Zachary E. Warren, Ph.D. - US grants

Project Summary Autism Spectrum Disorder (ASD) is a common and costly neurodevelopmental disorder with core deficits related to social communication. There is growing empirical support suggesting that early intervention can significantly improve specific early social communication skills (i.e., joint attention and social orienting) and that such improvements partially mediate improvements in other critical developmental areas, including broad social and language outcomes. Unfortunately, many families struggle greatly to access appropriate, effective early intervention services at young ages due to resource limitations, including: limited access to services, delays and waits related to identification/treatment, a limited number of expert providers, as well as barriers related to the cost of treatment. There is an urgent need for the development of new tools and paradigms that can help advance access to powerful interventions across resource-strained systems. One potential new paradigm involves attempting to harness the abilities of sophisticated technological platforms to address ASD concerns. The primary goal of the current research is to refine, augment, and test the ability of an innovative, intelligent technological architecture and system specifically designed to address early social communication vulnerabilities in children with ASD. We explicitly do not propose this technology as a stand-alone intervention nor as a replacement for existing necessary intervention and care at later ages. Rather, we propose it as a critical test of a new technological tool and paradigm for thinking about advanced early detection and action regarding early social communication concerns and ASD. We will test the ability of our innovative three- dimensional intelligent learning environment to detect and meaningfully respond to core social communication behaviors evidenced during caregiver interactions. Simply, the environment will be designed to help young children more often respond to their names, follow gaze and common gestures, and coordinate social attention in interactions with their caregivers. Future rigorous study and understanding of this system?s ability to bolster these early pivotal social communication skills could lay the groundwork for radically different future intervention approaches. In the current work, we will first refine and augment our existing intelligent technological architecture, currently capable of detecting and inferring social attention in real-time as well as providing prompts based on this detection, to create a robust and powerful blended technological and social learning platform. We will then evaluate the ability of this technology to operate fluidly and autonomously with a clinical sample. The expected benefit of this research could be tremendous. It would, for the first time, allow for the systematic evaluation of an innovative, technologically-mediated ASD intervention platform for very young children that could readily translate into meaningful real-world use at a critical point in development.

PROJECT SUMMARY Although there is evidence that Autism Spectrum Disorder (ASD) can be accurately identified during the second year of life, many children in the United States are not diagnosed with ASD until after age four years. This is especially true for children from traditionally underserved communities, such as children from racial and ethnic minority groups, children whose parents report low levels of educational attainment, and children from rural geographies, whom providers might not even screen for autism concerns when they cannot subsequently link children to appropriate diagnostic services. This creates disparities in diagnostic identification and care that may have harmful, long term consequences for children, families, and service systems. Diagnostic visits offered by expert clinicians generally take place in tertiary care centers that present barriers to access of travel, time, and resources. These barriers may be extremely amenable to the use of telemedicine methods and practices. No explicit tools for conducting early telemedicine based consultations are currently available. This project introduces two novel telemedicine tools, the TELE-STAT and TELE-ASD-PEDS, for assessing autism risk in young children within their medical homes. These tools have explicitly selected as (a) the TELE-STAT represents a specific tool previously successfully utilized in-vivo within rapid triage and teleconsultation settings and to (b) the TELE-ASD-PEDS represents clinically-informed application of a computationally sophisticated analysis of observations tools used in comprehensive settings to diagnosis ASD. Under the supervision of an expert remote clinician, these tools can be used to coach parents and naïve pediatric providers via distance in how to elicit the behavioral features marked as most indicative of autism risk. We will test the accuracy of expert clinician telemedicine diagnosis utilizing these tools as part of evaluation. This work will be conducted in two stages. First, we will explore tool implementation, feasibility, acceptability, and preliminary accuracy in an already identified sample of young children with or without specific ASD concerns. Based on this pilot we will refined and adapt final versions of each. We will then implement and compare the tools across two groups of clinically referred children. We will do this in a simulated telemedicine diagnostic setting that will be immediately followed by a gold standard, in-person diagnostic evaluation with a different set of blinded clinicians. These tools have been designed to be low cost, to be compliant with privacy rules, and to meet the pragmatic and financial needs of many community provider networks. If successful, our telemedicine tools could provide methodologies that rapidly link children to ASD experts within practice locations where they are currently receiving care, in partnership with their existing providers. In turn, these children, who without such assessment may wait months or over a year to access assessment and intervention, may now be able to do so within days or weeks of screening as at risk.

PROJECT SUMMARY Although there is evidence that autism spectrum disorder (ASD) can be accurately identified during the second year of life, and that early intervention can improve developmental outcomes, many children in the United States are not diagnosed with ASD until much later. Families seeking ASD evaluation often face barriers such as low availability of specialists, lengthy waitlists, and long distances to tertiary care diagnostic centers. This is especially true for children from traditionally underserved groups and communities. Without an innovative approach for prompt identification of ASD in young children, families and clinicians will continue to struggle with accessing and providing care. Telemedicine offers tremendous potential for addressing this need. However, there are few psychometrically sound, validated tools that can be administered remotely, via telehealth platforms, in order to guide service and action. We recently developed and evaluated a novel parent-administered/clinician- guided tele-assessment tool, the TAP (previously the TELE-ASD-PEDS; R21 MH118539). Initial successful studies within a controlled laboratory context yielded very high levels of family satisfaction with the experience, perceived utility by clinicians, and agreement regarding ASD risk classification with blinded comprehensive evaluation. Although promising, this work was limited by its reliance on controlled laboratory settings, a relatively small homogeneous sample, and use within a single research group. We have not yet validated the TAP for in- home use in a broader community sample, with novel groups of diagnostic clinicians, nor have we adequately demonstrated its value and impact for families representing traditionally underserved populations and areas. The current proposal represents the first rigorous evaluation of a tele-assessment tool for ASD with a large, diverse sample of children within their homes. In the current study, across two sites, we will recruit 360 toddlers (18-42 months) to participate in a home-based telemedicine evaluation using an ASD assessment tool explicitly developed for this purpose (i.e., the TAP). The sample will include children from the community who have screened positive on an ASD screener, as well as those referred from primary care and state early intervention systems. We will randomize families to receive assessment based on telemedicine alone or telemedicine plus an in-person evaluation, and then follow families over a 6-month time interval. This methodology will allow us to conduct rigorous psychometric evaluation of the TAP and comparison to widely used diagnostic tests, longitudinally evaluate service access and family engagement, and examine factors that affect diagnostic decision-making and family outcomes. This work has potential to transform the ASD evaluation process and dramatically improve access for traditionally underserved groups.