Training and Sustaining: Training and Learning Collaborative Outcomes Across a Statewide Network for Early Autism Diagnosis.

Abstract

Objective: The objective of this study was to describe the development of a primary care professional (PCP) autism diagnosis training model and to report on outcomes related to PCP training and sustained engagement in a longitudinal learning collaborative.

Methods: We developed Accelerating the Diagnosis of Autism with Primary care Training (ADAPT), a training program to prepare PCPs to develop independent competency in evaluation of autism in children aged 14 to 48 months. ADAPT includes didactic and case-based modules and practice-based coaching delivered by an expert diagnostic specialist; after training, PCPs participate in a longitudinal learning collaborative. Aligned with competency-based medical education standards, measures of autism evaluation knowledge and diagnostic competency are collected.

Results: From 2021 to 2023, 13 PCPs completed ADAPT didactic and practicum training to reach competency in independent autism evaluation. Clinicians demonstrated significant improvement in total autism knowledge after didactic training (p = 0.02). Scoring agreement on an autism observational assessment tool between clinicians and expert diagnosticians improved over case observations and practicum evaluations. Similarly, PCPs demonstrated improved evaluation competence, moving on average from Advanced Beginner to Competent Performer as rated by expert diagnosticians. After training, PCPs attended 57% of monthly learning collaborative sessions.

Conclusion: Training PCPs to deliver autism evaluations for young children as part of tiered community-based models of care is a promising solution to address access and waitlist challenges. ADAPT is an intensive, standardized PCP training model that results in achievement of independent competency and sustained engagement in autism evaluation. Effectiveness-implementation studies are needed to ensure scalability and sustainability of training models.