Challenging Everything We Know About Type 2 Diabetes Management, with Data

Diabetes is a disease of the unengaged. So, how do you influence the behavior of 100 million Americans(1) who are living with or are in the first stages of diabetes?

Leveraging Consumer Behavior to Influence Diabetics

Diabetes is a disease of the unengaged. So, how do you influence the behavior of 100 million Americans(1) who are living with or are in the first stages of diabetes? The answer may very well be in the consumer behavior patterns of this group.

At Wellsmith, we believe that personalized communication, accessibility to care providers, and the right tools can activate an unengaged person to manage their health. As the Director of Data Science at Wellsmith, I took on the exciting challenge of understanding the consumer by analyzing all our health data, and leveraging my experience in consumer behavioral influence and to look at it through the eyes of the consumer, to listen to them, which isn’t the traditional healthcare industry approach. One thing I know is in order to find an answer, you just have to listen for the signals. Through this, I discovered the problem with managing type 2 diabetes two-fold.

First, the way the care cycle is set. Doctors see their type 2 diabetic patients twice a year, at best, and because they are unengaged, their short appointments serve more as a check-up, than an opportunity to learn how they are managing their condition and their day-to-day behavior, so they can provide personalized health guidance.

Second, the unhealthy influences type 2 diabetics face exist everywhere, requiring them to make hundreds of decisions a day that directly impact their health – most lead a sedentary lifestyle and unhealthy food choices are more accessible.

How is Wellsmith challenging everything we know about diabetic management?

At Wellsmith, we care about results. We believe if you are more engaged in your own health, it will lead to better health. As an engagement platform, we use both active and passive engagement methods to influence behavior. The result – 65% of people on the Wellsmith platform are engaging consistently and experiencing reductions in A1C greater than 1 point. It all comes down to 4 things: listening, understanding, creating touch points within the normal daily life of our population, and using the inherent traits of a person to influence changes in their behavior. The application of data science to this problem is pretty straightforward.


An important part of any behavioral system is listening to what is actually happening. We listen to what people are doing, not saying. The data can tell you everything you need to know, but you have to collect it early and often. With thousands of data points, and by communicating with a person, we can listen to how their choices are affecting them. We take an interest in what is happening because that feedback loop is another listening channel that is imperative to leverage in our platform.

We take the security and privacy of our participants seriously, so all our data is aggregated and de-identified. When collecting this information, we have identified representative personas for our population, and this is what we use to drive our results. An individual’s information is additionally protected by creating look-a-like segments from collected data.


If you do not understand all this data, you run the risk of undervaluing the information. Many other platforms in the market, all they care about is the data. At Wellsmith we understand the inherent value, but more than that, we strive to discover:

  • How do the daily choices of a person affect their health;
  • What common traits exist within a diabetic population;
  • Specific to our platform, what behaviors are exposed and when;
  • Working with care providers, how can we separate medical from behavioral issues;
  • Everyone is different, what are the expectations of an individual.

As data flows in, we create workflows that help us to understand these items. We develop a mix of direct application and data exploration. In my opinion, good data science doesn’t just answer questions, but with the results, we create new, better questions and continue to improve our understanding of populations. Using the ever-evolving personas and segments within those personas, we are able to understand more about the people using and succeeding on the Wellsmith platform.

Creating Touch Points

Once we gain a better understanding of the behavior of our population- who and how they are using the platform, we can continuously provide the best features and services for the specific needs of a person on the Wellsmith platform. Every interaction is important, but identifying and focusing on the right interactions is more important for influencing behaviors.

Using data, we look at the patterns of behavior and identify ideal opportunities to interact with our population. By leveraging the same strategies that consumer-based applications use, we are able to influence our population, not as a patient but more as a consumer. People are more likely to do what is best for them if they have control over their own health.

Figure 1. Identifying opportunities to inject messaging at certain touchpoints within the consumer journey.

Feedback Loops and Success

Every touch point is cataloged and factors into our measure of success: engagement. All this data is then fed back into the platform and analyzed. When we find something that works, we implement and expand. When something doesn’t work, we adapt and change and use those failures to drive new hypotheses to test and learn. It is a never ending cycle that leads to our measure for success: better engagement.

Sources: 1 CDC report: More than 100 million Americans have diabetes or prediabetes:(