# Biostrap Kairos Wristband Versus Electrocardiography for Resting Heart Rate Variability Assessment

## TL;DR

The Kairos wristband accurately measures resting heart rate but has limited accuracy for heart rate variability metrics compared to electrocardiography.

## Contribution

This study evaluates the validity of the Kairos wristband for HRV assessment and identifies specific metrics with acceptable agreement against ECG.

## Key findings

- Resting heart rate showed excellent agreement between Kairos and ECG.
- SDNN had the strongest agreement for global HRV, while SD1 and RMSSD showed moderate agreement for parasympathetic activity.
- Higher HRV values were associated with greater measurement error and underestimation by the Kairos device.

## Abstract

The Kairos wristband offers on-demand heart rate variability (HRV) assessment through its “Spot Check” feature, enabling standardized recordings for clinical, research, or self-tracking purposes, but its validity is untested. Therefore, we compared the Kairos wristband to electrocardiography (ECG) for resting HRV assessment in young adults, and investigated the influence of skin pigmentation (M-index) on measurement accuracy. Simultaneous 3 min Kairos and ECG samples were obtained in the supine (n = 32) and seated (n = 30) position. Comparisons included resting heart rate (RHR) and time domain (root-mean square of successive differences [RMSSD], standard deviation of normal RR intervals [SDNN]), frequency domain (low [LF] and high frequency [HF]), and non-linear (standard deviation 1 [SD1] and SD2) HRV metrics. RHR showed excellent agreement whereas HF, LF, and SD2 showed poor agreement. For the remaining metrics, SDNN showed the strongest absolute and relative agreement, followed by SD1 and RMSSD. However, most HRV metrics exhibited heteroscedasticity or proportional bias, with greater error and underestimation at higher HRV values. M-index was unrelated to method difference scores, except for seated SD2 (p = 0.01). The Kairos wristband can be used to measure RHR, but HRV assessment should be limited to SDNN for global variability and SD1 or RMSSD for cardiac–parasympathetic activity. However, these metrics should be interpreted within the level of agreement identified in this study, and with consideration of the observed trend of diminished accuracy with higher HRV values.