The statistics in a medical paper decide what its numbers are entitled to say. A p-value indicates how compatible the data are with no effect — not the probability that a hypothesis is true, and nothing about how large an effect is. Confidence intervals carry both the estimate and its precision. A relative risk reduction and an absolute one describe the same result and sound entirely different.
We report these because they are where a health story most often goes wrong between the paper and the headline. The aim is not statistical instruction but a reader able to ask what the absolute numbers were.
A practical guide to reading medical research: study design, sample size, p-values, confidence intervals, absolute vs relative risk, bias, and conflicts of interest.
How randomized controlled trials work: randomization, blinding, control groups, intention-to-treat analysis, and why RCTs are the standard for testing treatments.
Medical diagnostics explained: laboratory tests, molecular and point-of-care diagnostics, imaging, sensitivity and specificity, screening, and companion diagnostics.