Background: Although the health benefits of physical activity are well established, it remains challenging for people to adopt a more active lifestyle. Mobile health (mHealth) interventions can be effective tools to promote physical activity and reduce sedentary behavior. Promising results have been obtained by using gamification techniques as behavior change strategies, especially when they were tailored toward an individual’s preferences and goals; yet, it remains unclear how goals could be personalized to effectively promote health behaviors. Objective: In this study, we aim to evaluate the impact of personalized goal setting in the context of gamified mHealth interventions. We hypothesize that interventions suggesting health goals that are tailored based on end users’ (self-reported) current and desired capabilities will be more engaging than interventions with generic goals. Methods: The study was designed as a 2-arm randomized intervention trial. Participants were recruited among staff members of 7 governmental organizations. They participated in an 8-week digital health promotion campaign that was especially designed to promote walks, bike rides, and sports sessions. Using an mHealth app, participants could track their performance on two social leaderboards: a leaderboard displaying the individual scores of participants and a leaderboard displaying the average scores per organizational department. The mHealth app also provided a news feed that showed when other participants had scored points. Points could be collected by performing any of the 6 assigned tasks (eg, walk for at least 2000 m). The level of complexity of 3 of these 6 tasks was updated every 2 weeks by changing either the suggested task intensity or the suggested frequency of the task. The 2 intervention arms—with participants randomly assigned—consisted of a personalized treatment that tailored the complexity parameters based on participants’ self-reported capabilities and goals and a control treatment where the complexity parameters were set generically based on national guidelines. Measures were collected from the mHealth app as well as from intake and posttest surveys and analyzed using hierarchical linear models. Results: The results indicated that engagement with the program inevitably dropped over time. However, engagement was higher for participants who had set themselves a goal in the intake survey. The impact of personalization was especially observed for frequency parameters because the personalization of sports session frequency did foster higher engagement levels, especially when participants set a goal to improve their capabilities. In addition, the personalization of suggested ride duration had a positive effect on self-perceived biking performance. Conclusions: Personalization seems particularly promising for promoting the frequency of physical activity (eg, promoting the number of suggested sports sessions per week), as opposed to the intensity of the physical activity (eg, distance or duration). Replications and variations of our study setup are critical for consolidating and explaining (or refuting) these effects. Trial Registration: ClinicalTrials.gov NCT05264155; https://clinicaltrials.gov/ct2/show/NCT05264155

The Experience Sampling Method (ESM) is gaining ground for collecting self-reported data from human participants during daily routines. An important methodological challenge is to sustain sufficient response rates, especially when studies last longer than a few days. An obvious strategy is to deliver the experiential questions on a device that study participants can access easily at different times and contexts (e.g., a smartwatch). However, responses may still be hampered if the prompts are delivered at an inconvenient moment. Advances in context sensing create new opportunities for improving the timing of ESM prompts. Specifically, we explore how physiological sensing on commodity-level smartwatches can be utilized in triggering ESM prompts. We have created Experiencer, a novel ESM smartwatch platform that allows studying different prompting strategies. We ran a controlled experiment ($N=71$) on Experiencer to study the strengths and weaknesses of two sampling regimes. One group ($N=34$) received incoming notifications while resting (e.g., sedentary), and another group ($N=37$) received similar notifications while being active (e.g., running). We hypothesized that response rates would be higher when experiential questions are delivered during lower levels of physical activity. Contrary to our hypothesis, the response rates were found significantly higher in the active group, which demonstrates the relevance of studying dynamic forms of experience sampling that leverage better context-sensitive sampling regimes. Future research will seek to identify more refined strategies for context-sensitive ESM using smartwatches and further develop mechanisms that will enable researchers to easily adapt their prompting strategy to different contextual factors.

Background: Financial rewards can be employed in mHealth apps to effectively promote health behaviors. However, the optimal reinforcement schedule—with a high impact, but relatively low costs—remains unclear. Methods: We evaluated the impact of different reinforcement schedules on engagement levels with a mHealth app in a six-week, three-arm randomized intervention trial, while taking into account personality differences. Participants (i.e., university staff and students, N = 61) were awarded virtual points for performing health-related activities. Their performance was displayed via a dashboard, leaderboard, and newsfeed. Additionally, participants could win financial rewards. These rewards were distributed using a fixed schedule in the first study arm, and a variable schedule in the other arms. Furthermore, payouts were immediate in the first two arms, whereas payouts in the third arm were delayed. Results: All three reinforcement schedules had a similar impact on user engagement, although the variable schedule with immediate payouts was reported to have the lowest cost per participant. Additionally, the impact of financial rewards was affected by personal characteristics. Especially, individuals that were triggered by the rewards had a greater ability to defer gratification. Conclusion: When employing financial rewards in mHealth apps, variable reinforcement schedules with immediate payouts are preferred from the perspective of cost and impact.

The Experience Sampling Method (ESM) is commonly used to understand behaviors, thoughts, and feelings in the wild by collecting self-reports. Sustaining sufficient response rates, especially in long-running studies remains challenging. To avoid low response rates and dropouts, experimenters rely on their experience, proposed methodologies from earlier studies, trial and error, or the scarcely available participant behavior data from previous ESM protocols. This approach often fails in finding the acceptable study parameters, resulting in redesigning the protocol and repeating the experiment. Research has shown the potential of machine learning to personalize ESM protocols such that ESM prompts are delivered at opportune moments, leading to higher response rates. The corresponding training process is hindered due to the scarcity of open data in the ESM domain, causing a cold start, which could be mitigated by simulating participant behavior. Such simulations provide training data and insights for the experimenters to update their study design choices. Creating this simulation requires behavioral science, psychology, and programming expertise. Large language models (LLMs) have emerged as facilitators for information inquiry and programming, albeit random and occasionally unreliable. We aspire to assess the readiness of LLMs in an ESM use case. We conducted research using GPT-3.5 turbo-16k to tackle an ESM simulation problem. We explored several prompt design alternatives to generate ESM simulation programs, evaluated the output code in terms of semantics and syntax, and interviewed ESM practitioners. We found that engineering LLM-enabled ESM simulations have the potential to facilitate data generation, but they perpetuate trust and reliability challenges.

Human Digital Twins (HDTs) are specifically designed to represent humans virtually. Despite their recency, HDTs are considered to be powerful tools for enhancing personalized healthcare, early disease detection, improving patient outcomes, and optimizing lifestyle. Building HDTs is challenging especially due to the complexity of modeling human behavior. To unravel such complexity, in this paper we focus on specific aspects of human behavior and emotions (valence and arousal) to enable a more informed HTD behavior modeling. We assessed the feasibility and performance of our data collection infrastructure with N=112 participants. During a science festival, our participants wore a smartwatch to self-report their emotions while simultaneously their physiological data were collected through smartwatch sensors. We explored predictive modeling possibilities once with all the collected data and then with only the sensor-related data. The former performed best with 25 features, modeled with the Quadratic Discriminant Analysis classifier, resulting in 72.4% accuracy, however, with only 4 features derived from the self-reports, the K-Nearest Neighbors Classifier was able to achieve an accuracy of 71.6%. The latter used 17 sensor-related features and the Quadratic Discriminant Analysis classifier estimator reaching a 51.3% accuracy.