Table 4. Risk of Class I Device Recall According to Predicate Recall Status.
| Characteristics | Cases (predicates with Class I recalls) | Matched controlsa | P value |
|---|---|---|---|
| Device population | |||
| Sample size, total No. | 64 | 64 | |
| Median months between clearance of cases and matched controls (IQR) | 0.7 (−2.4 to 2.5) | ||
| Descendant devices | |||
| Sample size, total No. | 177 | 121 | |
| Predicates per device, median (IQR) | 2 (1.5 to 3.5) | 1 (1 to 2) | |
| Descendant devices with recalls of any class, No. (%) | 120 (67.8) | 25 (20.7) | |
| Descendant devices with Class I recalls, No. (%) | 103 (58.2) | 11 (9.1) | |
| Statistical analyses, risk ratio (95% CI) | |||
| Relative risk of device recall of any class if authorized using a predicate subject to a Class I recall | 3.28 (2.28 to 4.72) | <.001 | |
| Relative risk of Class I device recall if authorized using a predicate subject to a Class I recall | 6.40 (3.59 to 11.40) | <.001 |
a
Matching was based on authorization date, manufacturer, and recall status for any class of recall.