How It Works
Personalized data analytics can help keep seniors safe so they can stay in their in their own homes as they get older. But how can relevant data be collected for people who don't carry smartphones, aren't interested in wearables, and may not worry about taking care of themselves?
Reliable Monitoring without Inconvenience
Traditional approaches to monitoring have relied on intrusive cameras, undesired wearables, and unwanted (and unreliable) behavioral changes. EmPowerYu is different. We place home security type sensors in the home to detect events that usually form patterns, like eating and sleeping, or watching a favorite TV show. When the events are reliable, like Grandmother always watching her TV around 5pm, the system assigns that event a high reliability score and knows when to expect that event. The system constantly compares the current activities of the day against Grandmother's historical pattern of activity for that day and time. When a reliable event is missing without an explanation (for example, activity elsewhere), we detect an abnormality and notify a caregiver.
Power of an open platform
Technology changes very quickly these days, and EmPowerYu is ready to adapt. Many companies look for applications for their existing hardware. EmPowerYu decides what data we need, then finds the best sensor to get that data.
Rather than relying on proprietary hardware that could become obsolete, we provide an open platform that works with a wide variety of commercially available devices. We partner with an industry-leading digital health platform, which supports more than 400 different medical and fitness devices. And as new devices or new generations of devices become available, they can be incorporated into the EmPowerYu ecosystem. This future-ready approach gives users and care providers the freedom to switch or expand the hardware, while maintaining the power of a continuous data set.
Anomaly detection through machine learning
What’s normal? What’s not? These simple questions can be quite tricky, but caregivers no longer need to struggle to answer them. EmPowerYu’s advanced anomaly detection analyzes complex data sets, identifies patterns*, and determines the difference between normal and abnormal conditions, while adapting automatically to changing circumstances. A calendar overlay supports different definitions of what’s normal for different days, and caregivers can adjust settings for alerts to minimize alarm fatigue.
*In technical terms, our machine learning algorithms apply distribution-independent measures of variance, and outer likelihood of range to determine what constitutes the normal pattern, and to detect when an anomaly occurs.
Activity Status at a Glance
The current status of a monitored person is presented to authorized caregivers using a simple display. A summary green checkmark indicates all patterns match the normal profile (everything is OK). If a significant change or missing event is detected, the status color changes to orange and an icon indicates the source of the possible problem. Alerts are sent according to the caregiver's selected settings. Healthcare and homecare providers can set specific alerts for personalized responsiveness and early intervention if needed.
More technically, EmPowerYu is a software analytics platform that uses behavioral state modeling, an inference engine, and sparse signal processing to understand the personal patterns of daily living, and calculates a Functional Living Signature™ for each person. We quantify behaviors like eating, sleeping, daily activity, and device usage, to show family and professional caregivers the status of a person at home. Current data are compared to personal historical patterns (for wellness monitoring) or pre-set thresholds (for post-acute recovery monitoring) to detect anomalies and generate alerts. We also analyze trends over time; for example, we can determine a baseline pattern before a planned medical intervention, then quantify progress during recovery. Or we can detect drift of sleep time, or change in the average time it takes to move between rooms.
Mrs Mac's behavior and her environmental readings are normal, but she has a medical reading outside of the defined normal range.
A concern about automated alert systems is the risk of inaccurate alerts, errors, or repeated alerts that could waste caregivers' time or cause unnecessary distress. EmPowerYu is pioneering Televalidation™, an innovative call service that will reach out to the care recipient and confirm the alert before proceeding to Alert Management protocols.
John's readings are all within normal range, so he gets a big green checkmark.