The lay abstract featured today (for Individualized Time-Varying Nonparametric Model With an Application in Mobile Health by Jenifer Rim, Qi Xu, Xiwei Tang, Yuqing Guo and Annie Qu) is from Statistics in Medicine with the full article now available to read here.

How to cite

Rim, J., Xu, Q., Tang, X., Guo, Y. and Qu, A. (2025), Individualized Time-Varying Nonparametric Model With an Application in Mobile Health. Statistics in Medicine, 44: e70005. https://doi.org/10.1002/sim.70005

Lay Abstract

The rise of wearable technology, such as smartwatches and the Oura ring, has revolutionized personalized healthcare by enabling the continuous tracking of health behaviors like physical activity and sleep patterns. This wealth of mobile health data offers exciting opportunities to develop more tailored health interventions. However, because people have unique lifestyles and ways of using these devices, the data often vary significantly between individuals. This highlights the importance of accounting for individual differences, or heterogeneity, in analyzing this type of data. By modeling these dynamic, individual-specific effects, researchers can pinpoint the best times to offer personalized interventions and activities. This research focuses on understanding how the effects of physical activities, such as exercise, change over time for each person. This method also identifies groups of individuals who share similar patterns, helping researchers gain broader insights while still tailoring recommendations to individual needs. Importantly, it addresses common challenges in mobile health studies, such as missing data, to ensure robust and accurate results. To illustrate its potential, the method was applied to data from pregnant women to explore how physical activity impacts deep sleep—a vital factor for recovery, energy, and overall well-being. For pregnant women, deep sleep is particularly crucial for their health and the development of their babies. Beyond pregnancy, this approach can be applied to studies examining how factors like diet or medical treatments influence health over time. By uncovering these dynamic relationships, the method offers a powerful tool for researchers aiming to deliver more effective, personalized health solutions.