An OMOP Common Data Model-Compliant Annotated Corpus for COVID-19 Clinical Trials
This README file describes the dataset accompanying the above publication.
It includes the original free text eligibility criteria (.txt) and annotations (.ann) for each trial. The annotation file (.ann）can be opened by Brat.
Six commonly used criteria across all COVID-19 trials about current age, high-risk status (e.g., hospital worker), COVID-19 status (e.g., yes, cleared, no), days since diagnosis (if relevant), current hospitalization or intensive care unit (ICU) admission, and pregnancy status are specified as “key criteria” and recorded in this .CSV table.
Metadata information of each trial are included in this .CSV file, such title, location, study type, phase and etc.
Structured eligibility criteria with normalized concepts are included in this .CSV file.
Main columns:
nct_id: clinical trial ID in clinicaltrials.gov
entity_source_text: segment of the original description of the entity in the clinical trial eligibility criteria source text
start_index: the position in the criteria source text where the entity starts
end_index: the position in the criteria source text where the entity ends
concept_id: mapped concept ID in the standard vocabulary of the entity
concept_name: mapped concept name in the standard vocabulary of the entity
domain: domain ('Drug','Condition','Procedure','Device','Measurement') of the concept
temporal_source_text: segment of the corresponded temporal attribute of the entity in original criteria text
days: normalized temporal attribute as the number of days
numeric_source_text: segment of the corresponded numerical attribute of the entity in original criteria text
numeric_att_min: lower bound of the normalized numerical attribute
numeric_att_max: upper bound of the normalized numerical attribute
is_exclusion: '1' for exclusion criteria and '0' for inclusion criteria