California-based Twin Health creates AI-powered "digital twins" to help people address the root cause of chronic metabolic conditions such as pre-diabetes, type 2 diabetes, and obesity. The company’s chief medical officer, Lisa Shah, M.D., recently spoke with Healthcare Innovation about the results it is seeing across its customer base, including employers such as Walmart. 

The company describes how it works this way: Using sensors like a smart scale, activity tracker, continuous glucose monitor and blood pressure cuff — along with quarterly bloodwork — Twin Health creates a “digital twin” of your unique metabolism. Then your digital twin makes highly personalized recommendations (via a smartphone app) regarding nutrition, sleep, physical activity and more to help you reach goals like a lower A1C, weight loss, and reducing reliance on medications like GLP-1s, which can be costly and unsustainable for many people.

Healthcare Innovation: Dr. Shah, could you explain the basic concept of Twin Health for us? 

Shah: We organize all of the internal biological data from wearable sensors and external behavioral data into a time series, and start looking at cause and effect and understanding what you are doing right now. For instance, I’m talking to you and I'm standing at my standing desk. That is a very different impact of my metabolism than if I'm sitting at my desk. My watch, which is one of the sensors we use, knows I'm standing right now. We map that, and we look at my blood sugar. You can see I'm wearing a sugar monitor. We use a Bluetooth scale. We use a Bluetooth blood pressure cuff. We track blood sugar, we track heart rate, sleep, stress levels, and then we organize all that and tell people with chronic metabolic diseases the little things they can do in their day — whether it be walking after a meal or meditating before bed or switching rice for a vegetable — all of these little things that are going to optimize their metabolic health biologically.

HCI: Is there some learning involved in how those nudges to people work or don't? Have you learned how to deliver those in a way that they'll act on them and not see them as an annoyance?

Shah: So much about what we do is how to maximize behavioral change through the right type of nudges and the right leverage. Members use an app and they get a human care team surrounding them. We have licensed providers, coaches, nurses, and medical assistants. As the person wears the sensors, it helps to feed and build their twin. The care team really builds that trust and leverage. We say, OK, what are your goals? A member will say, ‘Well, I want to lose 20 pounds. I want to get off my insulin. I want to live for my grandchildren.’ We collect all their goals and then use the digital twin in the data to immediately give them nudges and remind them that if they do this, they're going to see this outcome. 

If you told me you were sitting down each Chipotle right now, it would show you a curve and say this is what's going to happen to your blood sugar, but if you eat half of it, or if you walk to Chipotle instead of Door Dashing it, this is your new curve. Now if you add avocado on top, this is your new curve, and they can make choices. An hour later, the digital twin feeds back the curve. Look what you did. Members come to the app and they get these little nudges of small things they can do in choices, and then they get immediate feedback. Then over the week, we show them their average sugar went down this much and tell them that in six weeks they could get off their insulin, for instance. We remind them of how far they are from that goal. We find 92% of users are using the app seven days a week at one year, and 86% are using it five days a week at one year because they become really dependent on it for the little things.

HCI: Was there an initial health system, payer or employer partner that worked to validate all of this?

Shah: We actually did our research building years from 2016 to 2018 in India. Then we brought it to the U.S., and our first big partnership was with Blackstone Equity Healthcare. They are not an investor in us, but they chose to bring this to their portfolio companies. So now we have almost 35 of their portfolio companies having all of their employees on Twin Health, and they've been able to save massive healthcare costs for the companies that they're invested in. We now have over 200 partners. Walmart is one of our largest employer partners, and we have thousands of Walmart employees on Twin.

HCI: I understand that Cleveland Clinic conducted a controlled trial of the program and published the results. Can you summarize what they found?

Shah: The Cleveland Clinic trial was very important for us to demonstrate that we could replicate these outcomes in a controlled trial setting. In 2021 we built the protocol and in 2022 started enrolling patients. In 2023 we finished the study and in 2025 we published it. They randomized 50 people to get usual diabetes care at the Cleveland Clinic and 100 on the digital twin. The goal was at 12 months, could we achieve an A1C below 6.5 and eliminate all medications except Metformin?

The primary outcome was that 71% of the Twin group met the goal at 12 months. They eliminated all meds except Metformin, and got their A1C below 6.5. Only 2% in usual care were able to do that. In usual care, they actually added medication. Our value proposition, is getting people off of GLP-1s, expensive drugs, insulin drugs that have side effects, drugs that create problems. People want a chance at being medication-free. We can only do that in diabetes and weight loss when we can actually improve the baseline metabolism. So we were able to do that. 

We're getting at the root cause. We're addressing insulin resistance. We're figuring out how to reduce visceral fat around the organs, but also  improve muscle mass. When we bring Twin in, we not only eliminate the drug and have greater weight loss, but we actually build 8% more muscle back, and muscle cells metabolize faster. So when you have more muscle, your insulin resistance goes down, your inflammation in the body goes down. So that's really what Twin is focused on.

HCI: So when Cleveland Clinic sees those results, do they want to extend the use of it?

Shah: They do. They've been talking to us a lot about bringing it to their own employee population. They want to do more studies with us. But that trial was pretty landmark, because no one has done a randomized, controlled trial on diabetes reversal with digital twin technology of human metabolism. It really opened up our aperture. We are growing rapidly with health plans and more employers. We were really glad we did the study, but we really want to bring this to the commercial population. For us, this is about getting it into the real world.

HCI: Could you give an example of a payer that you partner with, and if they've had time yet to see to see some results from it?

Shah: We have a partnership with Health Care Services Corporation, which is five of the largest Blues plans in the country. We work with them to support employees as well as their fully insured population. We’re working with them on both obesity as well as diabetes and thousands of members are seeing similar results there.

HCI: What are some other potential directions for the use of this technology?

Shah: We are so focused on insulin resistance. We're so focused on inflammation and visceral fat that when we do what we do, everything gets better. We actually measure 10 systems on every member, and what we have seen is we're improving hypertension, we're improving cardiovascular disease and Framingham risk score. We're improving metabolic associated fatty liver disease. We're improving obstructive sleep apnea. We're improving chronic kidney disease. 

We would really like to get deeper into the cardiovascular space. That's where we really think we have the biggest power. We're able to really reduce cardiac events. I'd love to be able to support cardiac rehab areas where people have cardiac injury or cardiac insult, and you're trying to prevent another event. How can you really use sensor technology, wearables and the guidance of your digital twin to improve that, slowly and surely, and get to a place where, you've undone some of that damage? 

Liver disease is another great opportunity for us. We are really excited about continuing to work on demonstrating our impact on metabolic associated fatty liver. There are a lot of drugs that are about to come out in that market but that are really expensive. So that's a big opportunity for us. And of course, we see in obesity, a lot of people on GLP-1 medications. We think we offer a great opportunity for people who don't want to be on the drug forever. They want a strategy. And today's care is really not offering a standard of how long people should be on them. We're good at starting them, but we're not monitoring them and actually tracking when people should come off them or whether they are suffering from side effects. So there is an opportunity there to empower clinicians, empower the care team with real-time data, understand how people respond, and give them an opportunity to titrate off without rebound weight gain and actually continue to lose the weight and live a healthier way. 

HCI: Is there a way that payment models could be aligned to take better advantage of this? 

Shah: First of all, we're at risk. All of our partners pay us on outcomes. We only get paid for reducing A1C, helping somebody get off a medication safely, and weight loss — things that are tangible. I think we all need to shift into that sort of mindset. I think the applicability of Twin and digital twin technology for metabolism is really important, and one we're exploring for the Medicare population right now.