Risk of Acute Respiratory Distress Syndrome in Community-Acquired Pneumonia Patients: Use of an Artificial Neural Network Model

Table 4

Comparison of the ANNs model and LR model in predicting ARDS secondary to CAP in the test dataset.

Variable	ANNs model	LR model	Difference between models (95% CI)	value
Sensitivity	85.0%	89.2%	4.2% (1.2–6.9%)	0.093
Specificity	87.3%	88.1%	0.8% (0.4–1.4%)	0.405
Accuracy	86.5%	88.5%	2.0% (1.1–3.4%)	0.136
AUC	0.943 ± 0.025	0.942 ± 0.019	0.001 (0.000–0.002)	0.180

Note. ANNs, artificial neural networks; LR, logistic regression; ARDS, acute respiratory distress syndrome; CAP, community-acquired pneumonia; AUC, the area under ROC (receiver operating characteristic) curve.