Articles explaining common statistics, risk figures, test-accuracy claims and outcome measures, and how to read them in health studies. This page lists all 9 Everyday Prevention Evidence guides on the topic, each with its opening lines so you can pick the one that matches your question.
Everyday Prevention Evidence is an independent publication. These guides are for education and do not replace care from a qualified professional.
Guides in this section
Composite Endpoints: Which Events Actually Drove the Health Claim?
When a headline says a treatment "cut major health events by 30%," that number often isn't one event. It's a composite endpoint — several different events counted together as a single result.
Number Needed to Harm: Why the Time Period and Baseline Risk Matter
Imagine a friend tells you a new medication has a "number needed to harm" (NNH) of 50 — meaning one extra person out of every 50 treated experiences a specific side effect. That sounds precise. But precise about what? Fifty people treated for how long? Compared to what baseline risk?
Surrogate Outcome vs. Patient Outcome: What to Ask
A surrogate outcome is a stand-in measurement, such as a lab value or scan result, used because it is faster to track than a patient outcome like survival or quality of life. A surrogate outcome improving does not automatically mean the patient outcome improved too.
What a P-Value Can—and Cannot—Tell You About a Health Study
A p-value tells you how surprising a study's results would be if there were really no effect at all. It is not the probability the treatment works, and it does not measure how big or important an effect is. Reading it correctly means asking three follow-up questions: how big, how certain, and how repeatable.
How to Read a Health Test-Accuracy Claim Without Treating It as a Diagnosis
A health test's accuracy claim describes how the test performed in a research study — not what your own result means. Reading it correctly means separating the study-level statistics from your personal situation, and asking what a specific result would actually change before assuming what it implies.
Statistical Significance and Practical Importance: Two Different Questions
"Statistically significant" doesn't mean a treatment works well or a result matters. It means a study's finding probably wasn't random chance — nothing more. Whether that finding actually matters for real life is a separate question, answered by different numbers entirely.
Why the Outcome Measured Matters in a Health Study
An outcome measure is the specific thing a health study tracks to judge whether something worked — a lab value, a symptom score, a hospital visit rate, or a length of time.
What a Confidence Interval Can Tell You About a Health Study
When a health study reports a "95% confidence interval," it isn't giving you one exact number — it's giving you a range. A confidence interval shows how much a study's result might shift if the same research were repeated.
Relative Risk and Absolute Risk: What a Health Claim May Leave Out
Relative risk and absolute risk are two different ways of measuring the same health statistic. Relative risk compares two groups to each other. Absolute risk tells you how many people out of a group are actually affected.