Digital Health Beyond the Nudge

Adaptive personalization across different domains is important for effective personal health management.

First-generation digital therapeutic solutions did groundbreaking work around behavior change, using an important concept known as the nudge. A nudge is a prompt or a reminder, designed to elicit beneficial behavior. A nudge is not a command: the person being nudged can always decide to take the recommended action or not.

Nudges can be direct (“It’s time to check your blood glucose levels.”) or indirect (“Did you record what you ate for lunch today?”). They can even include facts about the benefits of healthy behaviors as a way of inspiring action.

In today’s world, nudges can be made by text message, phone call, in-app notification, or email. They can be scheduled in advance or automatically triggered by specific events.

The possibilities are far-reaching, and we’re finding that different types of nudges work better than others depending on who’s being nudged. Some people prefer daily nudging. Others just want to be reminded when they miss a reading or forget to log their exercise session for the day. Some people like to be praised when they’re being nudged, while others find that annoying or irritating.

Developing Dynamically Adaptive Care Journeys
No two people with the same chronic condition respond in the same ways, and individuals can even have different preferences within themselves from day to day, week to week, and month to month.

While the application of personalized nudges was a breakthrough in digital health, behavioral science offers many additional approaches to personalization that can lead to better outcomes for more people. Using data analytics, artificial intelligence, and psychological research, digital health providers like DarioHealth continue to optimize personalization for every individual.

Solutions that will succeed are those that adapt and pivot the fastest. Let’s say a diabetes patient is tracking her exercise and blood glucose levels on a regular basis, but then hurts her back.

A smart solution will notice the drop-off in tracking, change the tone and content of communications, and attempt to find the cause. Based on this particular patient’s profile, the initial prompt could be a text message, followed up by a personal call from the patient’s dedicated coach.

When the back pain issue is discovered, a digital health solution should be able to offer ways to address it, such as virtual physical therapy or self-guided exercises. This could happen during the call from the coach, but people who don’t respond to coaching could learn about their options in another way, such as within their app or via email.

As the patient in our example starts to manage her back pain, her digital health solution should be able to personalize the experience so that she stays engaged and eventually resumes the beneficial behaviors of regular blood glucose monitoring, diet, and exercise.

The Six Domains of Personalization
When people are going through stressful situations, communications must work to ease the tension, not make it worse. If we really know our customers, we should know who needs a pep talk and who doesn’t. We should know who likes to be contacted on the weekdays, on the weekends, at night, or at lunchtime. We should know who likes to be contacted when they make progress and who likes to be contacted when they hit a barrier. And, most importantly, we should understand which communications and recommendations will be the most effective for every person.

Users should be able to set and change their own preferences, but solutions need to be smart enough to anticipate those preferences before they even occur to the person.

Through a technique known as adaptive personalization, data can be used to build increasingly relevant customer relationships over time. Here at Dario, we look at six domains of personalization in our attempt to provide an optimal user experience under any given circumstance. These domains are:

  1. Timing – Which times of day generate the best responses from each member?
    Weekend mornings might be best for some, while others prefer contact in on weekday
  2. Tone – What tone and attitude resonate most with each member? Examples include
    friendly, caring, educational, and authoritative and their various combinations
  3. Content – What information makes sense given a member’s needs and interests at a
    particular point in time?
  4. Channel – Which media are most effective across each member’s care journey? Does
    a user prefer text messages, in-app notifications, or emails? How often do they like to
    be contacted by phone?
  5. Frequency – What level of day-to-day support drives optimal engagement?
  6. Intervention – Which coaching and clinical interventions support sustained behavior change?

Naturally, this is an ongoing process that requires continuous improvement based on information users provide and what can be learned from data based on user engagement over time. Advances in AI and data analytics are making it easier to meet the evolving motivations and needs of people with chronic conditions, so they can improve their quality of life on a daily basis.