Geometric Mean: When and Why It Matters in Clinical Research

Definition:The geometric mean is the nth root of the product of n values, used to summarize positive, skewed data. It is defined as:

Best Use Case:

• Log-normal data (e.g., concentrations, biomarkers, fold changes, viral loads)

• When multiplicative effects or relative changes are clinically relevant

• In meta-analysis of ratios (e.g., odds ratios, hazard ratios)
  

 Formulas

Arithmetic Mean (AM) AM = (x₁ + x₂ + ... + xₙ) / n

Geometric Mean (GM) GM = ⁿ√(x₁ × x₂ × ... × xₙ) —or— GM = exp[(1/n) × (ln x₁ + ln x₂ + ... + ln xₙ)]

⚖️ Why Not Arithmetic Mean?

Feature	Geometric Mean	Arithmetic Mean
Handles skewness	✅ Yes	❌ Sensitive
Allows log scale	✅ Yes	❌ No
Zero-tolerant	❌ No	✅ Yes
Best for growth	✅ Yes	❌ No

🧪 Clinical Example:

Suppose you're analyzing urinary albumin-to-creatinine ratios (UACR). Because these values are right-skewed, the geometric mean better represents the central tendency than the arithmetic mean.

If UACRs are 10, 20, 80 mg/g:

• Arithmetic mean = 36.7

• Geometric mean = ³√(10 × 20 × 80) ≈ 26.7
  

Interpretation: 26.7 better reflects the typical value in skewed data.

Let me know if you want a log-transformed regression interpretation or how to report GM with CI.