Back to dashboard
Randy Neil Biostats 22: USMLE Step 1, 2CK - Guaranteed Question on Step Exam
Episode Notes
USMLE purpose: Prepare for “guaranteed” board-style biostats questions by recognizing the high-frequency patterns quickly.
How to use this page
- Memorize the core pattern table.
- Drill the formulas until you can write them without hesitation.
- Use the practice question toggles to confirm recall.
- Treat this as a rapid review page before a question block.
Episode metadata
| Field | Details |
| Episode | Randy Neil Biostats 22 |
| Topic | Guaranteed exam question |
| Runtime | 11 min |
| Published | 2022-07-13 |
| Source | Open YouTube video |
One-liner
Most “guaranteed” biostats items are not conceptually hard; they test whether you can recognize the metric and avoid the wrong denominator.
High-frequency pattern table
| Question wording | Metric | Formula / move |
| Positive test, actually disease? | PPV | TP / (TP + FP) |
| Negative test, actually no disease? | NPV | TN / (TN + FN) |
| Disease present, test positive? | Sensitivity | TP / (TP + FN) |
| Disease absent, test negative? | Specificity | TN / (TN + FP) |
| Exposure followed to outcome | Relative risk | Risk exposed / risk unexposed |
| Disease vs no disease, look back | Odds ratio | AD / BC |
Common traps
⚠️ Trap: Using sensitivity when the question asks about a positive test result.
Fix: If the test result is already known, think predictive value.
⚠️ Trap: Putting the total sample size in every denominator.
Fix: Use the denominator that matches the subgroup being asked about.
Anki-style rapid recall
What is the formula for specificity?
TN / (TN + FP).
What study design gives odds ratio?
Case-control.
What study design gives relative risk?
Cohort or randomized trial.
Practice Question
A test is negative. The question asks the probability the patient truly does not have disease. Which metric is being tested?
- A) NPV
- B) PPV
- C) Sensitivity
- D) Odds ratio
Reveal answer & explanation
Answer: A) NPV
Negative test → truly disease-free is negative predictive value.