From The Editor | April 25, 2018

AI Provides Help For The Sandwich Generation


By Bob Marshall, Chief Editor, Med Device Online

AI Provides Help For The Sandwich Generation

Many of you reading this article may be a part of the sandwich generation.

The Pew Research Center provides this this profile of sandwich generation: “Adults who are part of the sandwich generation — that is, those who have a living parent age 65 or older and are either raising a child under age 18 or supporting a grown child — are pulled in many directions. Not only do many provide care and financial support to their parents and their children, but nearly four-in-ten (38 percent) say both their grown children and their parents rely on them for emotional support.”

As those of us in the sandwich generation provide emotional support for our aging parents, we also often need to provide healthcare assistance as they age and begin to decline. We worry about them as they try to maintain their independence and age in place. Fifteen years ago, I was filling pill boxes and calling to remind my father to take his pills as he battled congestive heart failure in his final years. After he passed away, my wife and I became caregivers for my mom, who had a stroke and was entering what would be a decade-long journey into dementia. She was a much more compliant patient than my father, but her memory loss intensified her need for care. One thing that I found interesting was that, as my mom’s ability to relate to and recognize people dissipated, she continued to watch television and also enjoyed holding stuffed animals.

All of these memories were in the front of my mind as I talked with Dr. Cory Kidd, founder and CEO of Catalia Health, about Mabu, the company’s personal healthcare companion. How might a product like this have helped my parents and made life easier for all of us? But first — just who or what is Mabu? See for yourself.

The Mabu platform learns about each patient’s personality, interests, and treatment challenges over time. This enables Mabu to create conversations that are tailored to each patient, and that resonate with their unique personality and circumstances. The structure of these conversations is based on proven behavioral models of psychology to promote behavior change. Unlike apps, which usually are used for only a short period of time, patients continue to use the Mabu personal healthcare companion. This allows Mabu to have conversations with patients every day to manage their treatment and gather data about their progress.

“The whole rationale for the robot is not about the technology, it’s about psychology,” Kidd explained. Mabu interacts with an individual and builds up knowledge by learning about the patient both medically and psychologically. This occurs in much the same ways that people get to know one another: implicitly in conversation, or explicitly, as in a patient/care-giver dialog.

The interaction typically begins with a greeting or perhaps a bit of small talk. The middle of the conversation is focused on how Mabu is helping the patient. Kidd offered the scenario of Mabu talking with a heart failure patient about their low sodium diet, if that’s not going so great. From there, the technology selects one of the “conversation trees” to control the structure of a small piece of conversation.

“If we contrast what we are doing to a lot of chat-bot technology, in that case the user is really driving the conversation; they say something or ask a question, and then the technology is essentially doing statistical analysis to figure out an answer. In our case, Mabu is always driving the conversation,” Kidd said.

The learning aspect of Mabu and the connection it forms with a patient suddenly made me think of a potential down-side to the relationship. What if something happens to a user’s specific Mabu unit? Especially in the case of older patients, who often have already lost a number of their family members and friends. Kidd assured me they had already thought about this when creating the technology, “The robot is simply the interface to the patient. All of the data resides on our HIPAA-compliant backend. So, if the robot falls off the counter or the cat decides it’s not happy with a robot usurping its place, we can ship out a new one and it will pick up right where the old one left off.”

I asked Kidd about the replacement of a broken unit because I watched a Mabu video and appreciated the bond that formed between the user and Mabu. Kidd explained that he realized the importance of this bond years ago.

“We designed the system that way based on what we have seen patients do. The first time we put these in front of patients was about 11 years ago in the greater Boston area. I had built all of those robots in the basement shops of MIT and they were a bit bulkier than they are today. But when I went back a few months later, it wasn’t a plain robot,” Kidd said. “People had dressed them up with hats, scarves, and one lady even put a red feather boa around her robot’s neck. Every single person who had one had named it. So we do see that bond forming very quickly and it’s all about the psychology and the medicine. The technology is just in support of those two priorities.”

Catalia Health also can connect a missing link for pharmaceutical companies that are a couple of steps removed from their users. In the early stages of drug development, pharmaceutical companies are heavily involved with patients through R&D and clinical trials. But, once the drug is out in the world, unless there are major problems, there is a general loss of visibility.

“We can help pharmaceutical companies gain insight into the lived experience of people using their products. We work with a lot with large healthcare systems and with big Pharma, those are our customers and we’re providing a care management program that can be scaled to large numbers of patients. We’re not selling a device. We’re not selling a piece of hardware or software. What we’re really selling is a service providing feedback on adherence and outcomes that happens to be delivered through a very innovative device,” Kidd concluded.