Screening Tool for Autism Risk Using Technology – START: Developing a Tablet-Based Platform to Detect Autism Risk in Low-Resource Settings

Background: Early diagnosis is key to early intervention and thereby to best outcomes in Autism Spectrum Conditions (ASC). In low-resource settings such as India, lack of access to early detection poses a significant barrier. Latest WHO data on India show that the ratio of specialists (psychiatrists) to patients can be as low as 0.30 per 100,000 people. In rural areas, furthermore, psychiatric expertise is non-existent. To fill the gap between the need for early identification and limited availability of expertise, it is important to develop alternative methods and tools that can be used by a low-skilled workforce, are inexpensive, and are portable so that they can be taken to the field without requiring any additional infrastructure.

Objectives: The “Screening Tool for Autism Risk using Technology” (START) project is developing an open-source, scalable autism screening tool in the form of an app that can be run on a tablet computer with minimal training.

Methods: This project is an interdisciplinary collaboration between neuroscientists, mental health professionals, public health researchers, computer vision scientists, app designers and app developers. Eight institutes/organisations (University of Reading, Birkbeck University of London, Nottingham Trent University, Indian Institute of Technology, All India Institute of Medical Sciences, Public Health Foundation of India, Sangath, and Therapy Box Ltd) from three countries (India, UK, and USA) participate in this consortium, funded by the UK Medical Research Council Global Challenge Research Fund. The project is divided into three work streams. The first stream is responsible for putting together a battery of behavioural tasks suitable for children between ages 2-5 years, to capture social, cognitive, and sensory-motor aspects of development relevant for autism. The second work stream is responsible for building a scalable pipeline to carry out data quality checks, data transfer to a central server, and maintenance of the research database. The third work stream will undertake field testing of the tool in India, collecting data from children with and without autism for validation.

Results: The app includes parent-report questionnaires and behavioural tasks to measure social preference, motor function, and sensory interest. A key innovation is tablet-camera based eye tracking, allowing capture of very low-resolution gaze direction data for coarse measures such as preferential looking and attentional disengagement, without the need for an external eye-tracker. Motor function tests are built in light of the emerging data from similar tablet-based tasks on the iOS platform (Anzulewicz, Sobota, & Delafield-Butt, 2016; Belmonte et al., 2016). The app also includes a task measure of social motivation previously associated with a group difference in children with autism (Ruta et al., 2017). Finally, the app includes a free-play section that provides an opportunity to video-record parent-child play interaction for subsequent analysis.

Conclusions: The working prototype establishes feasibility of a portable platform that can capture multiple task and interview measures, including eye-tracking and assays of motor function. The platform integrates mobile health technology for task-sharing application in low-resource settings by non-specialist community health workers, and offers the potential for scaling up as part of regular community health assessments.