Author: Walt Scacchi
When AI approaches to medicine rely on cloud-based big data for advanced analytical applications like precision chronic care decision making and self-managed telerehabilitation patient care at home, a core challenge is how to collect, organize, and interpret such data. In most situations, available data is heterogeneous, physically dispersed, and off-line. AI approaches like machine learning or cognitive computing excel when data are homogeneous, logically centralized, and online. So before AI approaches to long-term, personalized chronic care in home-based settings can be established, we need a scalable telerehabilitation infrastructure for experimentation, clinical trials, and evidence-based assessment that can reach from individual patient homes, to clinics and hospitals where physicians, therapists and others can monitor, measure, and adaptively tailor precise, patient-specific therapeutic care. Our efforts focus on development, deployment, and adaptation of a new generation of telerehabilitation care systems for stroke. In the U.S., stroke annually gives rise to 800K+ new stroke survivors, who generally live 6 years post stroke. In our view, stroke is an exemplar of a chronic disease that assumes self-managed chronic care that spans both clinical and home-based therapeutic rehabilitation treatments that are personalized to the individual patient, as well as monitored and clinically managed by therapists, attending physicians or others. Other chronic diseases or conditions that fall into this category include asthma, diabetes, neurodegenerative ailments, obesity and others, which often indicate different patient demographics. We further have chosen to develop and employ activity-focused, therapeutic computer games that are both instrumented to collect high-precision therapeutic activity-exercise compliance data, as well as designed to intrinsically motivate patients to engage and sustain their playful game-based therapy. We are currently conducting a nationwide, Phase II randomized clinical trial of this game-based telerehabilitation system consoles through a network of ten hospital partners and forty consoles. The consoles are deployed in patient homes, and connected through secured data network infrastructure to capture and provide patient-specific data streams of their game-based therapeutic activity, as designed and configured by the patient's physician and therapists. As individual patient performance, recovery, or decline data is logged and uploaded into our data analysis servers at UCI School of Medicine, our purpose-built software systems analyze, visualize, and report patient performance results to their designated physician-therapists team. Reports provide algorithmic recommendations on game selection that may improve patient recovery, based on standardized (upper extremity motor control) performance measures. However, the attending physician and therapists make final decisions regarding what therapeutic games or standard therapeutic exercises will be assigned to the patient. This presentation will describe the design, development, and deployment of our game-based telerehabilitation system and infrastructure for chronic stroke patient care, identify where and how we collect and analyze patient care performance and efficacy data, and how these capabilities help establish a foundation for future advances in AI in Medicine through cloud-based, big data infrastructure. We also identify how new technologies like Virtual Reality and Augmented Reality capabilities further complement telerehabilitation across chronic care diseases.
Co Author/Co-Investigator Names/Professional Title: Steve Cramer MD, Professor of Neurology; Anatomy & Neurobiology; Physical Medicine and Rehabilitation, UCI School of Medicine
Funding Acknowledgement (If Applicable): NIH 1U01NS091951-01A1; Telerehabilitation for Patients with Stroke (Steve Cramer, PI)