. 2018 Dec 21;1(8):e185097. doi: 10.1001/jamanetworkopen.2018.5097

Table 3. Model Discrimination for Multicenter Models Using Different Data and Analytic Methods.

Modeling Approach	AUC (95% CI)^a
Using highest and lowest of all laboratory values and vital signs, logistic regression (baseline)	0.831 (0.830-0.832)
Adding information from all observed laboratory values and vital signs^b	0.899 (0.896-0.902)
Adding NLP of clinical text^c	0.922 (0.916-0.924)

Abbreviations: AUC, area under the receiver operating characteristic curve; NLP, natural language processing.

^{^a}

Calculated using nested 10-fold cross-validation; 95% CIs were computed using bootstrapping.

^{^b}

Adds measures of distribution, variability, and trajectory of laboratory values and vital signs to models already using the highest and lowest values.

^{^c}

Adds NLP to models already using all observed values and measures of distribution, variability, and trajectory of laboratory values and vital signs.