Label the subgroup analysis as exploratory, report it with an interaction test and a caveat, and document the status change and its interpretation limit. The repair is a reporting and interpretation fix: the data stay the same, and the claim attached to them changes.
- Check the protocol and any registered analysis plan for a prespecified subgroup list. If the subgroup is absent, record that finding and mark the analysis as exploratory in the manuscript, so readers can see the status change.
- Report the subgroup comparison with an interaction test and a caveat. Cochrane guidance on interpreting results emphasizes confidence intervals, exact P values, and responsible conclusions, so present the subgroup effect with its uncertainty and describe it as a hypothesis-generating observation.
- Document the status change and its interpretation limit in the methods or limitations, stating that the subgroup was identified after the primary analysis and that its result needs confirmation in independent data.
Why the label matters
A prespecified subgroup carries a planned question; an unprespecified one carries a question raised by the data. The practical consequence is that the second cannot support the same strength of claim. Cochrane's meta-analysis guidance stresses careful interpretation of pooled results, and its interpretation chapter stresses applicability and certainty, which is why an exploratory subgroup should be framed as a lead for further study.
What to write
State the subgroup, the interaction test result with its confidence interval and exact P value, and one sentence on the interpretation limit. If the manuscript also rates certainty, GRADE's outcome-level judgements give a transparent structure for recording how the exploratory status affects the rating.