Editorial II: Solid as a ROC Galley HelenF. British Journal of Anaesthesia Volume 93, Issue 5, Pages 623-626 (November 2004) DOI: 10.1093/bja/aeh247 Copyright © 2004 British Journal of Anaesthesia Terms and Conditions
Fig 1 Test result for two imaginary patient populations with SIRS. The cut-off value is the test result above which patients are classified as having SIRS. In (a) the cut-off value is set fairly low (high sensitivity but low specificity), resulting in a low number of false negatives, but a large number of false positives. If the cut off value is increased (b), the sensitivity is decreased and the specificity is increased, so the number of false positives decreases but the frequency of false negatives increases, and the number of true positives decreases. British Journal of Anaesthesia 2004 93, 623-626DOI: (10.1093/bja/aeh247) Copyright © 2004 British Journal of Anaesthesia Terms and Conditions
Fig 2 Examples of what very good, good and useless ROC curves look like. British Journal of Anaesthesia 2004 93, 623-626DOI: (10.1093/bja/aeh247) Copyright © 2004 British Journal of Anaesthesia Terms and Conditions