Examqo › Examqo library › Family medicine › Evidence-based medicine, family and law · IMLE
Diagnostic test accuracy (sensitivity, specificity, predictive values)
Sensitivity and specificity belong to the test itself, but the predictive value depends on how common the disease is in the tested population — so the same "good" test gives many false positives when the disease is rare.
Why it matters in the exam
Every family medicine paper has calculations from a 2×2 table (sensitivity, specificity, positive and negative predictive value) and understanding questions: what happens to the predictive value when prevalence rises, which test is suited to ruling out and which to ruling in, and what a likelihood ratio means.
Newer questions describe a company that developed a "highly sensitive and specific" test or an artificial-intelligence algorithm, and ask why most positives in the general population are still not ill.
This understanding is also an everyday tool: before ordering a test, estimate the pre-test probability and ask whether a positive or negative result will really change the decision. A test whose result changes nothing only adds false positives and unnecessary work-ups.
An example worth remembering: in a thousand people with a prevalence of 1%, a test with 90% sensitivity and specificity finds 9 true cases and 99 false positives. So out of 108 positive results only 9 are ill — and that is the whole explanation of a low positive predictive value in screening. Other questions show an ROC curve and ask which cut-off suits the purpose of the test — screening or confirmation.
Key points
- Sensitivity = TP over all the ill; specificity = TN over all the healthy.
- Positive predictive value rises with prevalence; negative predictive value falls with it.
- A sensitive test — to rule out; a specific test — to rule in.
- A likelihood ratio above 10 or below 0.1 changes the decision considerably.
- When screening for a rare disease, most positives are false positives.
Summary table
| Measure | Formula (2×2 table) | What it tells | Depends on prevalence? |
|---|---|---|---|
| Sensitivity | TP / (TP + FN) | How many of the ill the test detects | No |
| Specificity | TN / (TN + FP) | How many of the healthy the test calls healthy | No |
| Positive predictive value | TP / (TP + FP) | The chance that a positive is truly ill | Yes — rises with prevalence |
| Negative predictive value | TN / (TN + FN) | The chance that a negative is truly healthy | Yes — falls as prevalence rises |
| Positive likelihood ratio | Sensitivity / (1 − specificity) | How much a positive result raises the probability | No; above 10 — a large change |
| Negative likelihood ratio | (1 − sensitivity) / specificity | How much a negative result lowers the probability | No; below 0.1 — a large change |
In the full version for subscribers
- A simple algorithm
- Memory aid
- Exam traps
- 5 review cards
- Enough for IMLE
The full version is open to subscribers of: Family medicine, Medical licensing exam (IMLE), Internal medicine, Public health.
Subscriber? Read the full topic Plans and prices 25 free questions
Sources
- Straus SE, Glasziou P, Richardson WS, Haynes RB. Evidence-Based Medicine: How to Practice and Teach EBM, 5th edition (2018) — Diagnosis and screening
- Gordis Epidemiology (Celentano DD, Szklo M), 6th edition (2019) — Assessing the validity and reliability of diagnostic and screening tests
- Centre for Evidence-Based Medicine, University of Oxford, current — Likelihood ratios and pre-test probability
Sources are given for further reading; the text is written afresh and not copied from them.
Reviewed by the Examqo AI Professor · 2026-11-06. The text may contain errors; when in doubt, check the source.