The sample size justification is the rationale documented in the protocol and comprehensible from a regulatory perspective for how many participants or events a clinical study requires. It links the clinical question, the primary endpoint and the planned statistical analysis with realistic assumptions about the study’s ability to provide meaningful results.
Content of the justification
A sample size is not merely a calculation result. The justification explains which study question is to be answered, which analysis population is intended for the primary analysis and which assumptions underpin the calculation. These include, in particular, a clinically relevant treatment effect, the expected variability or event rate, the analysis method, the number of comparisons and the planned control of the probability of error. For time-to-event endpoints, the required number of observed events may be more important than the number of randomized participants alone.
The assumptions must be derived from the development programme, reliable prior information or clinical knowledge and justified transparently. In addition, the planning takes into account expected dropouts, missing data, non-treatment after randomization and, where applicable, the duration of recruitment and follow-up. ICH E9 requires the study size to be sufficient to answer the questions addressed reliably. An insufficient sample size can lead to unclear results; an unnecessarily large study, on the other hand, increases burden and resource consumption without a corresponding gain in knowledge.
Incorporation into the protocol
ICH E8(R1) requires the protocol to describe the justification of the sample size for key primary and secondary endpoints. It must be consistent with the study objective, Estimand, hypotheses, endpoint definition, randomization and Statistical Analysis Plan. If one of these elements changes substantially, it must be assessed whether the original justification still holds. Subsequent adjustment of the sample size solely because of disappointing or favourable observed results would not be consistent with prospectively controlled planning.
For planned interim analyses or adaptive elements, the rules, including possible adaptations, their decision basis and control of the type I error, must be specified in advance. A sample size adjustment may be methodologically permissible if it is part of a planned adaptive design. However, it requires a comprehensible statistical justification and must not jeopardize the integrity of the study, blinding or the interpretability of the treatment effect.
Distinction from statistical power
Statistical power is a measure: under the assumptions of a particular test, it describes the probability of detecting a genuinely existing, prospectively defined effect. It is therefore one component of a sample size calculation. The sample size justification is broader in scope. It is the regulatory-required explanation in the protocol of why the selected study size can answer the clinical question with the planned design and planned analysis.
Power and sample size justification must therefore not be equated. A statement such as “the study has sufficient power” explains neither the relevant treatment effect nor the choice of endpoint, the analysis population, the assumptions regarding dropouts or the consequences of multiplicity and interim analyses. Conversely, formally calculated power cannot justify an implausible clinical assumption. The separate entry Statistical Power addresses the measure; this entry addresses its incorporation into the protocol justification.
The calculation should also not disregard practical feasibility. A theoretically sufficient number can only be achieved if inclusion and exclusion criteria, the number of suitable sites, recruitment capacity and expected follow-up are planned realistically. If these operational assumptions are not met, the amount of information available at the end may be lower than provided for in the justification. Sample size, recruitment strategy and data quality must therefore be planned together.
Relevance for clinical trials
A robust sample size justification is an interface between medical development, study operations and regulatory assessment. It supports the planning of sites, recruitment duration, supply quantities and data maturity. Investigators, ethics committees and authorities must be able to see that participants are not enrolled without a sufficient prospect of obtaining an informative answer. During review, the consistency of the assumptions and the prespecified analysis are assessed in particular.
Full-service CROs such as Mediconomics support scenario calculations, documentation of the assumptions and coordination of the protocol and Statistical Analysis Plan. Specific services include selection of the calculation method, planning of recruitment and follow-up, consideration of dropouts and interim analyses, as well as biostatistical programming and description of the sample size in the protocol and the clinical study report.
Frequently Asked Questions (FAQ)
Is a large sample size always better?
No. It must fit the clinical question and the endpoint. An unnecessarily large study can place a greater burden on participants and tie up resources without meaningfully improving the decision.
Can the sample size be changed during the study?
Only within a prospectively described, methodologically justified procedure, such as an adaptive design. The change must safeguard error control and the integrity of the study.
Why are dropouts included in the planning?
Because missing endpoint data and study discontinuations can reduce the information available for the primary analysis. Their expected frequency and how they will be handled must be taken into account.
Regulatory references
- ICH E9 “Statistical Principles for Clinical Trials” – requires a study size that enables reliable answers to the study questions.
- ICH E8(R1) “General Considerations for Clinical Studies” – requires the sample size justification to be included in the protocol.
- ICH E9(R1) “Addendum on Estimands and Sensitivity Analysis” – links the study objective with the Estimand and analysis.
- EMA Reflection Paper CHMP/EWP/2459/02 – addresses prospectively planned adaptations in adaptive confirmatory studies.