Validation of the ECTemp™ algorithm in construction workers: a comparison of photoplethysmography smartwatch and...

Background. Core temperature monitoring may provide information to mitigate heat-related illnesses among construction workers. This study evaluated the validity of the ECTemp™ algorithm and compared the use of a photoplethysmography-based smartwatch to a chest strap as the heart rate input for the algorithm. Methods. In 102 construction workers during the summer, gastrointestinal temperature was measured via ingestible telemetric pill (T _PILL), and heart rate was measured with a Polar H10 chest strap (n = 102) and a Garmin Instinct 2 smartwatch (n = 60) during a single work shift. Peak, mean and continuous time-series data from each device were used to compute core temperature estimates using ECTemp™ (ECTemp_GARMIN and ECTemp_POLAR). Agreement was evaluated with Bland-Altman analyses, Pearson correlations, and Kendall's Tau correlations. Results. Compared to T_PILL, ECTemp_POLAR showed a bias of -0.16 ± 0.30 °C for continuous data, with a median Kendall's τ of 0.39. ECTemp_GARMIN showed a bias of -0.09 °± 0.28 °C (median τ = 0.36). The two devices produced similar heart rate measures (bias: 1.6 ± 4.9 beats min^-1; median τ = 0.75) and ECTemp outputs (bias: 0.03 °± 0.06 °C; median τ = 0.83). Conclusion. ECTemp_POLAR and ECTemp_GARMIN are both adequate estimates of core temperature for construction work monitoring at a group level. The more practical ECTemp_GARMIN appeared to have less bias than ECTemp_POLAR, perhaps warranting future consideration for field use.