Evidence appraisal
Read the question, study design, effect size, uncertainty and bias before applying evidence to a patient.
Understand
A good appraisal starts with the clinical question, then examines design, bias, effect size, precision and applicability. A p-value alone cannot establish usefulness.
Assess
Ask whether the comparison and outcome matter to the patient, whether follow-up is sufficient and whether confounding or selection bias could explain the result.
Apply
Translate evidence into shared care with patient preference and local resources. A well-designed study can still be a poor fit for a particular population or intervention.
Key distinctions
Randomisation reduces some confounding but does not guarantee a study is free from bias; loss to follow-up, selective reporting and inappropriate comparators still matter. Observational evidence may be more practical for rare harms or long-term outcomes. Consider absolute effects and confidence intervals, not only relative percentages. Apply findings in light of patient values and feasibility.
Think through a case
A trial reports a statistically significant 1% improvement with wide uncertainty around patient-relevant outcomes. The practical benefit may still be small.
The distinction to remember
Do not equate a small p-value with a large effect or assume a trial population matches your patient.
Test your recall
Does statistical significance prove clinical importance?
Show answer
No; the size and relevance of the effect also matter.