Sensitivity and specificity are fundamental measures of test performance. They help evaluate how well a test identifies disease or rules it out, especially when a gold standard test is unavailable or impractical.
Clinical Context Example:
Calling a cardiology fellow for a cardiac catheterization (gold standard for myocardial ischemia) without first performing an EKG (screening test) would be inefficient and unnecessary.
Sensitivity (True Positive Rate)
The probability that a test correctly identifies patients with the disease.
Sensitivity=
Key Points:
- Measures only the distribution of diseased individuals.
- Uses data from the “sick” column of a 2 × 2 table.
- False Negative Rate = 1 − Sensitivity.
Clinical Example:
- Temporal arteritis (TA) always shows elevated ESR in patients >50 years.
- If ESR is normal, TA is effectively ruled out.
- Mnemonic: SNNOUT → Sensitive test Negative rules OUT disease
Illustration:
| Sick (Disease Present) | Test Positive | Test Negative |
|---|---|---|
| 200 patients | 160 (TP) | 40 (FN) |
Specificity (True Negative Rate)
The probability that a test correctly identifies patients without disease.
Key Points:
- Measures only the distribution of disease-free individuals.
- Uses data from the “not sick” column of a 2 × 2 table.
- False Positive Rate = 1 − Specificity.
Clinical Example:
- CT angiogram for pulmonary embolism (PE) has a specificity ≈ of 97%.
- A positive CT is highly likely to represent true PE.
- Mnemonic: SPIN → Specific test Positive rules IN disease
Illustration:
| Not Sick (Disease Absent) | Test Positive | Test Negative |
|---|---|---|
| 100 patients | 3 (FP) | 97 (TN) |
Trade-Off Between Sensitivity and Specificity
- Tests are rarely perfect; increasing sensitivity often reduces specificity, and vice versa.
- Graphical representation: Screening dimension curves, ROC curves.
- Goal: Balance to minimize false positives and false negatives while maximizing true detection.
Quick Reference Table
| Measure | Formula | Clinical Use | Mnemonic |
|---|---|---|---|
| Sensitivity | TP / (TP + FN) | Detect disease, rule out disease if negative | SNNOUT |
| Specificity | TN / (TN + FP) | Confirm disease, rule in disease if positive | SPIN |
| False Negative Rate | 1 − Sensitivity | Risk of missing disease | — |
| False Positive Rate | 1 − Specificity | Risk of labeling healthy as sick | — |
Learning Objective
After studying this topic, medical students should be able to calculate and interpret sensitivity and specificity, understand their clinical applications, recognize the trade-offs between these measures, and apply them to diagnostic and screening tests in real-world scenarios.
