Skip to main content
Fig. 2 | BMC Research Notes

Fig. 2

From: The clinical utility curve: a proposal to improve the translation of information provided by prediction models to clinicians

Fig. 2

ROC curves (a) and CU curves (b) for the prediction of patients with ≥1 coronary arteries with >50 % stenosis. Solid line represents men (n = 206) and dashed line represents women (n = 97). NNCOC, number needed to capture one patient with ≥1 coronary arteries with >50 % stenosis. C statistics were 0.846 (95 % CI 0.792–0.900) for men and 0.892 (0.812–0.973) for women, P = 0.348. The model achieved sensitivities of 80, 85, 90, 95 and 100 % at absolute risks of 46, 38, 32, 15 and 6 %, respectively, for men, and 36, 31, 22, 9 and 3 %, respectively, for women. The model identified all 114 men with ≥1 coronary arteries with >50 % stenosis in 199 men who underwent coronary angiography, with an NNCOC of 1.7, whereas when applied to women, the model identified all 25 women with ≥1 coronary arteries with >50 % stenosis in 80 women who underwent coronary angiography, with an NNCOC of 3.2

Back to article page