According to the CDC, autism is the fastest-growing developmental condition, with 1 in 44 children diagnosed in the United States. A fundamentally heterogeneous population, individuals vary in severity in communication and behavior. Comorbid medical concerns lead to a higher number of clinic, pediatric, and psychiatric outpatient visits, and hospitalizations when compared to children without autism. However, despite increased prevalence, disability and medical complexity, and healthcare utilization, children with autism receive an overall lower quality of patient- and family-centered care (PFCC). Kouo et al. (2021) and other researchers have reported families’ frustrations related to the lack of knowledge and strategies healthcare providers (HCPs) have concerning the individual complexity of autism. However, it is important to recognize the constraints of HCPs and healthcare settings. Providing care to this population is challenging due to stressed healthcare systems, limited time and resources, and deficient training in caring for patients with autism. Consequently, HCPs have reported concerns about effectively managing behaviors (e.g., self-injurious behaviors, aggressions) to prevent the overuse of pharmacological sedation or physical restraints to respond to such behaviors. Ultimately, HCPs are seeking applicable tools to better respond to the needs of this unique patient population, while also considering the constraints of healthcare systems.

Next Steps of the Project

The CARE app was originally designed for and piloted in the pediatric emergency department. Pilot findings and feedback indicate that HCPs require added support in using the Snapshot data to select the appropriate EBPs. Given their limited training on autism-specific EBPs, limited time during medical visits, and increasing census of patients seeking care, HCPs need a rapid, automated system that recommends specific EBPs to implement based on the patient profile. The application of machine learning (ML) is an innovative solution that has not been applied in this manner in healthcare and special education, and can significantly increase the impact of the CARE app.

Therefore, this site aims to gather data on individuals with autism and data on the impact of EBPs being implemented to inform and train a ML algorithm of effective and ineffective EBPs based on the unique profile of the individual with autism and additional contextual information. The aim would be to gather a minimum of 2,000 patient profiles via this site. Ultimately, this project will lead to the development of an accurate and comprehensive dataset that is not in existence and provide preliminary data to train a novel ML algorithm that can effectively recommend responsive and individualized EBPs to be implemented.

Expected Outcomes and Future Work

Understanding that the CARE app can evolve further through innovative solutions, the application of ML aims to support HCPs given their realistic constraints. Furthermore, the impacts can be broadened to support the practices of educators in classroom settings and families in the home and community, as this project aims to improve the speed and accuracy at which EBPs can be identified to be responsive to the needs of this highly heterogeneous population.

Your help would support the novel application of ML in healthcare, education, and in disability research. No dataset or solution currently exists to decrease the time and efforts needed to effectively identify EBPs to address the unique, individual needs of individuals with autism.