ROC curve for ureter and bladder
ROC curves in correlation to the median pain scores were calculated for the Enzian locations A, B, C, FB, FU, FA and FI. We only present the ROC curves with an area under the curve ≥0.6, since in that case this test can provide a discrimination. This was only the case for the ureter and bladder involvement (table 4).
Figure 2.1 shows the ROC curve for the ureter, with an area under the curve for dysmenorrhea of 0.52 (SE 0.04, CI 0.45-0.60, p=0.04), dysuria 0.51 (SE 0.04, CI 0.44-0.59, p=0.68), CPP 0.52 (SE 0.04, CI 0.44-0.60, p=0.67), dyspareunia 0.49 (SE 0.04, CI 0.41-0.56, p=0.70) and dyschezia 0.6 (SE 0.04, CI 0.53-0.67, p< 0.01) (table 4). We chose the cut-off point for dyschezia of NRS 6.50, while a high sensitivity is important regarding ureter involvement (cave silent hydronephrosis). Using this as a cutoff point, our sensitivity (true positive rate) would be 63% and our 1 – specificity would be 45% (false positive rate).
For the bladder (figure 2.2), with an area under the curve for dysmenorrhea of 0.49 (SE 0.04, CI 0.41-0.56, p=0.73), dysuria 0.62 (SE 0.04, CI 0.54-0.71, p<0.01), CPP 0.39 (SE 0.04, CI 0.31-0.47, p<0.01), dyspareunia 0.44 (SE 0.04, CI 0.36-0.52, p=0.13) and dyschezia 0.57 (SE 0.04, CI 0.49-0.66, p< 0.07) (table 4). We chose the cut-off point for dysuria of NRS 0.5. Using this as a cutoff point, our sensitivity would be 54% and our 1 – specificity would be 36%.