A 52-year-old man from rural Maharashtra presents with a 3-week history of persistent cough and low-grade fever. Chest X-ray shows a cavitary lesion in the right upper lobe. Sputum smear microscopy is positive for acid-fast bacilli (AFB). A rapid molecular test (GeneXpert MTB/RIF) is ordered. The test has a reported sensitivity of 98% and specificity of 99% in the hospital laboratory. Given the high pretest probability of tuberculosis in this clinical scenario (estimated at 85%), what is the approximate positive predictive value (PPV) of a positive GeneXpert result in this patient?

Explanation

## Calculating PPV in High Pretest Probability **Key Point:** PPV depends critically on pretest probability (prevalence), not just test sensitivity and specificity. A test performs best when pretest probability is high. ### Formula and Calculation PPV is calculated using Bayes' theorem: $$PPV = \frac{Sensitivity \times Pretest\ Probability}{(Sensitivity \times Pretest\ Probability) + [(1 - Specificity) \times (1 - Pretest\ Probability)]}$$ Given: - Sensitivity = 98% = 0.98 - Specificity = 99% = 0.99 - Pretest probability = 85% = 0.85 Substituting: $$PPV = \frac{0.98 \times 0.85}{(0.98 \times 0.85) + [(1 - 0.99) \times (1 - 0.85)]}$$ $$PPV = \frac{0.833}{0.833 + (0.01 \times 0.15)}$$ $$PPV = \frac{0.833}{0.833 + 0.0015}$$ $$PPV = \frac{0.833}{0.8345} \approx 0.998 \text{ or } 99.8\%$$ Approximate PPV ≈ **99.4%** ### Why PPV Is So High Here | Component | Value | |-----------|-------| | Sensitivity | 98% | | Specificity | 99% | | Pretest Probability | 85% | | **Resulting PPV** | **~99.4%** | **High-Yield:** When pretest probability is very high (>80%), even a moderately sensitive and specific test yields a very high PPV. This is why clinical suspicion (cavitary lesion + AFB smear) matters enormously. ### Clinical Pearl In this patient: - The clinical presentation (cavitary lesion + AFB smear positivity) already suggests TB with ~85% probability. - A positive GeneXpert in this context is almost diagnostic (PPV ~99.4%). - Conversely, a negative GeneXpert would have a high NPV (~99.7%), making TB unlikely but not impossible (especially in smear-positive, culture-positive cases with technical failure). **Mnemonic: PINT** — **P**retest probability **I**ncreases **N**egative Predictive value and **T**est utility overall. [cite:Park 26e Ch 11]