Frontiers in Digital Health (Dec 2021)
Data Quality and Network Considerations for Mobile Contact Tracing and Health Monitoring
Abstract
Machine Learning (ML) has been a useful tool for scientific advancement during the COVID-19 pandemic. Contact tracing apps are just one area reaping the benefits, as ML can use location and health data from these apps to forecast virus spread, predict “hotspots,” and identify vulnerable groups. However, to do so, it is first important to ensure that the dataset these apps yield is accurate, free of biases, and reliable, as any flaw can directly influence ML predictions. Given the lack of criteria to help ensure this, we present two requirements for those exploring using ML to follow. The requirements we presented work to uphold international data quality standards put forth for ML. We then identify where our requirements can be met, as countries have varying contact tracing apps and smartphone usages. Lastly, the advantages, limitations, and ethical considerations of our approach are discussed.
Keywords
