Digital health has quickly evolved to a point where we can now capture data on everything from step counts and hours of sleep to blood pressure and blood glucose – all day, every day and everywhere.

In addition to these advances in data science, digital health has more recently begun to tap behavioral science to better understand what people do relative to their health and why they do it, towards motivating them to make healthy choices.

The next frontier in digital health is happening right now at the intersection of those two sciences in a place called behavioral data science.

As a whole, AI-based digital health tools are valuable because they eliminate access barriers to healthcare and reach more people than is possible within the confines of a brick-and-mortar physician’s office. More importantly, and unlike telehealth solutions, AI can reduce the day-to-day cognitive load of managing a chronic condition.

For example, a person with Type 1 diabetes faces an average of 180 health-related decisions every day. AI-based digital tools put real-time data and recommendations based on that person’s day-to-day experiences at their fingertips – anytime they need it – to simplify decision-making and remove guesswork. 

This kind of continuous feedback, which delivers minute-to-minute guidance for achieving optimal results, is especially beneficial to people with chronic conditions. They also benefit from tools built on behavioral data science because proper management of a chronic condition requires monitoring health metrics and receiving feedback on how behaviors affect them.

Let’s take a closer look at how the integration of behavioral science and data science into digital tools helps people manage their chronic conditions better than tools based on just one of those sciences.

Integration is essential

It’s important to acknowledge that no single scientific discipline can or should delineate digital health solutions. A company might engineer the greatest data-logging and data-analyzing tool devised, but if engagement with the tool is low – meaning few people use it – then it offers little benefit. The best digital solutions are those that embrace the cross-disciplinary approach of behavioral data science.

To be sure, AI-based digital health tools must be firmly rooted in data. Data science-based tools collect and analyze all kinds of health metrics and data, then use those analytics to give the individual health information, such as blood pressure and dietary patterns over time, and recommend appropriate actions.

However, tools based solely on data science have two main drawbacks. First, the value of the recommendations depends heavily on how frequently the person interacts with the tool to provide the necessary data – for example, food consumption to gauge the sodium intake that affects blood pressure. Second, the data can indicate how much weight loss would be beneficial towards better regulating blood pressure, but can’t account for the behavioral barriers the person might have to overcome in order to lose that amount of weight.

The point is that digital tools have to do more than just feed people data. They must also motivate them to improve their health. This is where behavioral science comes in. It can support a level of engagement and retention with the tool that helps keep the individual motivated to stay healthy.

In the context of digital health tools, this science depends on a feedback loop between a person’s health data and their behaviors that continuously assesses their health status and modifies behavioral recommendations accordingly:

• Data science provides quantitative information on current health status and its near-term implications.
• Behavioral science determines how and when to “translate” that quantitative information into messages the person can understand and will most likely follow.
• Individual health metrics provide data on the effects those recommendations have on the person’s health status.
• Those metrics, plus the person’s level of engagement with the tool, provide insights into the interplay of behavior and data, enabling continuous refinement of the recommendations.

This feedback loop between behavior and data allows the digital tool to evolve with the individual as their health needs change over time. Digital tools based on behavioral data science can also reduce treatment burden by presenting people with chronic-condition recommendations that are already aligned with their behaviors – as learned by the tool – and therefore are easy to adopt.

Digital solutions built on behavioral data science are essential for people with chronic conditions to achieve their best health outcomes. These integrated solutions are also critical for digital health companies, because individual uptake and ongoing engagement will be higher for tools that provide important metric-based health information in ways that adapt to the “ups and downs” people experience while striving to improve their health.

Digital tools based on an understanding of the brain or body alone will leave many people with chronic conditions looking elsewhere for the support they need.

Dan Goldner is the executive vice president of advanced technologies, research and discovery at One Drop, which offers precision health solutions for people living with chronic conditions worldwide. He oversees the One Drop data science team transforming more than 31 billion health data points into predictive analytics solutions – most notably, CE-marked glucose forecasts and CE-marked blood pressure insights.

Before joining One Drop, Goldner served as a data science, modeling and simulation consultant, supporting strategic decisions for the Federal Aviation Administration, NASA and Fortune 50 corporations in pharmaceuticals, aerospace, technology and consumer packaged goods. He holds a PhD from MIT and a BA from Harvard University, both in physical oceanography, and an MEd in secondary mathematics from the Boston Teacher Residency.

Harpreet Nagra is a behavioral scientist, clinical researcher and licensed psychologist with over 15 years of experience in nonprofit organizations, academic institutions, medical clinics and private practice. At One Drop, Nagra guides the application of behavioral science frameworks and methodologies across key components of its AI-powered platform. Most recently, Nagra worked as a licensed psychologist and served as assistant director of Vista Counseling & Consultation, Simultaneously, Nagra held an adjunct faculty role at Purdue University Global. Prior to these positions, Nagra was an assistant professor at Oregon Health & Science University where she specialized in mental health and interdisciplinary care for pediatric and adult patients managing diabetes mellitus, including program development for the Young Adult Diabetes Clinic. Nagra holds a PhD in counseling psychology from the University of Oregon, with a master’s degree in counseling from Arizona State University.