The Center for Data Innovation spoke with Fahad Aziz, chief technology officer and co-founder of Caremerge, a Chicago-based tech company that uses AI to predict health concerns ahead of clinical diagnosis. Aziz discussed how AI is revolutionizing senior living communities by connecting residents with staff and family members.
Eline Chivot: What is your background, and what inspired you to create Caremerge?
Fahad Aziz: I have a background in technology and have been in this field for 20 years. During my college years, in 1997, I joined a startup for an internship. The startup grew from a team of six people in a small office to becoming a Nasdaq company in just three years. It was a great experience for me, being only 19 years old. I then moved to the United States and worked for a mid-size startup for 10 years. I had learned the ropes in sales, training, technology, and various aspects of how to run a company. In 2012, I quit the job and moved to Chicago to start Caremerge. With a co-founder, we initially thought we should build the “Facebook for healthcare” but soon realized that people don’t want to share about their sickness or medications on a social platform, so we dropped the idea and started to pivot.
I got the chance to spend a few weeks in a seniors’ community in Chicago, where I saw the challenges residents and staff members faces. As I shadowed nurses, and talked to residents and family members, I realized their biggest challenge was the lack of communication. I saw family members coming in, upset that their relatives were isolated, not taken care of, while the reason was that staff members, including nurses, while spending time with these residents, wanted to make sure they would step out of the room often enough to give these people enough time to meet their relatives. Caregivers were dedicated to their cause, and would give love to those residents, treating them like family members. This experience made me realize that my purpose had to be to serve this industry. Supply chain, automotive, financial sectors, I no longer wanted to be a part of that anymore.
Six years after founding Caremerge, we’re now serving over 400 communities across the country; 80,000 residents benefit from our software, and over 40,000 family members are connected to it.
Chivot: What are the challenges you’ve encountered while implementing Caremerge? Which features did you identify could improve the experience and increase its adoption among senior users?
Aziz: Back in 2017, I had the chance to go on a tour to visit customers. We had created a good mobile app and iPad experience, but were wondering why residents weren’t using our system more, or better. I came back with the realization that half of the population we catered to wouldn’t use our system. There are technology barriers. These users need additional support, they are not tech-savvy, they don’t know how to connect to WiFi or even turn on an iPad. So we had to do better than this, and get things right. Device technologies weren’t ready at the time, but they were getting better, and my team and I were convinced that it was through voice technology that we could connect to our senior customers. With voice technology, there is no need for training, you don’t have to login or connect to WiFi, it’s just there and all you need to do is ask in your natural language. You just need to use that invocation command. We started investing in device technologies to close this gap and facilitate getting seniors on board a system they otherwise wouldn’t embrace and use.
Yet voice was half of the solution. You can use Alexa and other smart speakers, which can understand what users are asking. But how do you respond to them? How do you give them an answer that’s context-based and intelligent. For instance, a user may be wondering if there’s a yoga class happening today. A non-smart speaker would say “No,” and that would be the end of the conversation. We wanted our technology to be able to say “No, there is no yoga today, however there are other activities similar to yoga available today.” Or: “You attended this event a few days ago, and it’s similar to yoga, would you like to know more about other options?” We wanted to offer suggestions, alternatives, recommendations— not just end up with a “Yes vs. No” situation. So we invested heavily on our back-end to create responses that would make our users feel as though they are having a conversation with a real person. Various communities set up this system in residents’ apartments, and this proved to work well for seniors as a touchpoint. Our research is looking into how to further improve our response models and better answer questions.
But there is another challenge to be solved in our industry: Falls. If someone falls in a given situation, in a community, it puts a lot of stress on the entire healthcare system. The community’s stakeholders don’t want to do anything about it, it’s too risky, so all they do is to call 911. An ambulance arrives, takes the patient to the closest hospital, but the staff in charge won’t be familiar with the individual and the complexity of his or her situation. The patient may stay for a few days, receive treatment or any other sort of medical procedure, then he or she comes back, and is taken care of differently by different people. The financial impact on the healthcare system can be up to $30,000, which is huge. At Caremerge we really wanted to figure out how we could play our part to prevent falls. With AI and devices, there is now a whole range of possibilities and opportunities.
To solve this, we looked at CarePredict as a partner. CarePredict provides a wristband that seniors wear and it monitors their activity and behaviors autonomously and can predict the likelihood of a fall before it occurs and can also alert if it detects a fall and shares that information with Caremerge. There are other technologies trying to do something similar, but the advantage is that Caremerge has so much data about its users’ medical situation and day-to-day activities, that we can combine it with Census data, and formulate predictions—for instance, about a particular user having 80 percent chance of falling. That’s one of the most exciting projects we’ve been working on.
Residents, staff members, and family members use Caremerge. Let’s take the case of a family member. Say your grandfather lives in a seniors’ community. In three month, no one has called you, you haven’t heard back from the community. And everytime you call, they try to redirect your call and it doesn’t always work. Then suddenly, your grandfather has a fall and that’s the only time you’re called by his community, and it’s the middle of the night—which is upsetting. With Caremerge, you get real-time notifications a few times a day: A message if your grandfather is taking medication, one if he’s going for a walk, and another when he’s playing a game with his friends. This gives you comfort in the knowledge that he’s fine. Similarly, if your grandfather is a little depressed, through a chat system, you are able to alert the community, staff members or your grandfather’s co-residents, to suggest they give him some company or to offer him something he likes, and keep his spirits up. You and your siblings can also have a combined group chat with other senior residents and staff members. These features enable conversations and build communities of interest, helping facilities and family members learn more about trends in residents’ well-being and wellness. That is also and really how Caremerge can create care-coordination solutions to empower patients, families, staff and caregivers.
Chivot: What kind of analytics, information, and insights are you able to generate and exchange that exemplify how your solution benefits them?
Aziz: We show communities using our system the feedback we’re getting from family members, by providing a score net promoter score (NPS). As a family member, you will receive a pop up in your app every other day, asking you whether you’re satisfied, scoring this from 1 to 10. You can see which community among all those we’re working with has the highest or lowest scores and how these evolve in time—decreasing or increasing. These types of analytics help communities figure out why people may not be satisfied with a particular community, what more can be done, what is the difference between various communities that display different scores.
Our users can access dashboards showing how many falls happened over the years to residents, which communities have experienced most falls, and the reasons why that might be. Sometimes it can be so simple to solve: It could be because the features of an environment have to be adjusted, such as furniture in a resident’s apartment. But to be able to do this—you need data.
There are other technologies that would be exciting for us to use and work with for face and movement recognition, such as cameras that only switch on in the event of a fall, with recordings showing what had happened two minutes before that event. Or infrared cameras could be a game changer. Sensors under the tiles as well—as you are walking, they can help identify who you are because of your weight, or because of how long you tend to spend in the bathroom, or how long you slept. Technologies are getting better and they will become invisible. When it comes to concerns over privacy issues—at some point in life, that becomes less of an issue. For seniors it’s more about safety and security. Isolation is the real challenge.
Caremerge is more an engagement system, a clinical system. We work with device companies to get the data we need, but we do not provide the devices. We work with Amazon as Alexa supports our voice technology system, we work with CarePredict to provide bands, and we get data from these providers to combine it with our engagement and clinical data, based on which we can make predictions.
Chivot: Which technologies do you see are most promising for healthcare and likely to disrupt this sector?
Aziz: I’m very optimistic: Technologies are getting better and better, more innovators are coming into this space. Because of AI, sensors, blockchain, and other emerging technologies, our healthcare systems will be very different, 15 or 20 years from now. There will be ways to prevent chronic diseases, people will be much more informed about illnesses and how to avoid them, and generations will have taken more ownership of their health. There’s a path to get there, and we’re just playing our little role. What excites me is the use of AI in particular. We talk about it a lot but with AI, the biggest possibility is the unknown—the things that it can do which people cannot think of or imagine can be done yet. These are actually being done. In my own area of the healthcare industry, falls are a huge problem. And if AI can predict them even with just 50 percent accuracy, it could probably prevent 10 or 15 percent of falls, which will make a significant difference in terms of costs and well-being.
Looking at the caregivers we work with: Staff members, nurses, and doctors also have families and a life of their own. They sacrifice a lot for their jobs, and are incredibly dedicated. That doesn’t always come with much or enough recognition. Technology can play a role in improving their experience, thereby changing the way this industry works. For instance, you could use AI to help other stakeholders recognize and acknowledge their role in our society. If two nurses work 30 days a month, and one of them particularly goes the extra mile, AI could detect this outstanding performance and suggest giving him or her more recognition, such as a shout out at a town hall meeting or even a gift card.
Chivot: What can tech developers, tech startups, and innovators do to “win the market,” ensure the adoption of new technologies like AI, thereby spreading its benefits?
Aziz: Based on my own experience, I would have three pieces of advice. First, you need to spend time with the people you’re building technology for, before, during, and after your product is released. Get to know them. Do not assume that you know what they want. Unless you ask them a lot of questions and understand the root cause of the challenges they encounter, you will not meet their needs and find the solution that’s right for them. A beautiful software won’t just do. Second, you need to think about your go-to market strategy before and as you’re building your product. This shouldn’t be an afterthought. You need to think about how to market it: Are you going to build a website? Do you intend to sell it on a website? Are you going to build it as an app, and if so, would it be an iOS app, etc.? Once you realize you’re going to sell this product on certain platforms, then do build the technology for those platforms. You don’t want to have to roll back and redesign the whole technology after. Third, this is about how to raise funding. The only thing that investors can’t refuse—other than team members and the product or the idea itself—is traction. If you go to investors and show that users are buying your product, and that more and more people are using it, from an investor’s perspective, it would be foolish to say no.