Anne M. Walling, Rebecca L. Sudore, Lisa Gibbs, Maryam Rahimi, Ron D. Hays, Chi‐Hong Tseng, Kanan Patel, Katherine Santos, Fernando J Sanz Vidorreta, Aaron J. Chau, Juan Carlos Antonio Lopez, Jamie Anand, Goolnora Marshall, Anna DePaolis-Dickey, Kirsten Inducil Buen, Douglas S. Bell, Christine S. Ritchie, Victor Gonzalez, Neil S. Wenger
BACKGROUND: Advance care planning (ACP) can improve communication of patients' preferences but is underutilized in health systems. OBJECTIVE: To compare approaches to increase ACP. DESIGN: Pragmatic, 24-month, comparative, cluster randomized trial. (ClinicalTrials.gov: NCT04012749). SETTING: Fifty clinics in 3 University of California health systems. PARTICIPANTS: All seriously ill primary care patients without an advance directive (AD) or a Physician Orders for Life-Sustaining Treatment (POLST) form in the electronic health record (EHR). INTERVENTION: Three ACP interventions, timed to primary care visits. The group 1 intervention included a letter with an AD sent via the EHR portal and mail. The interventions in groups 2 and 3 additionally included a link to PrepareForYourCare.org and a mailed pamphlet, and the group 3 intervention also included health navigator outreach. Clinicians received ACP training. MEASUREMENTS: Presence of AD or POLST form in the EHR at 12 and 24 months (primary outcome); documented ACP discussions and utilization (secondary outcomes). RESULTS: Among 5810 seriously ill patients (mean age, 71 years; 48% female; 50% in a racial or ethnic minority), by 24 months, 13.7% (95% CI, 12.1% to 15.3%) of patients in group 1, 12.7% (CI, 11.2% to 14.1%) in group 2, and 19.8% (CI, 18.1% to 21.5%) in group 3 had a documented AD or POLST form. After adjustment for patient and site factors, group 3 patients were more likely to have an AD or a POLST form compared with group 1 (adjusted difference [aDiff], 4.6% [CI, 0.8% to 8.4%]) and group 2 (aDiff, 5.5% [CI, 0.8% to 10.2%]); groups 1 and 2 did not differ significantly. Documented ACP discussions were higher in group 3 than group 1 (aDiff, 4.7% [CI, 1.4% to 7.9%]) and group 2 (aDiff, 4.2% [CI, 1.1% to 7.2%]); groups 1 and 2 did not differ. Utilization did not differ by group. LIMITATIONS: Academic health centers; no control group. CONCLUSION: Health system implementation of automated ACP interventions with clinician training and mailed materials increases ACP documentation and is enhanced with navigator outreach. PRIMARY FUNDING SOURCE: Patient-Centered Outcomes Research Institute.