One of the major goals in the healthcare industry for the next decade is the critical transformation from reactive damage control to proactive and personalized preventive care. This transformation can be illustrated by the P5 paradigm (Prediction – Prevention – Personalized – Participation – Public) in healthcare based on the seamless integration of the achievements in biology, technology, and computation to facilitate proactive screening methods and personalized preventive treatments.
The implementation of P4 systems boosts personalized medical instrumentation, enhances early diagnosis and prognosis for at-risk populations (e.g. critical care), lowers treatment cost, and subsequently reduces mortality and morbidity. While being adopted into the current healthcare systems, the full integration of P4 faces several challenges, including the insufficiency of analytical approaches to capture the complex dynamics of physiological and pathological mechanisms, the inadequate connectedness of patient-centered healthcare systems to efficiently harness the bio-signals and data, and the immaturity of analytic frameworks in providing reliable health indexes for personalized prognostics.
SPACHES LAb primary research interests encompass health informatics with a focus on connected healthcare systems and prognostic analytics to address the current challenges of the P4 paradigm. My interdisciplinary research directions, as shown in the figure below, include: 1) Data-driven and Sensor-based Modeling to characterize the coupling dynamics of the pathological processes via investigating the nonlinear lump parameter model of the biological processes driven by the collected sensor data; 2) Medical Device Manufacturing and Bio-signal Processing for deploying customized signals and data by considering the wearable, non-invasive, and point-of-cared designs with the integration of nonlinear bio-signal processing techniques and machine learning algorithms and 3) Predictive Analytics for Personalized Healthcare to forecast acute event onsets by qualifying the transition of the system dynamics from the normal to abnormal conditions via time series prediction models integrated with nonlinear dynamic system approaches.
The research settings range widely, from in-vitro data (from simulated data) and in-vivo (from online databases) to human subject data. Hence, the remaining goals are: to build systems and methods utilizing bio-signals and data, subsequently to process and analyze collected data for better understanding of the system dynamics related to disease diagnostics, and eventually to predict the onset, progression of and recovery from pending deterioration, which can in turn be embedded in the systems to help doctors make better informed decisions related to preventive treatment.