Exploring the Predictive Potential Behind One Drop with Director of Data Science, Ydo Wexler

Exploring the Predictive Potential Behind One Drop with Director of Data Science, Ydo Wexler

Have you ever wanted your app to know just the right time to send you a notification, whether it’s a reminder to check your blood sugar or log your meals for the day? But how can an app get to know you so well that it can predict what you need and when you need it? Ydo Wexler, director of data science at One Drop, explains the predictive models that can make this technology possible—and how they can transform the lives of people living with chronic conditions when applied to precision health solutions.

The Adaptive Support Models that Power One Drop

When using the One Drop app, you’re routinely met with notifications on everything from blood sugar trends to check-ins with your coach. Currently, most of these notifications need to be set up manually in the app, but Wexler and his team are working on a way for predictions to drive these prompts—meaning you’ll get your self-care reminders at the right times without ever having to think about it yourself.

“We’ve developed a system called the ‘adaptive support model’ (ASM), which learns when and how to nudge One Drop members for different behaviors (logging blood sugar, food, or blood pressure; taking medication; going for a walk) and does so in a way that’s customized to fit each and every individual,” explains Wexler.

Using machine learning—which allows computers and apps like One Drop to interpret data more deeply with human-like insights (think: recognizing patterns in how often you log your food or blood sugar, the numerical values that you’re logging, or the average number of steps you take per day)—the ASM can comb through these health data points and find similarities in not just your own health behaviors, but also the behaviors of everyone else using the app whose experiences resemble yours, whether it’s consistent blood sugar spikes after a particular form of exercise or forgetfulness with medication logging. With all of that data at its disposal, the ASM can use that information to provide precise self-care recommendations catered to each individual One Drop member.

“We built this system to customize our communication with our members in a way that will encourage them most and will be most effective for them,” explains Wexler. “We want to learn how our members will respond to different prompts or reinforcements, and how setting different goals adaptively can encourage them to achieve these goals and improve their health.”

In other words, the more someone engages with the One Drop app, the more the ASM (once implemented) will be able to learn about them (or adapt to them) and what works best for their unique self-care plan. Plus, since One Drop can sync with FitBit, Dexcom, the Apple Health app, and several other popular health apps and devices, that means we have tens of billions of data points to help refine these predictions and give members the right suggestions at the right time.

Of course, Wexler can’t accomplish this feat alone. “When I’m conceptualizing these projects, I’m always thinking about who we need to collaborate with—the product team, the behavior science team, whoever it may be—in order to create something that everyone really wants and that we have the resources to make,” shares Wexler. The behavior science team, for example, can devise the right language and timing of different prompts or reinforcements, while the product team can execute those ideas and bring them to life.

Along the way, continues Wexler, “we’re always thinking about how these projects relate to the vision we all have for One Drop as a whole: to improve precision health for people living with chronic conditions.”

How One Drop’s Approach Creates New Possibilities for People Living with Chronic Conditions

For the hundreds of millions of people in the U.S. living with a chronic condition (or, in many cases, multiple conditions), the healthcare system has largely relied on a complacent mindset of reactivity over proactivity.

One Drop’s approach, on the other hand, is about using precision health to get ahead of chronic health issues before they can get ahead of you—a goal that drives Wexler in every project he works on, whether he’s perfecting the accuracy of the app’s blood pressure insights or brainstorming ways to simplify food logging.

“Our company’s tagline is ‘reimagine possible,’ and I think that’s what data science is really all about: taking what seems impossible and figuring out a way to make it a reality.”

“With this technology, I hope that we’re creating something that helps people living with chronic conditions see where they’re going with a clearer vision,” adds Wexler. “But the real hope is that our ideas and our products—even if they’re not a cure for any of these chronic conditions—will allow every person to get the best care they can.”

Link copied to the clipboard. ×
Allie Strickler
Nov 17, 2021

Additional Reading

Why Handing Out More Genuine Compliments is Good for Your Health

Why Handing Out More Genuine Compliments is Good for Your Health

Read time: 5 minutes Studies in compliment research demonstrate the benefits for both the receiver and giver. Most people underestimate the value of a compliment. ...

Read more >
Your Guide to Strength Training

Your Guide to Strength Training

Read time: 4 minutes The idea of strength training can seem intimidating if you’re not one for the gym, but if you haven’t found your...

Read more >
Visit our store

Gorgeous gear. Supplies shipped to your door. On-demand support from diabetes experts.

Shop Now