MyHeart Counts was one of the inaugural ResearchKit apps launched in 2015 with the ambitious goal of putting together one of the largest cardiovascular health studies ever. Through MyHeart Counts, users completed surveys, shared activity data, performed assessments like the 6-minute walk test, and more.
Last month, the Stanford team released preliminary results from the cohort of patients enrolled up to October 2015. What was most striking then were the findings on user engagement. For example, the study included a seven day baseline motion data collection period. Only 9% of participants did all seven days. Similarly, 3% did the Heart Age Survey and 10% did the 6-minute walk test.
The new version of MyHeart Counts was developed with support from LifeMap Solutions, who have also worked on ResearchKit apps for asthma and COPD. Version 2.0 of MyHeart Counts adds, among other improvements, a Coaching Module that will hopefully change that by taking that first seven days of data and translating it into personalized health recommendations.
The coaching module will guide participants through a week of baseline measurements, followed by four one-week behavior-change interventions. One intervention, for example, suggests that sedentary participants take a moment to stand up or find ways to increase their daily step count, with the aim of helping them become more active. Participants are randomized into each of the interventions, so the coaching module can act as a randomized trial that will show which interventions are most effective.
The four interventions being tested will vary some of the core aspects of the user experience from the prompts and messaging delivered in the app to the graphics showing user data and the user interface itself. In some ways, this sounds a lot like the kind of basic testing that entities like Amazon and Google undertake, evaluating all kinds of variations on the user experience to maximize engagement. If all works as planned, version 2.0 of MyHeart Counts could not only inform the design of a more effective version 3.0, but could have implications for the design of a wide range of other digital health apps.