Personalized predictive care-at-home

Kuldeep Singh Rajput

Problem Tech Solves

In the very early weeks of the COVID-19 pandemic, we were contacted by research partners at the Queen Mary Hospital in Hong Kong inquiring if we had a remote patient monitoring (RPM) solution to help their clinicians more safely manage patients with the virus. Although we did not have such a solution at the time, we recognized the urgency of the situation. Our Biovitals platform is modular, and our engineers worked day and night during the next 2 weeks to build a solution that manages patients with COVID-19 and other acute conditions called Biovitals Sentinel. Like all Biovitals solutions, Sentinel is powered by the FDA-cleared Biovitals AI-powered Analytics Engine®, which learns the correlation between multiple vital signs during the patient’s daily activity and builds an individualized biometric signature that is dynamically and continuously updated based on incoming data. The engine then computes and continuously updates the Biovitals Index which reflects changes in the patient’s measured vital signs from their measured baseline. Due to the abundance of data available from the biosensors, the Biovitals individualized biometric signature for each patient produces more meaningful and actionable insight, alleviating the clinician from non-actionable notifications and irrelevant decision support messages. This technology has enabled patients with known or suspected COVID-19 to stay in their own homes to be monitored for the virus, with the technology flagging any need for intervention (including hospitalization) early on. The power of this technology meant more patients could stay in their own homes, with improved QOL.

Tech Brief

Biofourmis' end-to-end, turnkey solution includes an AI-based analytics platform; a medical-grade armband wearable that the patient wears 23 hours per day for continuous, real-time monitoring; a clinician dashboard; and a patient-facing app. The more the Biovitals(R) platform is used, the better it gets. At its heart, Biovitals Sentinel is an advanced and constantly evolving machine-learning platform that gets smarter and more effective over time. Each new patient treated using the technology adds valuable health data to our vast population datastore—at the rate of 20+ physiological signals per second. That means clinicians and researchers who use Biofourmis have a wealth of personalized, real-time patient data at their fingertips to help detect medical exacerbations so they can intervene earlier to keep patients safer, to improve outcomes, and to save costs. This capability has proved extremely valuable during the pandemic, since early interventions can save lives when it comes to COVID-19 disease progression. A personalized treatment algorithm runs patients' data against billions of data points from other patients with the same chronic or acute illness. The platform looks for patterns, so doctors and other clinicians can move beyond simply detecting serious medical events and into predicting and avoiding them, well in advance. That's the power of what Biofourmis calls personally predictive care.

Tech Differentiators

Biovitals Sentinel provides a solution at the intersection of care delivery, software technology, and artificial intelligence/machine learning. The continuous monitoring capabilities of our solution, the AI-based analytics and machine learning, and the ability of our medical-grade wearable to collect 20+ physiology parameters make our solution stand out. We are also device agnostic if clinicians prefer to use our analytics solution with their own biosensors. Many competing remote monitoring solutions rely on patients to input their own data, whether via bluetooth-enabled home medical devices or completely manual entry, and that data is based only on moments in time rather than 23 hours of monitoring per day that Biovitals Sentinel offers. In addition, Biovitals Sentinel reduces administrative burden for a patient's care team because it creates a personalized baseline for each patient within 6 hours, meaning it will only notify clinicians if there's a clinically significant change that may require intervention. The technology can notify clinicians up to 21 hours before a serious event might have otherwise occurred, enabling the clinician to take steps to prevent a medical crisis. Our solution benefits hospitals and large healthcare systems by improving clinical outcomes and reducing costs, while also improving patient satisfaction. Specific to the pandemic, Biovitals Sentinel's remote monitoring capabilities not only enabled patients to stay in their own homes when possible, it kept healthcare professionals safer during a time when 30-40% or more were testing positive for the virus.


Results of a study showing that Biovitals Sentinel delivered more accurate predictions across viral load, decompensation and hospital discharge was published in February this year in the journal Nature: The study concluded that Biofourmis' AI-powered platform for remotely monitoring patients with coronavirus, Biovitals® Sentinel, delivered more accurate and timely predictions about health decompensation than traditional in-person, manual methods. In the study, 34 patients with diagnosed mild COVID-19 were remotely monitored with Biovitals® Sentinel. A biosensor worn for 23 hours a day on the arm captured vital signs and patients were asked patients to also report symptoms into a smartphone app. For comparison, the same patients were monitored through traditional, in-person, manual vital signs capture. To measure decompensation, clinicians relied on the Biovitals® Index, a continuously generated, machine learning-derived health index that reflects the overall health status of patients. Clinical status was also assessed every morning using the National Early Warning Score 2 (NEWS2), a standardized, manual clinical evaluation endorsed by the British National Health System to early detect clinical deterioration in patients with sepsis. Researchers found that the Biovitals® Index was superior to NEWS2 at detecting increased viral load after taking respiratory samples from patients each day. The platform’s deterioration prediction accuracy rate was 93% with an average prediction time of nearly one full day (21 hours). The NEWS2 prediction accuracy rate was only 64%. Likewise, Biovitals® Sentinel was 78% accurate at predicting the patients’ length of stay while NEWS2 was only 58% accurate.

Why Us

Biofourmis Sentinel was one of the early technology solutions on the scene, hitting the market in February 2020, due to the modular build of our technology and the commitment of our team. Peer-reviewed studies demonstrate the effectiveness of our technology.

❮ Return to FULL LIST OF FINALISTSView Next Company❯