and High-Tech arms were prompted to inform their clinicians of the most important matter they wanted to address during the visit using online e-checkin.

Outcome measures: The primary outcome was a validated 3-item patient engagement measure (CollaboRATE) that measures patient perceptions of communication and decisionmaking during the visit. Secondary outcomes included patients’ confidence in managing their health and Net Promoter Score (likelihood of recommending this care provider to others).

Methods: A differencein-differences mixed-effect logistic regression with random intercepts for primary care clinician was used for analyzing differences in outcomes including superiority tests for High-Tech vs ASK, High-Touch vs ASK, and a non-inferiority comparison of High-Tech vs High-Touch.

Results: The 3 interventions did not differ significantly in CollaboRATE: High-Tech vs. ASK, odds ratio (OR) 1.011, 95%CI 0.700, 1.460, p=0.952; High-Touch vs, ASK, OR 1.077, 95%CI 0.741, 1.564, p=0.699; Hi-Tech vs. High-Touch, OR 0.939, 95%CI 0.649, 1.360, p=0.138. After 3 months, compared with patients receiving care from clinicians at

High-Touch clinics, patients receiving care at High-Tech clinics had less confidence in managing their health (odds ratio [OR]=0.174; 95% CI: -0.339, -0.009; p=0.039), and were less likely to recommend their clinician to others (OR=0.579; 95% CI: 0.342, 0.978; p=0.041).

Conclusions: The primary outcome showed no evidence of differential effects of the interventions. Some secondary outcomes suggested positive effect of clinicians receiving in-person coaching. Alternative measures of patients’ engagement in shared decision-making may be needed.