(332a) (Invited Plenary Talk): Towards Ubiquitous Physiological Monitoring

Modern medicine is largely reactionary. We wait for patients to exhibit symptoms and then we try to treat them. Yet we know physiological changes precede clinical deterioration. Two of the most basic and critical physiological parameters are respiration and hemodynamics. Even short interruptions of either have catastrophic consequences. Remarkably, there is currently no way to automatically and continuously assess respiratory effort, minute ventilation, or even blood pressure.

In response, we have developed Band-Aid© like sensors for continuous and quantitative monitoring of both respiratory and hemodynamic waveforms. Using commodity shrink film to create nanostructured thin metal films, we have developed a lift off process to transfer these thin films into any elastomeric polymer to create scalable sensors for physiological monitoring. From these waveforms, we can continuously monitor key parameters – and importantly, how these parameters change in real time – including respiratory rate, respiratory flow, heart rate, and blood pressure. Both patches can be applied and left on for > 24 hours, and has been correlated with FDA approved medical grade devices (spirometer and finger-cuff volume clamp beat to beat blood pressure monitor, respectively). The biggest advantage of our sensor system compared to other solutions (e.g. pulse oximetry, capnography, blood pressure cuff, EKG leads) is its ability to provide automatic and continuous untethered ambulatory monitoring to alert of early indicators of acute distress. Additionally, because our sensors are compatible with roll to roll mass manufacturing, they are extremely inexpensive to produce and can be disposable.

The data from the wireless sensors can be transmitted wirelessly to a tablet, phone, or workstation. With these sensors, we have various human subject tests in different clinical populations including congestive heart failure, stroke, and asthma patients. We are working on analyzing the data using a variety of methods to uncover promising digital biomarkers. Real-time machine learning models are being used to identify and classify the signals.

We are also developing complementary sensors for sweat and saliva testing to provide a more holistic approach to patient monitoring. We are working towards syncing all the data together to provide comprehensive and real time information of the patient's well-being, to alert of (and eventually hopefully prevent) potential deterioration.