- Published
- Updated
Revisiting sample size assumptions during a trial differs from examining treatment effects and choosing a convenient size. Define the information used, permitted modifications and upper bounds prospectively so the reassessment has a clear scientific and operational role.
Define the question and assumptions
Identify the uncertain assumption being revisited, such as variability or event frequency. Explain how comparative treatment information is excluded and consider additional risks in open-label settings. Use specialist review to assess consequences for error control and estimation; a blinded label alone does not establish a valid noncomparative decision process.
Hypothetical study example
In a hypothetical study, outcome variability exceeds the planning assumption. The team applies the prespecified noncomparative rule alongside an upper sample limit and operational constraints. It does not choose the increase after seeing a favourable treatment difference, and preserves the data and calculation trail used for the reassessment.
Design review checks
- Define the targeted assumption and the permitted range of design modification.
- Assess whether treatment comparisons could indirectly influence the reassessment decision.
- Review consequences of the revised size for resources, duration and participant burden.
Record the decision and its limits
An accepted decision follows the specified method under the actual information conditions. Explain the scientific rationale separately from budget approval. Reassessment cannot repair every planning problem: eligibility and endpoint-definition issues need their own review. Report the implemented adaptation and remaining limitations transparently in the final study account.
Frequently asked questions
- Is blinded reassessment risk-free in every setting?
- No. Information content, methodology and access still need assessment.
- Does increasing sample size repair a poor design?
- Not alone. Measurement and scientific-question problems can remain.