Back to dashboard
Randy Neil Biostats 20: USMLE Step 1, 2CK - Biostats Viewer Requested Video
Episode Notes
USMLE purpose: Work backward from sensitivity, specificity, and prevalence to build a 2×2 diagnostic table.
How to use this page
- If no sample size is given, choose an easy number like 1000.
- Use prevalence to split disease+ vs disease−.
- Use sensitivity to fill the disease+ column.
- Use specificity to fill the disease− column.
- Calculate PPV or NPV from the completed table.
Episode metadata
| Field | Details |
| Episode | Randy Neil Biostats 20 |
| Topic | Viewer requested biostats review |
| Runtime | 13 min |
| Published | 2022-03-05 |
| Source | Open YouTube video |
One-liner
When prevalence is given but sample size is missing, choose N = 1000 , build the diagnostic table, then calculate PPV.
Worked PPV setup
Example: sensitivity 95%, specificity 80%, prevalence 20%, N = 1000.
| Disease + | Disease - | |
| Test + | 190 | 160 |
| Test - | 10 | 640 |
| Total | 200 | 800 |
PPV = 190 / (190 + 160) = 190 / 350 ≈ 54%
Board Exam Buzzwords
| Given | Use it to fill | Formula |
| Prevalence | Disease+ column total | Prevalence × N |
| Sensitivity | True positives among disease+ | TP / disease+ |
| Specificity | True negatives among disease− | TN / disease− |
| Positive test actually infected? | PPV | TP / TP+FP |
Anki-style rapid recall
If no N is given, what number can you choose?
Choose 1000 to avoid awkward decimals.
How does prevalence help build the table?
It gives the disease-positive column total.
What does specificity fill?
True negatives among patients without disease.
Practice Question
A test has sensitivity 95%, specificity 80%, and prevalence 20%. If N = 1000, what is the PPV?
- A) 23%
- B) 54%
- C) 80%
- D) 95%
Reveal answer & explanation
Answer: B) 54%
Disease+ = 200, disease− = 800. TP = 190, FP = 160. PPV = 190 / 350 ≈ 54%.