The cost-effectiveness of three thyroid function testing strategies for suspicion of hypothyroidism in a primary care-setting.

OBJECTIVE: To determine the sensitivity and specificity of thyroxine (T4) and the cost-effectiveness of three testing strategies in the diagnosis of hypothyroidism in a primary care setting.
DESIGN: 1) A retrospective chart review to determine sensitivity and specificity of T4 in diagnosing hypothyroidism; a cost-effectiveness analysis comparing ordering an initial T4 test alone, an initial thyroid-stimulating hormone (TSH) test alone, and T4 and TSH tests together in diagnosing hypothyroidism; a sensitivity analysis was performed on critical assumptions.
SETTING: Primary care adult practice of a health maintenance organization. PATIENTS: Eight hundred sixteen consecutive patients suspected of having hypothyroidism who had both T4 and TSH tests performed.
INTERVENTIONS: None.
RESULTS: The sensitivity of a T4 cut-off of 7 micrograms/dl (90.3 nmol/L) in diagnosing primary hypothyroidism was 93% (95% confidence interval = 85-100%) and the specificity was 68% (95% confidence interval = 65-71%). The cost-effectiveness ratios of using an initial T4 or TSH test were about the same across a wide range of test characteristics and disease prevalence estimates. As the ratio of T4 to TSH test charges declines from 0.6 to 0.2, the marginal cost of the TSH-first method increases from $3,500 to $18,000 for each additional hypothyroid patient identified. Ordering both tests together was very costly compared with the single test methods ($125,000 for each additional case diagnosed) and remained so under a wide range of assumptions.
CONCLUSIONS: When hypothyroidism is suspected, a TSH-first testing approach is generally preferable due to its greater sensitivity and, under most assumptions, only small increment in average or marginal cost per case compared with a T4-first method.