The 'receiver operating characteristic' (ROC) curve is used to:

• A Calculate the 95% confidence interval of the odds ratio
• B Select the optimal cut-off for a continuous diagnostic test by plotting sensitivity against (1 − specificity) at various thresholds ✓
• C Estimate the number needed to treat from a clinical trial
• D Measure agreement between two continuous measurements (Bland-Altman alternative)

Explanation

A ROC curve plots true positive rate (sensitivity) on the y-axis against false positive rate (1 − specificity) on the x-axis at every possible cut-off threshold. The area under the curve (AUC) quantifies overall discriminative ability: AUC 0.5 = no discrimination, 0.7–0.8 = acceptable, 0.8–0.9 = excellent, > 0.9 = outstanding. The optimal cut-off is typically the point closest to the top-left corner (maximizing Youden's index = sensitivity + specificity − 1).

Reference: Park's Textbook of Preventive and Social Medicine, 27th ed.

High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP

Written and medically reviewed by the StethoPrep medical team.