Table 6.
Identification of High Performance Hospitals by “No Date Stamp” Models
| Condition | n | Sensitivity | FNER | PPV | FPER |
|---|---|---|---|---|---|
| Coronary artery bypass | 123 | 0.54 | 0.46 | 0.76 | 0.24 |
| Coronary angioplasty | 162 | 0.82 | 0.18 | 0.84 | 0.16 |
| Carotid endarterectomy | 311 | 0.99 | 0.01 | 0.99 | 0.01 |
| Abdominal aortic aneurysm repair | 301 | 0.97 | 0.03 | 0.87 | 0.13 |
| Hip replacement | 360 | 0.96 | 0.04 | 0.97 | 0.03 |
| Acute myocardial infarction | 378 | 0.85 | 0.15 | 0.77 | 0.23 |
| Stroke | 389 | 0.95 | 0.05 | 1.00 | 0 |
n, number of hospitals; FNER, false-negative error rate (proportion of high-performance hospitals “missed” by the “no date stamp” models); PPV, positive predictive value; FPER, false-positive error rate (proportion of hospitals misclassified as high-quality hospitals by “no date stamp” model).