Table 5.

Response of electrocardiographic artificial intelligence and Cox proportional hazards regression models to follow-up electrocardiograms

Scenarios	ECG time	TP	FP	TN	FN	Specificity	Sensitivity	Negative predictive value	Positive predictive value
1	Baseline	116	764	1819	2	0.7042	0.9831	0.9990	0.1318
	Follow-up	116	764	1819	2	0.7042	0.9831	0.9990	0.1318
2	Baseline	108	515	2068	10	0.8006	0.9153	0.9952	0.1734
	Follow-up	116	528	2055	2	0.7956	0.9831	0.9990	0.1801
3	Baseline	93	261	2322	25	0.8990	0.7881	0.9893	0.2627
	Follow-up	111	258	2325	7	0.9001	0.9407	0.9970	0.3008
4	Baseline	77	127	2456	41	0.9508	0.6525	0.9836	0.3775
	Follow-up	95	123	2460	23	0.9524	0.8051	0.9907	0.4358

ECG, electrocardiogram; FN, false negative; FP, fasle positive; TN, true negative; TP, true positive.