Talk Highlight

Bio

Jaideep Srivastava is a Professor in Computer Science and Engineering and in the Hubbard School of Journalism and Mass Communication at the University of Minnesota. He is a Distinguished Visiting Professor at the Gangwal School of Medical Sciences and Technology at IIT Kanpur. His research is in the application of AI for Healthcare and Computational Social Science. He is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE), has been an IEEE Distinguished Visitor, a Distinguished Fellow of Allina's Center for Healthcare Innovation, and has been awarded the Distinguished Research Contributions Award of the PAKDD, for his lifetime contributions to the field of machine learning and data mining. He has authored over 450 papers and advised over 50 PhDs. A number of his papers have won best paper awards. His research has been funded by various US federal agencies, as well as the industry. He has held advisory positions with the State of Minnesota and is advisor to the Aadhaar project of the Government of India, the digital governance platform which provides biometrics-based identification to the 1.40+ billion citizens of India. He has significant industry and entrepreneurial experience, having advised over a dozen large corporations and multiple startups. He has been on the founding teams of three startups, with experience in raising capital, product building, and acquisition. He received his BTech from IIT-Kanpur, andhis MS and PhD from UC Berkeley, all in computer science. 

Abstract

We envision a future care scenario where a patient's health status is being continuously monitored, with a range of care types - tele-visits, home visits by professionals, hospital visits by patients, and family members responsible for care; each providing information as needed. This will allow care to be targeted and timely, with the type of care decisions made by the care team using data-driven tools vs. the present family/patient-driven decision to visit the provider. This will not only make care provisioning timely and appropriate, but will also reduce cost and effort for both patients and care providers, and improve care access. We call this approach continuous care at home (CC@H). Success of CC@H requires accurate models to predict general patient deterioration (e.g. frailty) or impending critical clinical events (e.g. cardiopulmonary arrest). In addition, the prediction should be timely, providing sufficient time for proactive and preventive interventions at home or the hospital. Potentially the biggest blocking factor for achieving the vision of CC@H is the provider team's lack of visibility into what is happening in their patient's life once they have left the hospital and go about their daily lives. Visits back to the hospital/clinic are episodic, triggered by the patient or a family member, neither of them with expertise in making this triaging decision. Some triggers are timely, e.g. the patient not feeling good, but many are not, e.g. a medical emergency requiring an ambulance and ICU. At each provider visit, information about life at home since the last visit is collected. Unfortunately, it is often a (questionable) qualitative recollection of daily life and significant events. There are attempts towards systematic collection of this data, e.g. via questionnaires, but the information provided is usually of low accuracy and outdated. The revolution in (i) body wearable sensors, either generic commercial ones like Apple Watch/Fitbit/all.health, medical grade ones like Actigraph wGT3X-BT, (ii) other home devices like smart beds and smart weighing machines, and (iii) medical devices that can both record and send mobile cardiac telemetry data (for those requiring them), provides the opportunity for detailed, quantitative visibility into a patient's life and medical condition outside provider visit. This, in turn, opens up the exciting and transformative opportunity to move from the current episodic care approach, which is reactive, to a proactive, targeted and personalized continuous care approach. While the overall vision is broad, our objectives for the proposed project are narrower, focused on the problem of obstructive sleep apnea (OSA), which, if left unchecked, can have consequences for cardiac health and Alzheimer's. We present results from our work, which is being done in collaboration with hospitals (MHealth/Fairview, Allina and Mayo Clinic), a biomedical instrument manufacturer for OSA (Inspire Medical) and a Smartbed manufacturer (Sleep Number). We will also present the vision and status of a much larger effort in Digital Health, led by the University of Minnesota, and involving medical device manufacturers (Medtronic, Boston Scientific, various companies in the Medical Alley), hospitals (UMN, Mayo, Allina, etc.), insurers/payers (United Health Group/Optum, etc.), and with appropriate participation from private, State and Federal agencies.