A meta-analysis of the diagnostic test accuracy of artificial intelligence predicting emergency department dispositions

Kuo, Kuang-Ming; Chang, Chao Sheng

doi:10.1186/s12911-025-03010-x

BMC Medical Informatics and Decision Making

Table 6 Summary estimates for sensitivity and specificity of admission studies

From: A meta-analysis of the diagnostic test accuracy of artificial intelligence predicting emergency department dispositions

Characteristic	Covariate	Metric	n	Estimate	95% C.I.		p value
Data	Public dataset	Sens	6	0.94	0.57	0.99	0.284
	Public dataset	Specs	6	0.90	0.70	0.97	0.581
	Private dataset	Sens	33	0.79	0.73	0.84	[Reference]
	Private dataset	Specs	33	0.86	0.79	0.91	[Reference]
	Unstructured	Sens	3	0.72	0.70	0.74	0.387
	Unstructured	Specs	3	0.80	0.74	0.86	0.521
	Combined	Sens	5	0.74	0.61	0.84	0.362
	Combined	Specs	5	0.82	0.68	0.91	0.510
	Structured	Sens	31	0.83	0.75	0.89	[Reference]
	Structured	Specs	31	0.88	0.81	0.93	[Reference]
Sample	Adult	Sens	21	0.81	0.72	0.87	0.308
	Adult	Specs	21	0.87	0.76	0.93	0.448
	Youth	Sens	6	0.82	0.69	0.90	0.218
	Youth	Specs	6	0.80	0.67	0.88	0.064
	Elder	Sens	3	0.76	0.74	0.79	0.541
	Elder	Specs	3	0.75	0.70	0.79	0.027
	Mixed	Sens	5	0.70	0.55	0.82	[Reference]
	Mixed	Specs	5	0.92	0.84	0.96	[Reference]
Artificial intelligence technique	Machine learning	Sens	28	0.80	0.72	0.86	0.535
	Machine learning	Specs	28	0.87	0.80	0.92	0.770
	Deep learning	Sens	11	0.85	0.71	0.93	[Reference]
	Deep learning	Specs	11	0.85	0.73	0.92	[Reference]
	Random forest	Sens	10	0.82	0.61	0.93	0.147
	Random forest	Specs	10	0.92	0.77	0.98	0.720
	eXtreme Gradient Boosting	Sens	8	0.77	0.69	0.84	0.022
	eXtreme Gradient Boosting	Specs	8	0.86	0.77	0.91	0.788
	Deep neural network	Sens	7	0.70	0.65	0.75	0.002
	Deep neural network	Specs	7	0.78	0.74	0.82	0.360
	Recurrent neural network	Sens	2	0.97	0.31	1.00	0.604
	Recurrent neural network	Specs	2	0.93	0.58	0.99	0.899
	Convolutional neural network	Sens	2	0.99	0.15	1.00	[Reference]
	Convolutional neural network	Specs	2	0.99	0.02	1.00	[Reference]
	Ensemble	Sens	26	0.78	0.71	0.84	0.301
	Ensemble	Specs	26	0.86	0.78	0.91	0.713
	No-ensemble	Sens	13	0.86	0.73	0.93	[Reference]
	No-ensemble	Specs	13	0.88	0.76	0.95	[Reference]
	Cross validation	Sens	26	0.84	0.76	0.90	0.170
	Cross validation	Specs	26	0.86	0.76	0.92	0.541
	No cross validation	Sens	13	0.74	0.65	0.82	[Reference]
	No cross validation	Specs	13	0.88	0.82	0.93	[Reference]

Note: C.I. = Confidence Interval, Sens = Sensitivity, and Specs = Specificity

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com