COVID-19 literature is growing exponentially. Evidence appraisal and summarization has become a daunting task for clinicians and the public. Leveraging natural language processing research and open source tools that our lab has developed, we propose an automated method to structure the information in published articles on COVID-19 to enable fine-grained search and characterization of COVID-19 phenotypes, summarization of COVID-19 evidence, and publication retrieval based on population similarity, intervention, or outcomes.
