Consultants' forum: should post hoc sample size calculations be done?

Area of Science:

• Clinical Trials
• Biostatistics
• Medical Research

Background:

• Pre-study sample size calculations are mandatory for clinical trial protocols.
• Inadequate patient recruitment can lead to underpowered studies.
• Investigators may face non-statistically significant results with limited power.

Purpose of the Study:

• To debate the desirability of post-hoc power calculations.
• To discuss performing power calculations after data collection and analysis.

Main Methods:

• Literature review and critical analysis of statistical practices.
• Exploration of the implications of post-hoc power analysis in clinical research.

Main Results:

• The article presents arguments for and against post-hoc power calculations.
• It highlights the potential for misinterpretation and bias.

Conclusions:

• Post-hoc power calculations after data analysis are generally discouraged.
• Focus should remain on robust pre-study sample size determination and interpretation of results based on the original protocol.