Visually assessing work performance using a smartwatch via day-to-day fluctuations in heart rate variability

To optimize workplace health promotion, a simple method for quantifying allostatic load response is needed. This study examines the feasibility of optimizing objective anxiety and presenteeism monitoring using daily smartwatch-measured ultra-short heart rate variability (HRV).

Office workers without diagnosed disease prospectively performed 30 s HRV self-measurement each morning for two months and responded to the State-Trait Anxiety Inventory (STAI) and Work Limitation Questionnaire (WLQ). Logistic regression analysis examined daily HRV parameters in the high-trait anxiety group (HTA, STAI ≥ 40) using mean and variance HRV, age, self-reported gender, and body mass index (BMI). The ideal cutoff value enabled comparison of WLQ using the Mann-Whitney U test. Heart rate variability data were collected for 279 participants (male ratio, 83.9%; age, 42 ± 10 years) who completed questionnaires and monitored HRV for 30+ days.

Compared to the low-trait anxiety group, HTA exhibited higher variance of the log-transformed coefficient of component variance of high-frequency component (LnccvHF) and low-frequency per HF (Lnccv L/H), in addition to differences in the means of these HRV parameters. In addition to BMI (odds ratio [OR] = 0.92, p = 0.02) and mean LnccvL/H (OR = 10.75, p < 0.01), the variance of Lnccv L/H was an independent predictor of HTA (OR = 2.39E + 8, p = 0.011). The daily Lnccv L/H dispersion group had a lower WLQ productivity loss score (p = 0.02, r = 0.17).

By focusing on HRV dispersion status, this simple and instantly applicable daily HRV monitoring system enables optimized quantitative monitoring of anxiety and productivity.