Use of CDISC Standards in the RECOVERY Trial of COVID-19 Treatments
Dr. William Stevens, Ph.D, Nuffield Department of Population Health, Oxford University
The RECOVERY COVID-19 platform trial has identified several effective treatments (including dexamethasone), and several ineffective treatments. Data comes from an eCRF and electronic healthcare systems, this is mapped to SDTM and ADaM. Existing tools were used, and datasets were ready early in the pandemic.
Treatments are added and removed over time, protocols and CRFs are under continual revision, and analyses and regulatory submissions are needed for some treatments while others continue. Workflow and version control considerations are important.
Data sharing agreements mean that dates cannot be included in SDTM. Not all SDTM date variables have corresponding day variables, and some validation rules do not work using day variables alone. Validation was carried out in two steps.
Sometimes disagreements are expected in electronic healthcare data, this means that SDTM cannot faithfully represent what happened to the participant – a best estimate is obtained when deriving ADaM data.