Research findings presented orally at the 99th Annual Meeting of the Japan Society for Occupational Health. Expected applications in personal monitoring and health guidance -

Tech Doctor Inc. (Representative Director: Kazumichi Minato; Headquarters: Chuo-ku, Tokyo; hereinafter “Tech Doctor”) analyzed wearable data collected over approximately one year and three months from approximately 2,000 employees of research partner companies and established age- and sex-specific reference values for physical activity, sleep, and heart rate variability. The findings of this study were presented orally at the 99th Annual Meeting of the Japan Society for Occupational Health (held from May 27–30, 2026).
In this study, more than 500,000 person-days of data obtained from Google Fitbit devices were utilized to establish reference values (median values and 95% prediction intervals) for 16 indicators related to physical activity, sleep, and heart rate variability among working-age Japanese individuals. In addition to changes associated with aging, sex differences, and temperature, associations with medical history and metabolic syndrome were also suggested, demonstrating the potential application of wearable data for personal monitoring and health guidance.

◾️Research Background

In recent years, the use of daily-life data obtained from wearable devices has advanced, and numerous studies have reported associations between such data and health conditions or disease risks.
For example, a meta-analysis of 15 international cohort studies reported that increased daily step counts were significantly associated with reduced risks of all-cause mortality and chronic diseases (Paluch et al., Lancet Public Health, 2022). In addition, associations between Fitbit-derived step count data and reduced risks of diabetes, hypertension, and obesity have also been reported (Brittain et al., Nature Medicine, 2022).
Meanwhile, in order to utilize wearable data for health management and disease prediction, reference values are required to determine whether a given value falls within the normal range. In particular, for indicators that vary depending on age and sex, the development of reference values based on Japanese populations is needed.
While large-scale studies involving tens of thousands to millions of participants are progressing overseas, research in Japan has primarily focused on studies involving only several dozen participants, and large-scale reference value studies targeting Japanese populations have been limited.

◾️Research Objective

Using long-term wearable data obtained from Google Fitbit devices worn by employees of research partner companies, this study aimed to establish age- and sex-specific reference values for indicators related to physical activity, sleep, and heart rate variability.

Primary Objectives:
Construction of age- and sex-specific reference values (median values and 95% prediction intervals) for biometric indicators obtained from Fitbit devices
Construction of Bayesian statistical models considering the effects of aging, sex differences, and temperature

Secondary Objectives:
Examination of associations between the established reference values and health checkup data
Evaluation of the usefulness of the reference values through associations with diseases and health conditions

◾️Research Methods

Subjects
1,996 employees aged 25–65 years working at companies
Male: 1,260
Female: 736

Devices and Measurement Period
Device: Google Fitbit
Measurement period: January 2023 – November 2025
Average wearing period: 435 days

Collected Data
Minute-by-minute physical activity data
Sleep stage classification data (Wake / REM / Light / Deep)
Heart rate variability during sleep

Number of Participants Included in Analysis
*Filtered according to analysis conditions
Physical activity indicators: 1,920 participants
Sleep duration: 1,899 participants
Sleep stages and heart rate variability: 1,934 participants

■Research Results

Reference Values for Heart Rate Variability (SDNN)
Among younger participants, the median values for males were higher, with males showing 32 ms and females 27 ms at age 25. However, the values declined with aging, with a greater decline observed in males, and the sex difference tended to diminish around age 60.

Reference Values for Physical Activity
Step counts consistently tended to be higher in males. Meanwhile, no major age-related changes were observed. Total physical activity (TPA) showed an increasing trend with aging, suggesting an association with increased activity time due to reduced sleep duration.

Reference Values for Sleep Indicators
Sleep duration decreased with aging in both males and females. At age 25, sleep duration was 7.1 hours for males and 7.7 hours for females, whereas at age 65, it was 6.6 hours for males and 6.7 hours for females. In addition, the proportion of deep sleep also declined with aging, confirming a trend toward lighter sleep.

Associations with Diseases and Health Conditions
In addition to wearable data, associations with health checkup results were also analyzed. Participants with a medical history or those meeting the criteria for metabolic syndrome showed reduced heart rate variability and elevated heart rate. These findings suggested that deviations from the established reference values may be associated with health risks.

■Summary of Research Findings

In this study, age- and sex-specific reference values were established for 16 indicators related to physical activity, sleep, and heart rate variability using wearable data from more than 1,900 participants and over 500,000 person-days.
In particular, the study demonstrated that:

Heart rate variability and sleep indicators decline with aging
Physical activity shows relatively small changes with aging
Sex differences exist in heart rate variability and physical activity
Deviations from the reference values may be associated with health risks

■Social Significance and Future Prospects

In this study, reference values for physical activity, sleep, and heart rate variability were established based on more than 500,000 person-days of wearable data collected from approximately 2,000 participants, representing one of the largest wearable data studies conducted in Japan. Furthermore, through analyses of associations with health checkup results, the possibility that deviations from the reference values are associated with health risks was also confirmed.
These findings are expected not only to serve as a foundation for more objective assessment of individual health conditions, but also to contribute in the future to early detection of lifestyle-related disease risks, personalized health guidance, applications in occupational health, and the development of digital biomarkers.
Tech Doctor will continue to pursue the realization of new healthcare and medical solutions based on daily-life data through ongoing research and analysis utilizing wearable data.

■Conference Presentation Overview

Conference Name: 99th Annual Meeting of the Japan Society for Occupational Health
Presentation Date: May 30, 2026
Presentation Format: Oral Presentation
Official Website: 99th Annual Meeting of the Japan Society for Occupational Health

■References

Paluch AE, et al. Daily steps and all-cause mortality: a meta-analysis of 15 international cohorts. Lancet Public Health. 2022.
DOI: 10.1016/S2468-2667(21)00302-9

Brittain EL, et al. Association of step counts over time with the risk of chronic disease in the All of Us Research Program. Nature Medicine. 2022.
DOI: 10.1038/s41591-022-02012-w