TSH — Hypothyroidism
Test Characteristics
| Metric | Value |
|---|---|
| False-negative rate | 2% (sensitivity 98%) |
| False-positive rate | 8% (specificity 92%) |
| Bayes factor (positive test) | 10× |
| Bayes factor (negative test) | 1/50× |
| Base rate | 5% of US adults (overt hypothyroidism) |
Interpreting Results
| Scenario | Prior | + Result | − Result |
|---|---|---|---|
| Unselected primary care patient | 5% | 10 × 5% = 50%5%× 1050% | 1/50 × 5% = 0.1%5%÷ 500.1% |
| Fatigue, weight gain, cold intolerance, family hx | 15% | 10 × 15% ≥ 100%15%× 10100% | 1/50 × 15% = 0.3%15%÷ 500.3% |
10 × 15% ≥ 100%15%× 10100%: exact posterior is 64%. TSH is exceptional: strong rule-in (10×), extraordinary rule-out (1/50×). A normal TSH essentially excludes primary hypothyroidism at any reasonable prior. Performance data from specialty settings; in unselected populations, false positives may be higher (acute illness, medications).
- + result: risk goes from vegetarian (5%) to coin flip (50%)
- − result: risk drops to appendicitis this year (0.1%) — essentially ruling it out
Sources:
- AAFP. Screening for thyroid disease.
- USPSTF. Screening for thyroid dysfunction.
- J Endocr Soc, 2022. PMC 9706417.