Table 2 Distribution of labeled breaths by event type and dataset
From: Real-time detection of respiratory circuit events in mechanical ventilation using deep learning
Respiratory event type | Model development (n = 18,136) | Internal validation (n = 8632) | External validation (n = 30,528) |
|---|---|---|---|
Normal | 5320 (29.3%) | 3120 (36.1%) | 7976 (26.1%) |
Fluid-accumulation-like patterns | 9344 (51.5%) | 3952 (45.8%) | 10,984 (36.0%) |
Leakage-like patterns | 3472 (19.1%) | 1560 (18.1%) | 11,568 (37.9%) |
