A Comprehensive and Curated Dataset of Covid-19 and Epistemonikos Evidence

The emergence of COVID-19 has highlighted the importance of reliable information for clinical decision-making and public health policies. Evidence-based medicine (EBM) seeks to identify and evaluate scientific documents related to novel diseases, and biomedical text classification is crucial in accurately categorizing such documents. To aid this process, we present a comprehensive dataset of COVID-19-related documents, with 18,854 labeled documents that include document type, title, abstract, and metadata such as pubmed id, authors, journal, and publication date. The dataset is labeled by collaborators from Epistemonikos and contains five document types: systematic reviews (SR), primary study randomized controlled trials (PS-RCT), primary study non-randomized controlled trials (PS-NRCT), broad synthesis (BS), and excluded (EXC). Additionally, we offer an evidence dataset with 399,737 non-COVID-19 documents labeled with SR, PS-RCT, PS, BS, and EXC, which can serve as a baseline for future research. We believe that this open-access dataset and accompanying resources will help advance the field of evidence-based medicine and facilitate further research.