An AI-Built ‘Digital Twin’ Could Improve Your Health, This Startup Says – CNET [CNET]

Do you know what your metabolism looks like? I don’t know mine, either. If we did, it could help with our overall health.

Our metabolisms are the internal processes that turn food and drink into energy. Research shows that no two metabolisms are quite alike. And that complicates the treatment of metabolic diseases like Type 2 diabetes and obesity, which impact roughly 140 million Americans.

But a 6-year-old startup called Twin Health is using AI to create replicas of its members’ unique metabolisms and to generate personalized health plans to help them lose weight and get off medications like GLP-1s, which include names you might recognize, such as Ozempic, Wegovy and Victoza.

Or as Chief Medical Officer Lisa Shah put it, Twin Health is looking at “what inputs drive what outputs … to precisely impact each individual person with a set of recommendations that is very specific to them.”

One of the greatest promises of AI is its potential to solve problems we haven’t been able to successfully tackle on our own. Research shows that nearly half of Americans are trying to lose weight through exercise and/or diet changes, but as many folks know, it’s a struggle. And so AI now offers hope to help curb the obesity epidemic that began in the US nearly 50 years ago.

Twin Health’s metabolism replicas are called Whole Body Digital Twins.

Each digital twin is created using data from sensors and clinical labs, as well as your preferences, which it learns based on the choices you make regarding food and activity.

When you become a Twin Health member, you start wearing a glucose monitor so your blood glucose readings can be taken every five minutes, as well as an activity tracker to monitor activity levels, heart rate, heart recovery rate and sleep. You also use a Bluetooth scale to track weight and muscle mass and a Bluetooth blood pressure cuff to track stress and hypertension.

You also have to log what you eat daily and get lab work done every 90 days.

Twin Health uses all this data to learn about each of its members and to offer increasingly precise advice thanks to the machine learning algorithm powering the platform.

In addition to a personalized program, the goal is to encourage behavior changes that members might actually stick to. The digital twin generates recommendations regarding nutrition, sleep, activity and stress, but it also offers choices.

For example, if you want to eat two slices of toast for breakfast, the app will show you the subsequent spike in your blood glucose level, as well as how much your weight will go up. It then offers choices to mitigate these spikes, such as cutting down the serving, eating an egg beforehand or walking for 15 minutes after you eat the toast.

You can choose to follow one, all or none of these recommendations. Over time, the AI model learns what you’re willing to do and uses that information to provide choices you’re more likely to accept, Shah said. Then ongoing lab work helps Twin Health understand how these recommendations impact your organ systems.

“What we know about health care is that health care is behavioral. A lot of it. And it’s environmental,” Shah said. “And so if you ask somebody to do something they’re never going to do, maybe they’ll do it for two or three months, but they’ll get tired of it.”

Twin Health has nearly 9,000 members, Shah said. The company works with employers and health plans to cover the cost. (She didn’t disclose that cost.)

According to a spokesperson, consumers can’t sign up individually yet. 

Twin Health closed a $50 million Series D round in December 2023. Investors include Temasek, Iconiq Growth, Sofina, Peak XV and Helena.

Early results show promise.

A yearlong study published in the American Association of Clinical Endocrinology found that the program helped improve fatty liver disease in patients with Type 2 diabetes.

“Our vision is, right now, metabolic disease, obesity, Type 2 diabetes, prediabetes,” Shah said. “Our hope is someday everyone has a digital twin and can learn about themselves and can prevent disease and improve longevity and really understand specifically beyond generic advice — just eat better, stress less, sleep more, work out — what they can do and control and have the power of choice in their health care.”

Other Services & Software