Aligning Noninferiority Assumptions and Decision Rules in a Protocol of Adjunctive Acupuncture for Late-Life Depression (Preprint)

UNSTRUCTURED Background: Noninferiority trials require strict alignment between the primary endpoint, assumptions used for sample-size justification, and prespecified decision rules. Objective: To request clarification and improved transparency in a recently published protocol of adjunctive acupuncture plus citalopram versus citalopram alone for late-life depression. Key Points: The protocol defines the primary endpoint as HAMD-17 response at week 12, yet the assumed control event rate appears to be justified using STAR*D figures commonly cited as remission rates. Because response and remission are distinct endpoints, this may undermine endpoint-consistent design assumptions. In addition, the protocol would benefit from an explicit noninferiority decision rule for the binary endpoint (effect measure, confidence-interval approach, and comparison to the prespecified margin) and harmonization of one-sided versus two-sided alpha statements. Finally, given potential sensitivity of noninferiority conclusions to analysis populations and missing data, prespecifying the roles of intention-to-treat and per-protocol analyses and defining missing-data handling consistent with the primary endpoint and decision rule would reduce interpretive ambiguity. Conclusion: These clarifications would strengthen methodological transparency and reproducibility before trial results are reported.