Sample size estimation for the sorcerer's apprentice. Guide for the uninitiated and intimidated

Area of Science:

• Clinical research methodology
• Biostatistics
• Primary care science

Background:

• Sample size determination is essential for robust clinical study design.
• Accurate sample size ensures reliable and valid research outcomes in primary care.

Purpose of the Study:

• To review the importance and practical application of sample size determination.
• To guide primary care researchers in estimating appropriate sample sizes for their studies.

Main Methods:

• Literature search of MEDLINE (1966-1998) using keywords: sample size, sample estimation, study design.
• Consultation of medical statistics texts, article references, and expert colleagues.
• Selection of resources offering clear, simple approaches to sample size estimation relevant to primary care.

Main Results:

• Key statistical parameters include alpha error (usually 5%), beta error (commonly 20%), and study power (typically 80%).
• Examples provided for sample size estimation for comparing means, event rates, odds ratios, correlation coefficients, and meta-analyses.
• Pre-study sample size estimation enhances statistical power and minimizes bias, thereby increasing study validity.

Conclusions:

• Sample size estimation is an accessible process for novice researchers.
• Proper sample size determination is fundamental for maximizing study quality.