An Automated Telehealth System for Long-Term Monitoring of Meltdowns in Children with Autism Spectrum Disorder: Epxautism.

Background:  For patients with autism spectrum disorder, wait times often exceed three months for scheduling specialty clinic visits. Moreover, current attempts to ascertain data on longitudinal behavior patterns in autistic patients during this interim rely primarily on retrospective interviews or questionnaires, which suffer greatly from recall bias. The lack of accurate and effective monitoring of clinically relevant behaviors such as meltdowns can lead to less than optimal care provided to the patient. We have created an automated text-message-based intervention to bridge these temporal gaps in care, giving providers data and automatically triaged alerts on patients’ behaviors between visits.

Objectives:  To monitor patients with autism through a novel telemedicine system which facilitates early detection of exacerbation of problem behaviors, providing a tool for healthcare providers to improve treatment recommendations and medication titration.

Methods:  EpxAutism sends daily/weekly messages to parents of children with autism asking about quantity and duration of meltdowns, instances of aggression/disruption/self-injurious behavior during meltdowns, and whether there is a known antecedent to meltdowns. The system synthesizes data received through text responses for each patient, triages patients into one of three risk categories based on their change in behaviors, and sends alerts to providers upon recognizing significant upward deviations in recorded problem behaviors.

Results:  We achieved an 85% overall response rate to messages through an initial phase of pilot study involving 6 enrolled patients. Each patient has a diagnosis of autism spectrum disorder. Duration of enrollment spans from 18 days to 110 days. Response rate has not decreased significantly for patients enrolled for more than 100 days.

Conclusions:  We present EpxAutism, a novel, cost-effective monitoring system for behaviors of children with autism between clinic visits. Prompt reporting both reduces recall bias as compared to paper surveys/retrospective questionnaires and allows providers in this study to remotely alter subjects’ plan of care. Data suggest that our intervention is a promising method of tracking meltdowns and problem behavior in patients with autism, allowing physicians to more precisely treat patients and also be alerted more promptly about ongoing behavioral problems. Results from our pilot warrant further study in a larger population, which is ongoing.