Posted: Sep 09, 2019 7 min. read

STEP UP: When paired with gamification and a science-based engagement program, wearables can boost (and sustain) activity levels

By Greg Szwartz, Life Sciences data science practice lead, Deloitte Consulting LLP

At the Temple Mount in Jerusalem, the ancient construction is strikingly precise. Every stone in the outer wall appears to have been meticulously placed, and all of the stairs are exactly the same size and shape. The last flight of stairs leading to the entrance, however, is noticeably different. The width and height of the of the steps are erratic—some are narrow while others are broad. To keep from stumbling, visitors must watch their feet closely as they walk. Historians suspect the arrangement is intentional.1 As visitors ascend, they are nudged to bow in respect.

Thousands of years ago, the people who designed this temple understood there are ways to nudge people to do something they might not do on their own. But you don’t have to go all the way to Jerusalem to see behavioral science in action. Some of the same behavioral-science techniques used by architects 12,000 years ago can be incorporated into wearable devices and companion apps to nudge people to be more active.

About 600 Deloitte employees (from 40 states) recently participated in a 36-week randomized clinical trial to determine if a wearable activity-tracker—combined with gamification—would increase physical activity among overweight and obese adults. The STEP UP study was led by the Perelman School of Medicine at the University of Pennsylvania. Deloitte funded the study as a part of its ongoing investment in developing data and science-driven approaches to engage patients. By tapping into pre-existing behavior patterns, we can help ensure that the healthy choice is also the easy choice for people who are trying to be healthier and more adherent to a treatment plan.

Here’s how it worked: Participants wore a fitness-tracker for two weeks to establish an activity baseline (the average number of steps taken per day). They were then asked to set a goal that was 33 percent, 40 percent, or 50 percent higher than their benchmark.

The 150 people who made up the control group were asked to use the fitness tracker and strive to meet their daily step goal for 36 weeks. The other 451 participants were divided into three intervention arms—support, collaboration, and competition. Prior to the start of the study, these participants were asked to sign a pre-commitment pledge that they would strive to meet their step goals. Commitment devices have been shown to motivate behavior change.2 Participants also earned 70 points (ten points for each day of the week) based on whether they met their weekly goals. Participants lost points if they didn’t meet their goals. Here’s how it worked:

  • Support arm: Participants were asked to select a family member or friend who they could email each week and report their progress—including the number of points earned and level.
  • Collaboration arm: Participants were placed into three-person teams that communicated via email. Each day, a team member was randomly selected to represent the team. If that person met his or her goal the previous day, the team kept its points. The team lost 10 points if the selected member did not meet his or her goal.
  • Competition arm: Three-person teams received an email each week that ranked all teams by the number of points they had accumulated.

While participants in all three groups had higher activity levels than the control group, those in the competition arm saw the largest and most sustained improvements in activity. Participants in this group had 1,166 more steps per day than people in the control group, and 609 more steps per day three months after the study concluded.

The findings from the STEP UP study are exciting, and they demonstrate—scientifically—that engagement programs, if designed correctly, can move people toward healthy behaviors. Some of the core principals needed for behavior change (and effective wearable/app design) were on display in the trial. They include:

  • Choice architecture: This is the science of sequencing choices in a way that encourages better decisions. A lunch buffet, for example, might nudge you put more salad on your empty plate by placing the salad bar at the beginning of the line. Similarly, we would like to help people increase their activity levels by making the healthy choice (going to the gym, walking, or getting outside) a little easier. This requires an understanding of human behavioral tendencies and is an area where data science can be used to identify patterns in a person’s historical activity data.
  • Loss aversion: In the field of cognitive psychology, loss aversion is the idea that people would rather avoid losing something than acquire something of identical value. Most people would rather avoid losing $5 to finding $5. Some studies suggest that losses can be twice as psychologically powerful as gains.3 Each Monday, STEP UP participants received 70 points (ten points for each day of the week). Those who did not meet their step goal had 10 points deducted from their balance.
  • Self-efficacy: Significant science (behavioral and data) went into determining how to determine the most effective goals and how to ramp up those goals to promote self-efficacy. Pre-commitment contracts around goals were used by participants to connect their “current self” with their “future self.” Science-based studies such as STEP UP can help researchers better understand how to encourage change and build confidence. Establishing the goal and commitment contracts early can be critical.
  • The fresh-start effect: People are more likely to adopt new habits or change behaviors at natural transition points. Consider New Year’s Eve. As this fresh start approaches, many of us make resolutions (e.g., exercising more, eating healthier, putting more money aside for savings). Other fresh starts could be the start of a school year, or a new job, even the start of a new project. The feeling of having a clean slate makes people more open to goal-setting. For the study, five goal-setting levels were established. At the end of each week, participants who had at least 40 points were allowed to move up to a more challenging tier. They were also allowed to drop to a lower level. We established a fresh start after each time period, but those periods were varied. If people know when to expect a fresh start, they might become less engaged as the new period approaches.

Why should health and life sciences companies consider behavioral economics?

The results of this study illustrate that data science and behavioral economics—if applied correctly—can have a positive impact. The findings go beyond measuring steps to helping us understand how behaviors can be affected, improved, and even sustained (as we saw in the competitive arm). If we can combine the right types of devices and incentives, we could have a positive impact on health. The same strategies that led to increased activity levels might also be used to encourage people to eat healthier food or to improve medication adherence.

“Human behavior is the final common pathway for the application of nearly every advance in medicine,” explains Mitesh Patel, M.D., director of Penn Medicine’s Nudge Unit and the study’s primary investigator. Dr. Patel is challenging us to understand that outcomes are increasingly about the “last mile” of adherence as much as they are about the molecule or therapeutic intervention itself. In addition, benefits from behavior change can have an effect across multiple disease areas for improved health. Results from the STEP UP study demonstrate the value of gamification in a wearable companion app—particularly when coupled with a broader science-based choice architecture. Similarly, life sciences companies should consider ways to bring data and behavioral sciences into their companion applications and patient services programs. This can align incentives of the entire health care ecosystem around effectiveness rather than just efficacy.

STEP UP is part of Deloitte’s continued investment in understanding how to drive predictable behavior change. This study reinforces the idea that if you set things up correctly, behavioral nudging can be effective. You can’t just slap an activity band on someone and expect that activity levels will automatically increase. That device can be paired with informed incentives and nudges that motivate participants to change behavior without their having to make a conscious effort. While behavioral “nudging” is thousands of years old, it can still be paired with modern technology and data science to help us make the best choices.

1. Ophel and Hulda Gates of Jerusalem,
2. Rogers T, Milkman KL, Volpp KG. Commitment Devices: Using Initiatives to Change Behavior. JAMA.2014;311(20):2065–2066. doi:10.1001/jama.2014.3485
3. Kahneman, D. & Tversky, A. (1992). “Advances in prospect theory: Cumulative representation of uncertainty”. Journal of Risk and Uncertainty. 5 (4): 297–323. CiteSeerX doi:10.1007/BF00122574.


Return to the Health Forward home page to discover more insights from our leaders.

Subscribe to the Health Forward blog via email