Can a Novel Telemedicine Tool Reduce Disparities Related to the Identification of Preschool

Innovative telemedicine practices could address many existing traditional barriers to ASD identification. Telemedicine has the potential to put linguistically and culturally competent clinical expert virtual providers directly into communities during a critical window in which many vulnerable families may approach, or re-approach, care systems about developmental concerns that become more prominent with age. In the current proposal, the investigators will evaluate the potential clinical and familial value of a telemedicine-based ASD assessment tool, the TAP-Preschool (TELE-ASD-PEDS-Preschool), designed to overcome traditional barriers to diagnosis and service access in underserved preschool populations that may not be readily identified or engaged by early screening and intervention systems.

A growing body of literature supports the use of telemedicine-based approaches to ASD assessment and intervention. This includes remote activities to assess infant social communication skills in the first year of life, coaching parents through administration of gold standard diagnostic tests, and provider coding and analysis of interactions that are video-recorded by caregivers. Results reflect high levels of caregiver and provider satisfaction and satisfactory agreement with traditional in-person evaluations, illustrating both the promise and feasibility of tele-assessment. However, most existing approaches are limited by a focus on screening rather than diagnosis, protocols requiring specific materials, or asynchronous analysis of submitted videos, which require resources that preclude broader use. Moreover, in current form, these tools are not designed or intended to yield a quantitative formal assessment of core ASD symptoms to support diagnostic decision-making. Further, although providers are increasingly exploring telemedicine approaches to ASD assessment in the context of the COVID-19 pandemic, there is limited published work to date on the use of real-time, caregiver-led ASD assessment in home settings with diverse populations. Prior work from the investigators' team has focused almost exclusively on toddlers in this regard. Ultimately, very few viable tools with validated psychometric properties are available for use in current models of telehealth assessment and care.

In earlier work, we applied machine learning models and the principles of feature engineering to a phenotypically rich clinical research data set of preschoolers with ASD and other developmental concerns (N=914: 594 ASD and 320 Non-ASD) to identify key behavioral targets for tool development. We then engaged in a rigorous adaptation and translation approach to optimize design of the TAP-Preschool. We included leading ASD assessment experts, providers dedicated to caring for underserved populations, and parents of preschool children with ASD from underserved racial/ethnic and linguistic groups. We lead these groups in a collaborative design process in order to yield a set of interactive and play-based, parent-coached tele-assessment activities designed to 1) elicit observations tied to key computational features, 2) be deployed within a 30-minute timeframe, 3) employ inexpensive materials found in most homes, and 4) use accessible assessment instructions for real-time coaching of parents.

The investigators will now deploy the preliminary TAP-Preschool with a small sample (n = 30) to assess acceptability/feasibility, potential clinical value for remote observation, and challenges that warrant revision. These data will then be used to modify the TAP-Preschool and the refined tool will be deployed with a new sample of clinically referred children (n=120). The investigators will evaluate its ability to facilitate accurate telemedicine supported diagnostic decision-making.

Initial deployment, evaluation, and refinement of the TAP-Preschool: