Satish Movva is the Founder and CEO of CarePredict, a company that uses Artificial Intelligence to help seniors live independently, economically and longer by triggering just in time care for elders, anywhere.
In this interview, Mr. Movva shares insights on how AI will become more and more central in areas of healthcare system such as senior care, remote monitoring, and disease prediction.
What is your background and how did you start CarePredict?
Prior to CarePredict, I founded ContinuLink, the first SaaS Homecare and Hospice platform and took it to commercial success before an exit. I was fortunate to have more than my share of firsts by creating the first mobile EMR device on a Palm Pilot in 1998 and the first SaaS EMR for Neonatal ICUs in 2000. In total, I have over 35 years of experience in technology, 28 of which have been in healthcare technology.
I founded CarePredict to meet my needs as the primary caregiver to my 81 and 91-year-old parents who live independently a few miles away from me. I call them every day and visit them once a week. Five years ago, I found most of my weekly visits were being spent in the hospital or ER waiting rooms. The unpredictability surrounding my parent’s health was leading to unpredictability in my life as I was juggling a young family and a career as well. I began to look for a solution that would provide me with early insights into their well-being. There were only two types of solutions available in the market. One was the personal emergency response system (PERS) which only alerted me after something had already happened such as a fall and the second was ambient monitoring systems that used repurposed burglar alarm components such as motion sensors and door/window switches on everyday objects like a refrigerator door and toilet flush handles to detect activity and infer wellbeing. Unfortunately, these didn’t work for my situation where I had two parents living in the same house and motion sensors could not tell me who was doing what. And more importantly, they were unreliable as I could not assume that my parents ate, just because the system told me the refrigerator door had opened and closed.
So, I decided to create a solution that would let me know when I had to intervene before something happened. I knew that if I had caught the signs of decline earlier on, I could have prevented declines in my parents’ health by proactive action. I observed that the first signs of changes in their health often appeared as changes in their daily activities and behaviors. In order to catch these subtle changes as soon as they appear, I would have to continuously observe their everyday activities such as eating, sleeping, walking and toileting. That’s what CarePredict does: allows a senior’s activity and behaviors to be observed autonomously and without any self-reporting to find the earliest signs of a decline and give the family enough time to act.
What is CarePredict’s mission?
Population aging is one of the biggest social concerns of this century and it comes with far-reaching consequences. The number of people over the age of 65 years is projected to be 2.1 billion in 2050. Conversely, the number of caregivers needed is not keeping pace. We are developing the technology to continuously observe, learn, and trigger just in time care for aging seniors, anywhere. Along the way, we are proving these technologies in group living, home care and seniors aging at home. As a result, we are producing rich data sets that are not currently available anywhere.
Walk us through CarePredict’s products, and what problems they solve.
We are solving the problem of over-reliance on in-person observation to catch declines in the health of seniors aging in a group living or at home. As the number of elderly increases and the number of caregivers declines, a force multiplier is needed to deliver just in time care to intervene at the earliest signs of declines to save money and extend life. Group living, health insurance systems and families will benefit when an elder gets the care that allows them to live in their familiar care setting as they age without escalation to a higher care/cost setting.
Which product are you currently focusing the most on?
CarePredict AI-powered product focuses on the senior regardless of the setting they choose to grow old in — senior group living communities, at home being taken care of by home care aides or aging independently at home assisted by family caregivers.
Which product do you think is making the biggest impact, and why?
Our AI-powered solution uses a smart wearable, Tempo™ to unobtrusively and autonomously collect data of the daily activity patterns such as eating, sleeping, walking, toileting, and grooming of the individual. Machine learning and artificial intelligence create an individual-specific baseline and the system alerts at the first sign of a deviation indicative of a decline.
We are focused on preventive senior care and the goal is to avert hospitalizations and prevent issues before they happen.
How is your product improving patient experience?
We are focused on preventive senior care and the goal is to avert hospitalizations and prevent issues before they happen. In the senior living communities, we are deployed at, we have reduced hospitalizations by 20% by predicting the probability of falls, malnutrition, social isolation, depression, and other declines in wellbeing.
In group living facilities, we have reduced instances of fall by 22% within the first four months of use. By taking a proactive approach, we can ensure robust, holistic wellness for our loved ones.
Our smart wearable is packed with features like a real-time location system and geofencing to ensure the safety and security of seniors along with a concierge services button and a two-way voice communication system. It understands the gestures of the dominant arm to recognize activities such as eating, drinking, brushing and has a host of environmental sensors as well. This allows our system to discreetly understand the activity and behavior patterns and alert when things are not as they usually are.
How do you measure the performance of your products?
We do it at many levels:
For seniors, it is by reducing the instances of falls, detecting falls and catching the early signs of declines to reduce hospitalizations. Today we are able to predict the probability for falls a week ahead and probability of an UTI/GI issue 4 days ahead and probability of signs and symptoms of depression a week ahead. Because we monitor activity and behavior patterns, the rich data sets we are collecting can be used in the future to detect the first signs of mild cognitive impairment such as short-term memory loss which can be a precursor to more serious cognitive declines.
For caregivers, it is this predictive ability of our AI-powered solution that allows them to intervene in time and prevent a more serious issue.
For families, it is the ability to look at our app and instantly know How is mom doing today or see how she has been doing overtime.
For senior group living operators, it is the ability to have visibility into the productivity of their workforce as the staff wear the Tempo wearable as well to allow real time location tracking within the building. Facilities using our system see a 61% improvement in staff response rate to alerts and requests for assistance.
For home care agencies, it is the ability to provide peace of mind to the families of their clients while providing timely care.
For caregivers, it is this predictive ability of our AI-powered solution that allows them to intervene in time and prevent a more serious issue.
Tell us about the CarePredict team.
CarePredict is composed of health technology veterans, nurses, engineers and data scientists and advised by leading geriatricians and physicians, one thing that sets us apart, however, is that every member of our team is here because they were affected personally by the aging of a loved one.
What are the primary issues in healthcare that AI can help to solve?
Senior care:
AI can directly address the rapidly aging demographic that is putting enormous pressure on our already overstretched healthcare systems. Just in the US alone, 10,000 people are turning 65 every day* and this is expected to continue into the 2030s. By 2050, a fifth of the world, 1.5 Billion people will be aged. Super aging societies will become the norm as the number of elderly increase, the number of caregivers decrease and population trends to negative or zero growth in these societies. All this means that we need technology augmentation as force multipliers to take care of this massive demographic to scale senior care planet wide. The key benefit of AI in senior care lies in its predictive abilities to identify issues before they become issues.
Remote monitoring:
Many conditions require patients to rehabilitate at home. Without constant monitoring, it is common for the condition to relapse. AI can be used to provide remote patient monitoring and reduce visits to the emergency department, minimize readmissions, and avoid hospitalization. Devices such as smart wearables provide a continuous data stream and allow early interventions.
Preventive care:
Much of the healthcare system, especially in senior care is designed to Detect and Treat. But longer life and cost savings come from the paradigm shift to Predict and Prevent. AI can facilitate personalized care that predicts and alerts in real-time the minute the first signs of a decline appear, even before they are apparent to the individual. So instead of focusing on the injury treatment plans after a fall has occurred, AI can predict and allow an intervention when a higher probability of fall is detected.
By predicting and preventing, we can ensure the overall wellness and safety of our growing aging populations.
How will AI affect the work of the doctors?
Today physicians only know the state of the person at their periodic in-person or virtual encounters. They are completely oblivious to the wellbeing of the person when they are at home. A way around this is continuous machine observation of an at-risk person. But it is not feasible for humans to absorb or monitor the massive data streams generated by continuous telemetry from an individual – But AI can sift through this rapidly and only surface signals of interest for follow up by a human.
There needs to be a conscious movement away from the ‘diagnose and treat’ method as its too late. We have been doing this for years and it has resulted in an overburdened healthcare system and poor quality of life for seniors.
How can doctors prepare for AI-powered healthcare?
At its core, the act of “Diagnosis” is merely recognizing patterns based on past individual experiences. AI excels at this as it can also take into account collective experience. I see Diagnosis as an area where AI will rapidly assert itself and treatment being the more human element. AI is an augmentation technology and not a replacement technology. The health care community must understand that instrumentation of every individual is the only way to save the health care systems costs and extend lives and to be open to continuous data streams and knowing and reacting to the person outside of in person or virtual encounters. It is also now known that the Social Determinants of Health (SDOH) are just as important, if not more so, than physiological measurements such as blood pressure, etc. But SDOH is based on activity and behavior patterns and understanding the whole of the person. We can only get there by instrumentation of the person and continuous observation of activity and behavior patterns.
Do you think that AI will drastically change the way patients are cared for?
Absolutely. Because we are beginning to catch the signs earlier, we are able to intervene at an earlier stage. Take, for instance, falls. One out of four seniors over the age of 65 years falls every year. If we are able to identify a senior at an increased risk for falls, we are able to address the situation in an entirely different perspective by focusing on muscle strengthening routines, putting in place other precautionary measures, getting to the root cause of what and why they are at an increased risk. As opposed to focusing on emergency procedures, hospitalizations and long-term immobility.
Data will allow the predict and prevent model to become the dominant healthcare method. It will allow preventive medicines, advices and services to be delivered to you even before you know you needed it. This prescience will lead to healthier and longer lives for all and a reduction in costs to the healthcare systems worldwide.
Which healthcare process has the biggest potential for improvement in the coming years and why?
There needs to be a conscious movement away from the ‘diagnose and treat’ method as its too late. We have been doing this for years and it has resulted in an overburdened healthcare system and poor quality of life for seniors. The biggest potential for improvement lies in the widespread adoption of the ‘predict and intervene’ method.
What are the current technological limitations of AI which, once overcome, could bring massive improvements to the healthcare system?
Data, Data, Data, and most importantly, labeled data. For AI to work well, the models need to be trained on massive amounts of data. Collecting, classifying and learning from data takes time before actionable insights can be produced. As AI becomes more mainstream, the confidence level in the predictions will continue to get refined.
Which AI-related technology trend do you think will have the biggest impact on healthcare in the coming years, and why?
The continuous collection of individual data will become common as our everyday smartphones, wearables start collecting more and more granular data. This data will allow the predict and prevent model to become the dominant healthcare method. It will allow preventive medicines, advice and services to be delivered to you even before you know you needed it. This prescience will lead to healthier and longer lives for all and a reduction in costs to the healthcare systems worldwide. This is only possible with AI and its ability to consume massive quantities of population-wide data and to distill into actionable individual predictions.
