CliniGA CRO

CliniGA CRO Blog

Continuing safety review after coding version changes

Separate coding revisions from new clinical events and preserve the meaning of comparisons and earlier assessments.

Representative research photograph: Continuing safety review after coding version changes
Photo: Official U.S. Navy Page from United States of America · License: Public domain · Cropped and resized.
Published
Updated

Updating a coding dictionary may change displayed terms or event groupings. This does not establish that new clinical events occurred. Safety review should explain how revised codes affect trends and previous decisions when the underlying source information remains unchanged.

Scope of the decision

Keep the coding version, data cut and scope of recoding visible. Suitably competent staff should assess relevant automated changes instead of assuming clinical meaning is preserved. Review involves more than completion of a technical conversion: consider comparability and any need for further assessment.

Worked review

For example, events previously reported under one heading are split between two groups in a new version. A chart now shows a decrease in one group. Before presenting this as improved safety, the team compares old and new classifications using the same source records and explains the contribution of the definition change.

Practical checks

  • Preserve relationships between source text, previous codes and revised codes.
  • State the version and recoding scope used in trend reports.
  • Refer important grouping changes for assessment by the appropriate clinical expert.

Acceptance and limits

Acceptance requires traceable conversion and explained interpretation limits. A dictionary update neither automatically invalidates nor confirms an earlier clinical judgement. Relevant experts should separately assess whether the revised classification provides information important to the existing decision.

Frequently asked questions

Should a coding update change source wording?
Preserve source authenticity; coding and correction of source information are different activities.
Does a lower chart value show real improvement?
Not by itself. Distinguish a classification change from a change in actual events.

Sources

  1. EMA — ICH E2A clinical safety data management
  2. MHRA — GxP data integrity