A method for intelligent allocation of diagnostic testing by leveraging data from commercial wearable devices: a case study on COVID-19

This study introduces an Intelligent Testing Allocation (ITA) method for optimizing diagnostic testing resource allocation during disease outbreaks. Leveraging data from the CovidInMyLife and MyPHd studies (15,345 participants, including 1,265 with smartwatch data), the researchers developed a model using resting heart rate (RHR) and step count data to predict COVID-19 infection. RHR proved a more sensitive early indicator than step count. The ITA model demonstrated an AUC of 0.77 in cross-validation and an AUC-PR of 0.55 ± 0.21 in independent testing, identifying both symptomatic and asymptomatic individuals. The study suggests that large-scale deployment of ITA could improve testing efficiency and reduce the strain of resource shortages.

Md Mobashir Hasan Shandhi, Peter J Cho, Ali R Roghanzadeh, Karnika Singh, Will Wang, Oana M Enache, Amanda Stern, Rami Sbahi, Bilge Tatar, Sean Fiscus, Qi Xuan Khoo, Yvonne Kuo, Joseph Hsieh, Alena Kalodizati, Amir Bahmani, Arash Alavi, Ustab Ray, Michael P Snyder, Geoffrey S Ginsburg, Daniel K Pasquale, Christopher W Woods, Ryan J Shaw, Jessilyn P Dunn