Assessing Accuracy of a Wrist-Worn Wearable Device Over Changing Levels of Activity

The widespread popularity of wearable devices has offered new avenues for continuous heart rate (HR) monitoring. A variety of other health related metrics including stress, sleep, and daily energy are being derived by heart rate and other sensors on the device. These or similar metrics could prove useful in critical care and other inpatient settings. This study evaluates the accuracy of a commercial smartwatch's photoplethysmography (PPG) sensor in tracking HR changes across various physical activities, comparing its performance to an electrocardiography (ECG)-based chest strap. Eighteen participants, aged 20 to 50, engaged in a series of tasks including rest, walking, and stair climbing, while HR data were collected via both devices. During rest, the smartwatch provided HR measurements within 5% of the chest strap values for a significant majority of readings (73.0% for the right wrist and 78.8% for the left). However, accuracy declined during physical activities, with the disparity becoming more pronounced during stair climbing (62.0% for the right wrist and 69.6% for the left). Excluding the initial 60 seconds of each activity from the analysis improved smartwatch accuracy, particularly during walking and the second rest period, as the smartwatches often failed to detect sharp changes in heart rate when the activity effort suddenly changed. This study highlights smartwatch reliability in static conditions and its limitations in accurately capturing rapid HR fluctuations during dynamic activities. These outcomes will impact their potential clinical utility in some cases. Copyright (C) 2024 The Authors. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)