Skip to main content
. 2023 Dec 22;6:239. doi: 10.1038/s41746-023-00974-w

Fig. 1. Workflow of study to track COVID-19 symptoms and physical activity of healthcare workers over a 600-day period.

Fig. 1

Physical activity data was collected using HealthKit on iOS devices (iPhones and Apple Watches) from the first day of COVID-19 symptom onset, serology test or PCR diagnosis. Current and previous symptoms were self-reported at clinical visits. Unsupervised machine learning methods were applied to the longitudinal activity (e.g. heart rate variability (heartRateVariabilitySDNN)) and symptoms (loss of taste) to classify each individual healthcare worker as having high/low physical activity (light green/dark green) and long/short COVID-19 symptoms (dark blue/light blue). We can subsequently taste whether these activity and symptom patterns are associated.