Skip to main content
. 2023 Jun 2;6:104. doi: 10.1038/s41746-023-00838-3

Table 2.

Performance breakdown of binary models: External validation.

Target Model Sensitivity (CI95) Specificity (CI95) AUROC (CI95)
Control Baseline 0.241 (0.139–0.372) 0.830 (0.742–0.898) 0.703 (0.622–0.785)
DeepBreath 0.770 (0.645–0.868) 0.840 (0.756–0.904) 0.887 (0.839–0.936)
Pneumonia Baseline 0.000 (0.000–0.079) 1.000 (0.968–1.000) 0.315 (0.226–0.405)
DeepBreath 0.469 (0.325–0.617) 0.839 (0.760–0.900) 0.739 (0.656–0.821)
Wheezing Disorder Baseline 0.419 (0.245–0.609) 0.693 (0.605–0.772) 0.614 (0.513–0.715)
DeepBreath 0.636 (0.451–0.796) 0.701 (0.616–0.777) 0.743 (0.651–0.836)
Bronchiolitis Baseline 0.542 (0.328–0.744) 0.933 (0.876–0.969) 0.896 (0.843–0.949)
DeepBreath 0.500 (0.291–0.709) 0.923 (0.867–0.961) 0.870 (0.796–0.943)