Does the Result Matter? Moving from P Values to Patient-Relevant Effects
A small p value does not prove that an effect matters to patients. Evaluate effect size, confidence intervals and minimally important differences.
Two different questions
Statistical significance asks how compatible the data are with a specified null model. Clinical importance asks whether the magnitude matters to patients, clinicians or health systems. A huge sample can make a trivial effect statistically significant, while an important effect can remain uncertain in a small study.
Report an effect estimate and uncertainty
Use an appropriate measure—mean difference, risk difference, risk ratio, odds ratio, hazard ratio or standardised effect—with a confidence interval. The p value does not communicate magnitude.
What is an MCID?
A minimally clinically important difference is the smallest change considered important enough to influence patient experience or decisions. It is context-dependent and can vary by population, baseline severity, measurement instrument, treatment burden and elicitation method. Anchor-based information, distribution-based evidence and patient input should be combined where possible.
Compare the confidence interval with the clinical threshold
An interval entirely beyond the benefit threshold supports an important effect.
An interval excluding zero but lying below the MCID suggests statistical significance with limited clinical importance.
An interval spanning harm and important benefit indicates uncertainty.
An interval confined to clinically trivial values can provide evidence against a meaningful effect.
Absolute effects matter
A 50% relative risk reduction from 2% to 1% is a 1-percentage-point absolute reduction and an NNT of about 100. The same relative reduction from 40% to 20% has a very different clinical meaning. Benefits should be considered alongside harms, cost and patient preferences.
Reporting example
“The intervention reduced the score by 2.1 points (95% CI 0.9–3.3; p=0.001). Because the prespecified minimally important difference was 5 points, the effect is statistically significant but unlikely to be clinically important.”
References
Jaeschke R, Singer J, Guyatt GH. Minimal clinically important difference. 1989.
Wasserstein RL et al. Moving to a World Beyond p<0.05. 2019.
Accessed 20 June 2026.