Reliability, Accuracy, and Timeliness of Autism Telehealth Evaluations by Pediatricians

Early diagnosis of autism spectrum disorder (ASD) can facilitate timely intervention and improved developmental outcomes, but many children wait over a year for ASD evaluation. Telehealth assessments by pediatricians may address geographic and workforce barriers for children awaiting ASD evaluation. The goal of this study was to assess the reliability, accuracy, and timeliness of ASD telehealth evaluations by pediatricians. We hypothesized this approach would demonstrate reliability with specialist evaluation, concordance with in-person diagnosis, and wait-time reductions.

Thirty-two pediatricians received standardized training, completed fidelity/reliability testing, and utilized a novel evaluation model employing validated screening, parent-interview, and child observation tools. Interrater reliability of 32 pediatricians who administered 200 telehealth cases was assessed relative to blinded clinical supervisors. The diagnostic accuracy of 27 pediatricians who watched 19 video-recorded telehealth cases (494 total assessments) was compared to in-person diagnosis for the 19 children. Wait time for a telehealth assessment was compared to wait time for an in-person assessment for 2,483 children.

There was 91.5% scoring agreement (95% CI: 0.893-0.932) between pediatricians and clinical supervisors (p < 0.001, K = 0.206) on 200 cases. On 19 video-recorded cases, there was 93.5% accuracy (PPV = 0.957; NPV = 0.811) between the in-person diagnosis and the 494 diagnoses provided by 27 pediatricians. The average wait time for 2,483 children who received telehealth evaluations (11.7 days) was shorter (W = 3.08 × 10⁶, p < 0.001) than the average in-person wait time (11.8 months).

These results demonstrate that telehealth evaluations by general pediatricians offer a reliable, accurate, and timely approach to ASD assessment through a scalable, nationwide model.