Evidence and research methods
How the USPSTF Built Its Adult Depression Screening Recommendation
A plain-language walk through the USPSTF Grade B for adult depression screening, the evidence links behind it, and why follow-up systems count.
Health & Evidence Library
Evidence and research methods
Tools for reading studies, interpreting tests and effect estimates, recognizing uncertainty, and asking whether evidence applies.
207 guides · Page 4 of 4
Evidence and research methods
How Preregistration and Registered Reports Curb Bad Science
A reader's guide to preregistration and Registered Reports, and how the order of plan versus results helps you judge a study's claims.
Evidence and research methods
How GRADE Turns Evidence Into a Recommendation
How GRADE rates certainty of evidence and uses the Evidence-to-Decision framework to reach strong or conditional guideline recommendations.
Evidence and research methods
How Doctors Decide Which Test to Order for Chest Pain
Why stable chest pain leads to CT angiography, stress imaging, or no test at all, based on the estimated likelihood of coronary disease.
Evidence and research methods
How Depression Rating Scales Are Scored and Why the Cutoffs Are Debated
How depression rating scales like the Hamilton and MADRS are scored, what a clinically important difference means, and why their cutoffs are debated.
Evidence and research methods
How Decision Rules Rule Out Pulmonary Embolism Without a Scan
How Wells, PERC, age-adjusted D-dimer, and YEARS rule out pulmonary embolism safely, and the failure-rate evidence behind skipping the scan
Evidence and research methods
How Clinical Guidelines Get Updated, and How to Read the Change
Guidelines change when a panel decides the advice no longer fits the evidence. How to read an update by its date, its trigger, and its dissent.
Evidence and research methods
How Choosing Wisely Decides a Test Is Low Value
How specialty societies build Choosing Wisely lists, the three tests each item must pass, and why the evidence behind them varies enough to deserve appraisal.
Evidence and research methods
How Basket Trials Support Tumor-Agnostic Drug Approvals
Why basket trials group cancers by shared mutation, how NTRK-fusion drugs earned tumor-agnostic approval, and what single-arm evidence cannot settle
Evidence and research methods
How and Why a Trial Chooses Its Primary Endpoint
Why a trial commits to one primary endpoint, the trade-offs between hard outcomes and surrogates, and what the choice reveals to a reader.
Evidence and research methods
How Adaptive and Platform Trial Designs Work, and What They Trade Off
A plain-language guide to adaptive randomization, arm-dropping, and platform trials, and the efficiency they buy against the discipline they demand.
Evidence and research methods
How a Perioperative Cardiac Risk Score Is Derived and Validated
How the Revised Cardiac Risk Index was built and tested, and what discrimination, calibration, and prediction versus causation actually mean.
Evidence and research methods
High, Moderate, Low, Very Low: How GRADE Rates the Certainty of Evidence
A plain-language guide to how GRADE rates evidence certainty across five domains, and why a strong recommendation and strong evidence are not the same thing.
Evidence and research methods
External Control Arms: Reading a Trial That Never Randomized Its Comparator
A plain-language guide to single-arm and externally controlled trials, when regulators may accept them, and how comparability and timing can distort the result.
Evidence and research methods
Efficacy Versus Effectiveness, and Why a Careful Reader Checks Which One Is Being Claimed
Efficacy is whether a treatment works under ideal conditions. Effectiveness is whether it works in everyday care. Check which one a claim means.
Evidence and research methods
Does the Type of Talk Therapy Matter? Reading the Comparative Evidence
An evidence appraisal of the dodo bird debate and comparative meta-analyses on whether CBT outperforms other talk therapies
Evidence and research methods
Does Advance Care Planning Improve End-of-Life Care?
Randomized trials show advance care planning rarely makes end-of-life care match patients' wishes, a lesson in trusting countable proxies.
Evidence and research methods
Confounding and Causation: Why a Real Association Can Still Lie to You
Association is not causation. A confounder is a third factor tied to both cause and effect, and study design exists to rule it out.
Evidence and research methods
Composite Endpoints: When Studies Combine Outcomes
What a composite endpoint is, why trials use them, and how to read one without being misled by an impressive combined number.
Evidence and research methods
Basket, Umbrella, and Platform Trials: What Master Protocols Actually Do
How basket, umbrella, and platform trials differ, and how master protocols lead to tissue-agnostic FDA approvals.
Evidence and research methods
When the headline and the study disagree
A practical way to check a health news story against the research behind it, so you can tell solid reporting from overstatement.
Evidence and research methods
What a nationwide registry study can and cannot tell us
Why large national health registries are a research powerhouse, what makes their findings strong, and where their limits lie.
Evidence and research methods
Statins and cancer outcomes: a lesson in reading the evidence
A careful look at what observational studies on statins and cancer can and cannot tell us, and why this topic is a good lesson in reading evidence well.
Evidence and research methods
Correlation, causation, and confounding, with real examples
Why two things moving together does not mean one causes the other, what confounding is, and how to think clearly about observational findings.
Evidence and research methods
Cohort, case-control, trial: a field guide to study types
A friendly guide to the main kinds of medical studies, what each is good for, and why some sit higher than others when judging cause.
Evidence and research methods
Absolute versus relative risk: the difference that changes everything
Why a claim that a treatment cuts risk by 50 percent can mean a lot or almost nothing, and how to tell which, in plain terms.
Evidence and research methods
How to read a medical study without a statistics degree
How to read a medical study without a statistics degree. A calm, six-step checklist for absolute risk, confidence intervals, p-values, and confounding.