Table 5.

The number of correctly and incorrectly classified patients for ASA; POSPOM; logistic regression model and DNN model with 87 features; logistic regression model and DNN model with reduced feature set and ASA; and logistic regression model and DNN model with reduced feature set and POSPOM at different thresholds. Results for best thresholds chosen by 1) highest % of observed mortality and 2) highest F1 score.

ASA					POSPOM
Threshold	# True Negative	# False Positive	# False Negative	# True Positive	Threshold	# True Negative	# False Positive	# False Negative	# True Positive
3	5,244	6,666	2	85	10	2,741	9,169	6	81
5	11,894	16	73	14	20	10,782	1,128	56	31
Logistic Regression with 87 features					DNN with 87 features
Threshold	# True Negative	# False Positive	# False Negative	# True Positive	Threshold	# True Negative	# False Positive	# False Negative	# True Positive
0.00015	37	11,873	1	86	0.05	2,414	9,496	1	86
0.1	11,788	122	59	28	0.3	11,734	176	52	35
Logistic Regression with Reduced Feature Set & ASA					DNN with Reduced Feature Set & ASA
Threshold	# True Negative	# False Positive	# False Negative	# True Positive	Threshold	# True Negative	# False Positive	# False Negative	# True Positive
0.0034	7,578	4,332	3	84	0.22	7,669	4,241	3	84
0.1	11,795	115	57	30	0.4	11,875	35	72	15
Logistic Regression with Reduced Feature Set & POSPOM					DNN with Reduced Feature Set & POSPOM
Threshold	# True Negative	# False Positive	# False Negative	# True Positive	Threshold	# True Negative	# False Positive	# False Negative	# True Positive
0.002	5,772	6,138	2	85	0.2	7,550	4,360	3	84
0.1	11,790	120	61	26	0.4	11,897	12	82	5

POSPOM: Preoperative Score to Predict Postoperative Mortality; DNN: Deep neural network.