Machine-based Algorithm for Increased Physical Activity and Sustained User Engagement

We partnered with Telus Health Inc. to begin to examine the effectiveness of a supervised ML algorithm integrated into their Telus Wellbeing corporate wellness app (iOS and Android versions), formerly known as "Sprout-At-Work". As of April 29, 2025, Telus Wellbeing was the employee wellness solution for more than 50 companies across 70 countries (i.e., approximately 90% of users registered within USA). At the time they reported over 400,000 registered users and about 100,000 monthly active users (MAUs; i.e., users who engage with the platform at least once a month). Download and use of the app is voluntary and of no financial cost to employees. The multi-component app is rooted in the core components of the three-layer Behaviour Change Wheel (BCW), a comprehensive framework for designing and evaluating behaviour change interventions.

Telus Wellbeing operationalizes the COM-B model, for example, by providing educational content and PA self-monitoring opportunities to build knowledge and skills (i.e., capability), offering features like peer-to-peer sharing and team challenges to foster social support (i.e., opportunity), and using FI (i.e., motivation) to drive goal achievement (see Supplementary File A for more details). Telus Wellbeing also addresses structural barriers (e.g., built environment), systemic challenges (e.g., socioeconomic disparities), and contextual factors (e.g., social norms) through the integration of promising BCTs. For example, by offering offline functionality (i.e., environment restructuring [12.1]), subsidies for wellness programs (i.e., material incentive [6.1]), and connecting users to local health services (i.e., social support [3.1]), Telus Wellbeing supports diverse user needs (Supplementary File A).

After downloading Telus Wellbeing, users can open the app and navigate through the landing and home page. On the home page, the number of steps completed that day and the users daily step goal are shown. Participants can click on two icons at the top of the home page. If the users click on the left icon, the history page is displayed. The history page allows participants to track their performance over the past week by showing their daily steps and daily goals on a color-coded bar graph. The green bar indicates the accomplishment of achieving the step goal on the corresponding day, and the red bar indicates failure to achieve the step goal on the corresponding day. The built-in health chip in the iPhone and Android devices collects the step data, and the accuracy of step counts collected by the iPhone and Android health chip has been validated in a number of studies to have comparable accuracy to an ActiGraph. The push notification for the app is also activated (if activated by the user), and the standard iOS and Android push notification is used. The push notification is visible in the landing page and in the recent notifications tab on the phone.

Eligible participants will start a five-week run-in period after downloading the app. The purpose of the run-in period is to collect baseline daily steps, and assess if the participant is able to comply with the requirements needed to regularly use the Telus Wellbeing app. During the run-in period, all participants in the control and the intervention groups will receive a traditional personalized daily step goal based on their historical step data. The machine learning algorithm will not be used to compute step goals for participants in the intervention group during the run-in period. Dynamically increasing step goals will be used in the run-in period to engage participants in using the app regularly. In addition, all participants will receive a push notification at 8:00 AM that provides the day's step goal, and if the participant accomplishes the goal before 8:00 PM, then another push notification will be sent to congratulate that participant on reaching their step goal for the day. The identical goals between the 2 groups during the run-in period is used to establish a reference level of initial physical activity, which will be used in the statistical analyses to compare the difference in daily steps between run-in and 12 weeks for the 2 groups. Data collected during the run-in period will be used by the machine learning algorithm to compute step goals for the intervention period. This is a valid approach because run-in data will be indicative of the preference of different participants. All participants will have been placed into one of two groups. The allocation of app users to groups will be implemented by Telus Health after the run-in period. After the five-week run-in period, participants in the control group will be provided with a personalized static daily step goal hrough the Telus Wellbeing app. Participants will receive a push notification at 8:00 AM every day that provides that day's step goal (10,000 steps), and if the participant achieves the goal before 8:00 PM, then another push notification will be sent to congratulate the participant on reaching their step goal (of 10,000 steps) for the day.

After the five-week run-in period, participants in the intervention group will receive adaptively personalized step goals through the Telus Wellbeing app. The daily step goals will be computed using machine learning on the complete history (past steps and goals) of the user. Machine learning will be applied every day to reduce variance in future steps and goals. Participants will receive a push notification at 8:00 AM every day that provides today's step goal, and if the participant accomplishes the goal before 8:00 PM, then another push notification will be sent to congratulate the participant on reaching their step goal for that day. Machine learning will adaptively compute personalized step goals that are predicted to maximize future physical activity for each participant based on all their past steps' data and goals of each participant. Machine learning is applied to each participant individually, and it consists of two main steps. The first step is to use all of the participant's data to construct a quantitative model that predicts how many steps the participant will take in the future, given a prescribed set of step goals, and an important aspect of the model is a component that describes how achieving goals in the present can increase the likelihood of achieving goals in the future. The second step is to use this quantitative model to select a sequence of step goals that maximizes the predicted future number of steps. To make the process of updating step goals adaptive, machine learning is applied each day (using all the users' past data) to generate step goals for the coming day. Moreover, the step goals computed by machine learning for the coming day are not constant but increase or decrease based on the model prediction. The Telus Wellbeing app will automatically track the participants' step counts each day and will provide goals regardless of their level of engagement within the app over the 12-week study period. As the data is being analyzed retrospectively, no "end of study" letter will be provided to participants. However, upon registering for the studies procedures, all users provided written informed consent for their data to be collected and analyzed.