Defining the Habitome: Phenotypes of Routine and Their Relationship to Health Outcomes

Human health is deeply interconnected with behaviors that are dictated by daily routines, including activities such as sleep, physical activity, and dietary habits. Wearable technologies like smartwatches provide continuous, real-world insights into these behaviors and have presented a new opportunity to study habits and their relationship to health. Leveraging data from the NIH All of Us Research Program, we analyzed over 13.9 million days of Fitbit data from 15,389 participants spanning 5 years to define "routine-otypes"—quantifiable phenotypes of behavioral routines. Unsupervised learning identified distinct patterns of device wear across seasonal, weekly, and daily dimensions, uncovering seven long-term wear patterns, six circadian wear patterns including behaviors like daytime-only wear, nighttime-only wear, and precise charging times, and fourteen weekly wear patterns. Ultimately, we reveal five dimensions of routine-otypes, including circadian, weekly, and seasonal patterns, as well as overall wear levels and habituality. Routine-otypes were significantly associated with demographic factors such as age, sex, and income. Importantly, routine-otypes showed strong associations with health outcomes like depression and chronic disease, opening the possibility of developing new digital biomarkers for these conditions. This study introduces routine-otypes as objective, scalable, and modifiable human characteristics and highlights their potential to act as digital biomarkers of disease. By proposing the "habitome" framework, this work lays the foundation for developing deeper insights into routines and their role in health.