Table 4.
Results of the ROC analysis for each participant split showing the number of participants (N) who were classified as ‘DP’ or ‘non-DP’ based on the CFMT & CFMT-A, the AUC value and its 95% confidence interval (CI) and the optimal PI20 cut-off and associated sensitivity (% of DPs correctly identified) and specificity (% of non-DPs correctly identified).
| CFMT cut-off | N DPs | N non-DPs | AUC |
optimal PI20 cut-off |
|||
|---|---|---|---|---|---|---|---|
| value | CI | cut-off | sensitivity | specificity | |||
| 65% | 95 | 289 | 0.75 | 0.69–0.80 | 60.5 | 77% | 67% |
| 60% | 60 | 324 | 0.76 | 0.70–0.83 | 60.5 | 82% | 63% |
| 55% | 23 | 361 | 0.88 | 0.82–0.94 | 75.5 | 78% | 82% |