Table 4 Positive predictive value per 100,000 visits (95% CI) of prediction models from split-sample and entire sample estimation approaches in the development dataset and prospective validation dataset
Split-sample prediction model | Entire-sample prediction model | ||||
|---|---|---|---|---|---|
Risk percentile cutpoint | Testing set, Development | Prospective validation | 5-fold cross-validation, Development | Bootstrap optimism correction, Development | Prospective validation |
≥ 99% | 27.5 (16.7, 39.3) | 27.6 (14.3, 46.3) | 30.0 (21.9, 39.3) | 41.7 (21.0, 58.9) | 23.7 (11.4, 40.7) |
≥ 95% | 19.3 (13.3, 26.5) | 12.6 (8.6, 17.5) | 16.0 (12.4, 19.7) | 22.6 (19.8, 25.2) | 13.9 (9.4, 19.9) |
≥ 90% | 12.2 (9.0, 15.9) | 7.9 (5.8, 10.5) | 12.0 (9.9, 14.4) | 15.2 (13.8, 16.3) | 9.1 (6.7, 12.0) |
≥ 75% | 7.1 (5.8, 8.7) | 4.8 (3.9, 5.9) | 7.2 (6.3, 8.2) | 8.1 (7.8, 8.4) | 4.9 (3.9, 6.0) |