07 · FEMETRIC REFERENCE
Reading Research
Study design, effect estimates, uncertainty and applicability without a single evidence score.Start with PICO
Identify the population, intervention, comparator and outcome. Then check randomization, blinding, duration, attrition, prespecified endpoints, multiplicity and funding.
Estimate before threshold
Read the effect size and confidence interval before the p-value. A wide interval can include materially different possibilities even when a threshold is crossed.
Common misreads
A preclinical mechanism does not prove a human outcome. A biomarker change does not automatically establish clinical benefit. A subgroup finding may be exploratory. Absence of evidence is not proof of no effect, but it is also not positive evidence.
Apply the result carefully
Ask whether the enrolled population, tested formulation, route, duration and comparator resemble the question you are asking. Keep sponsor updates separate from peer-reviewed publications.