This Small Business Innovation Research (SBIR) Phase II project will build an end-to- end platform around the ambulatory monitoring device proposed in Phase I, for continuous health monitoring of a human. The Phase I wearable device measures multiple noninvasive biosignals from a person in their daily home routine (or in the hospital), providing unprecedented visibility into health or disease status outside a critical care setting. The Phase II platform will comprise an "ecosystem" of software for providing automated, scalable intelligent monitoring of the signals from the device using advanced machine-learning algorithms, and exception-based alerting of medical staff upon early indication of deteriorating health of an ambulatory patient.
If successful this platform will provide a substantial improvement in the capability of the healthcare system to proactively manage the health of the large population of patients with costly chronic diseases. Current methods for remote (home) patient monitoring "while better than a complete lack of monitoring" involve extremely sparse data (once per day) and require proactive patient compliance to make manual measurements, typically of weight or blood pressure. These methods do not handle ambulatory variation; in contrast, the proposed algorithms uniquely detect health anomalies otherwise hidden in ambulatory variation. This Phase II project not only has the potential to fundamentally improve healthcare with continuous automated visibility into patient health in the home environment, but also stands to provide unique insight into new signatures of disease heretofore not recognized by medical science. The advanced detection algorithms are able to learn empirically the normal physiological variation (e.g., variations in blood pressure, metabolic activity, etc., throughout the day) of the human system, and reveal incipient anomalies from normal behavior which are not visible upon a plain, univariate inspection of the data. Moreover, the device itself provides data from human activities not customarily encountered in the static conditions of a medical facility, where patients are supine and sedated. It is highly likely that this new approach to multivariate analysis of human biosignals will unveil new signatures providing early warning of disease progression, for example, decompensation in a heart failure patient.
