Effectiveness of the Extension for Community Health Outcomes Model as Applied to Primary Care for Autism: A Partial Stepped-Wedge Randomized Clinical Trial.

Importance: The Extension for Community Health Outcomes (ECHO) model is a widely adopted technology-based model for training primary care physicians and practitioners (PCPs) to care for patients with complex conditions. Despite its popularity, to our knowledge, direct effects of ECHO on clinical practice have not been tested in a large-scale study. Objective: To test the effectiveness of the ECHO model as applied to primary care for autism and whether it resulted in improved clinical practice, knowledge, and self-efficacy regarding autism screening and comorbidity management. Design, Setting, and Participants: Primary care physicians and practitioners were recruited to participate in a 6-month ECHO Autism program delivered by 1 of 10 academic medical center sites. A sequential, staggered rollout of ECHO Autism was delivered to 5 cohorts of participants (15 per site; 2 sites per cohort). Sites were randomized after recruitment to cohort/start time. Cohorts launched every 3 months. The ECHO Autism program used videoconferencing technology to connect community-based PCPs with interdisciplinary expert teams at academic medical centers. There were 148 participants (PCPs [family practice physicians, pediatricians, nurse practitioners, and physician assistants] providing outpatient services to underserved children) studied between December 2016 and November 2018. Interventions: The 6-month ECHO Autism program included twelve 2-hour sessions connecting PCP participants with an interdisciplinary expert team. Sessions included didactics, case-based learning, guided practice, and discussion. Main Outcomes and Measures: Coprimary outcomes were autism screening practices and comorbidity management (assessed by medical record review). Secondary outcomes were knowledge (assessed by direct testing) and self-efficacy (assessed by self-report survey). Assessments were conducted at baseline, mid-ECHO, post-ECHO, and follow-up (3 months after ECHO). Results: Ten sites were randomized to 1 of 5 cohorts. Participants were 82% female (n = 108), 76% white (n = 100), and 6% Hispanic or Latino (n = 8); the median age was 46 years (interquartile range, 37-55 years). Significant changes in autism screening and treatment of comorbidities in children with autism were not observed. Participants demonstrated significant improvements in knowledge (9%; 95% CI, 4-13; P < .001) and self-efficacy (29%; 95% CI, 25-32; P < .001). Conclusions and Relevance: The ECHO model was developed to increase access to high-quality health care for underserved patients with complex conditions. Study results provide support for the model in improving clinician knowledge and confidence but little support for achieving practice change. Trial Registration: ClinicalTrials.gov Identifier: NCT03677089.