In recent years, commonly-accepted statistics have estimated that approximately one in sixty-nine children will be diagnosed with Autism Spectrum Disorder (ASD). While the number of therapies and treatment plans available to families of such children continue to grow, so does the amount of data required to initiate, sustain, and evaluate several treatment options. This data poses a burden to parents, therapists, teachers, and physicians alike, who must both collect and manage data, and in the case of medical professionals, also successfully analyze it to refine treatment strategies. While computational tools such as data mining and machine learning have been leveraged in ASD research, the majority of these applications have been in the areas of diagnosis, assistive and educational technology, and drug discovery. For those tasked with caring for and treating children diagnosed with ASD, effective data management continues to be a challenge, especially as the volume of data generated grows, aided in part by the realization of ubiquitous computing.
To ease the burden of collecting and managing data in support of the treatment of autism spectrum disorder, we have built AMP (Autism Management Platform), a health care information system consisting of a mobile application, web client, secure data repository, and analytics engine. Together these components simplify the means by which multimedia data can be captured, disseminated, navigated, and analyzed by caregivers and clinicians. The result is a fully integrated, intuitive system that aims to improve data sharing and facilitate data mining, with the ultimate goal of streamlining the treatment process for families and professionals alike. Additionally, because AMP has been designed as a modular platform, features can be easily augmented and evolved in response to new data needs, new users, or new perspectives on treatment plans. Though still in its infancy, AMP has generated interest from parents, therapists, teachers, and clinicians who have grown frustrated with antiquated paper-based and electronic methods that fail to take full advantage of data generated during the course of treatment.
The figure below provides an overview of AMP’s architecture, which is currently deployed in a fully-functioning beta version that implements the features we feel are most important in improving the data management and mining process for clinicians and caregivers. Each of AMP’s components provide several avenues for additional research, but my group is currently focusing on improving the mobile application, which has not undergone user testing, and needs to be validated by real-world users in clinical and educational settings before AMP can be widely leveraged in the treatment of ASD.
AMP is significant in that it is the only data management and analytics platform to date that allows real-time collaboration between caregivers and clinicians involved in ASD treatment, while at the same time providing big data analytics capabilities. The intent is to also make AMP open source software at the conclusion of our research activities, with the hope that making it freely available will have a tangible impact on ASD caregivers who are currently lacking effective tools for data collection, sharing, and analysis.