An Analysis of COVID-19 Recovery Duration: Smartwatch Activity and Self-Report

Area of Science:

• Digital epidemiology
• Wearable technology in health
• Post-COVID-19 recovery

Background:

• Understanding COVID-19 recovery is crucial for patient management and public health.
• Objective assessment of recovery duration is challenging using traditional methods.

Purpose of the Study:

• To investigate the correlation between smartphone-derived activity data and self-reported COVID-19 recovery.
• To develop predictive models for COVID-19 recovery duration using digital and self-reported data.

Main Methods:

• Development of the COronaVIden app to collect activity data (iOS Health, Google Fit) and patient-reported outcomes (symptoms, well-being).
• Analysis of data from over 5,000 participants, with 177 meeting inclusion criteria for COVID-19 recovery.
• Creation of two predictive models incorporating physical activity, demographics, and post-infection symptoms, evaluated against baseline recovery duration.

Main Results:

• Discrepancy observed between objectively measured recovery (23.04 days) and self-perceived recovery (16.51 days).
• Predictive models showed significant improvement over baseline (29% and 37% MAE reduction).
• Female sex, older age, and fatigue were associated with longer recovery times; higher pre-infection physical activity showed a trend towards shorter recovery.

Conclusions:

• Smartphone and wearable data offer a viable method for tracking disease recovery patterns.
• Digital health tools can provide valuable insights for public health strategies and personalized patient care.
• Further research is needed to address confounding factors and refine predictive accuracy for COVID-19 recovery.