| Amy Zhang |
During a clinical trial, it is common to extract data from a live database for various purposes, such as medical monitoring,
TLFs development, or evaluation of the safety profile for an investigational drug. These data extractions can happen
multiple times throughout a trial lifecycle, and it is beneficial for statistical programmers to understand the overall context
of the data at each data transfer. Some particular questions we want to answer are how many datasets were included in
this transfer and were there any missing or new datasets? Have there been changes to the dataset structure? Were any
variables or their controlled terms introduced or removed? Was new value-level metadata added to the finding domains?
Due to the interdependent nature of SDTM and ADaM codes, without awareness of all the updates, such changes can
result in potential mapping errors that impact ADaM programming or the contents of the final outputs. Analyzing all these
aspects manually can be time-consuming. To address this problem, we developed a utility tool to check the contents of
SDTM datasets and produce a summary report of differences between the newest delivery and previous deliveries. We
assembled a customized list of common changes that we expect to occur or need to pay close attention to across the
SDTM domains, and implemented SAS® macros to automatically summarize and output the information we are
interested in. |