To the Editor.
—In trying to replicate the results of the article on screening for mild thyroid failure by Mr Danese and colleagues,1 I found a mathematical error in one of the assumptions of the cost-effectiveness model. Consequently, the analysis overestimates the effect of levothyroxine sodium on symptom relief and underestimates the costs of screening.Danese and colleagues assumed there was "symptomatic improvement in 28% of patients with mild thyroid failure." This estimate comes from 2 randomized trials of treatment for mild thyroid failure.2,3 In one of these trials,2 treatment with levothyroxine sodium relieved symptoms in 8 of 17 patients vs 3 of 16 patients in the placebo group. The difference of 8 of 17 (0.47) less 3 of 16 (0.19), or 0.28, is the net benefit or effect size of treatment. Its reciprocal, approximately 4, is the number needed to treat to benefit 1 patient. However,
Helfand M. More on Screening for Mild Thyroid Failure. JAMA. 1997;277(6):458-459. doi:10.1001/jama.1997.03540300026021