Managing Data Changes - Clinical Study Level

The sponsor pharma selected a CRO to monitor and process the clinical study data and database. The CRO now must maintain an audit trail to document any data changes that occur after data entry in the clinical study database. What should the data audit trail contain? At the very least the data audit trail should contain:

• sponsor pharma name/protocol number
• investigator number
• patient number
• dataset name
• CRF page
• variable name
• old value
• new value
• date/timestamp of change
• person who made change
• reason for change.

There are generally two types of data changes. Data changes arising from DCFs generated by a CRO monitor, CRA, CRC, DM, QC, QA or quality control process and "obvious errors". Be careful making data changes pertaining to "obvious errors". Be careful of "obvious errors".

Obvious errors are corrections that result in data changes that can be made "without" a confirmatory DCF issued. These corrections are self-evident based on the review of corresponding CRF data. There must be Generic Guidelines as well for Handling Obvious Errors and most importantly, the sponsor pharma is the only authority that can approve Obvious Error corrections, not the clinical personnel at the site, not the investigator. Final approval is only made by the sponsor pharma which is ultimately responsible for the integrity of the final clinical study data and clinical study database.

Monitor the monitor, monitor the data, monitor changes to the data, monitor the audit trail - as a sponsor take an active role in the quality control of the data and database at the clinical study level.