USMLE purpose: Master the “guaranteed points” setup for sensitivity, specificity, predictive values, study design, odds ratio, relative risk, cutoffs, prevention, power, and standard deviation.
This is the core Randy Neil “Biostats Basics” episode: draw the box, label the study design, recognize the formula, and do the arithmetic without overthinking.
| Concept | Rule | Stem clue |
|---|---|---|
| Sensitivity | TP / (TP + FN) | “Among people with disease, how many test positive?” |
| Specificity | TN / (TN + FP) | “Among people without disease, how many test negative?” |
| PPV | TP / (TP + FP) | Positive test → chance truly diseased |
| NPV | TN / (TN + FN) | Negative test → chance truly disease-free |
| Case-control | Odds ratio | Starts with disease vs no disease |
| Cohort | Relative risk | Starts with exposure and follows outcomes |
| Power | 1 − β | Probability of detecting a true difference |
| Normal distribution | 68–95–99.7 rule | Mean ± SD questions |
| Level | Timing | Examples |
|---|---|---|
| Primary | Before disease occurs | Vaccination, counseling, risk reduction |
| Secondary | Early/asymptomatic disease | Screening tests, BP check, colonoscopy |
| Tertiary | Established disease | Treatment, rehab, complication prevention |