Ferritin (<30 ng/mL) — Iron Deficiency Anemia
Test Characteristics
| Metric | Value |
|---|---|
| False-negative rate | 8% (sensitivity 92%) |
| False-positive rate | 2% (specificity 98%) |
| Bayes factor (positive test) | 50× |
| Bayes factor (negative test) | 1/10× |
| Base rate | 5% of adult women; 2% of adult men |
Interpreting Results
| Scenario | Prior | + Result | − Result |
|---|---|---|---|
| Woman with fatigue | 5% | 50 × 5% ≥ 100%5%× 50100% | 1/10 × 5% = 0.5%5%÷ 100.5% |
| Menstruating woman with heavy periods | 10% | 50 × 10% ≥ 100%10%× 50100% | 1/10 × 10% = 1%10%÷ 101% |
| Man, no obvious risk factors | 2% | 50 × 2% ≥ 100%2%× 50100% | 1/10 × 2% = 0.2%2%÷ 100.2% |
All positive-test entries saturate. Exact posteriors: woman with fatigue = 72%, heavy periods = 85%, man = 51%. One of the strongest rule-in tests in primary care (50×) and also a good rule-out (1/10×). Critical caveat: ferritin is an acute-phase reactant — in patients with concurrent inflammation, a normal ferritin does not rule out iron deficiency. AGA recommends a higher threshold of <45 ng/mL in such patients.
- + result: risk goes from vegetarian (5%) to 72%
- − result: risk drops to identical twin (0.5%) — one of the best tests in primary care
Sources:
- JAMA Network Open, 2024. Ferritin cutoffs in primary care.
- AAFP. Iron deficiency anemia: evaluation and management.
- AGA guidelines, 2021. AAFP summary.