Sample size calculations for pilot cluster-randomised controlled trial: a review 2010-2020

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Health Services Research

Background:

• No standardized guidance exists for pilot cluster-randomised controlled trials (cRCTs) sample sizes.
• Pilot trials should focus on feasibility, not efficacy, and power calculations are inappropriate for sample size justification.
• CONSORT guidelines require sample size justification for pilot trials.

Purpose of the Study:

• To review sample size choices and justifications in pilot cRCTs.
• To analyze trends in pilot cRCT sample sizes over time.
• To explore the impact of CONSORT guideline publications on study practices.

Main Methods:

• Systematic search of PubMed and Web of Science for pilot/feasibility cRCTs published from 2010 onwards.
• Data extraction focused on planned sample size (clusters per arm), participants per arm, and enrolled sample size.
• Descriptive and graphical analyses were employed.

Main Results:

• 170 pilot or feasibility cRCTs were identified.
• The median sample size was consistently four clusters per arm.
• Sample size remained stable regardless of justification, planned analyses, or study characteristics.

Conclusions:

• Pilot cRCT sample sizes have been static between 2010-2020.
• Stated study aims and justifications for sample size do not appear to influence the chosen sample size.
• Sample sizes may be determined by convenience or practical considerations rather than research objectives.