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 8 Summary estimates for sensitivity and specificity of mortality 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	2	0.90	0.64	0.98	0.446
	Public dataset	Specs	2	0.95	0.70	0.99	0.799
	Private dataset	Sens	31	0.85	0.80	0.88	[Reference]
	Private dataset	Specs	31	0.94	0.90	0.96	[Reference]
	Combined	Sens	4	0.82	0.66	0.91	0.591
	Combined	Specs	4	0.85	0.74	0.92	0.189
	Structured	Sens	29	0.86	0.80	0.90	[Reference]
	Structured	Specs	29	0.95	0.91	0.97	[Reference]
Sample	Adults	Sens	24	0.85	0.79	0.89	0.853
	Adults	Specs	24	0.94	0.89	0.97	0.759
	Youths	Sens	2	0.99	0.00	1.00	0.663
	Youths	Specs	2	0.89	0.83	0.92	0.458
	Elders	Sens	2	0.88	0.77	0.94	0.689
	Elders	Specs	2	0.90	0.77	0.96	0.540
	Mixed	Sens	5	0.84	0.67	0.93	[Reference]
	Mixed	Specs	5	0.96	0.82	0.99	[Reference]
Artificial intelligence technique	Machine learning	Sens	25	0.86	0.80	0.90	0.709
	Machine learning	Specs	25	0.95	0.90	0.97	0.442
	Deep learning	Sens	8	0.83	0.74	0.90	[Reference]
	Deep learning	Specs	8	0.91	0.84	0.96	[Reference]
	Random forest	Sens	4	0.91	0.56	0.99	0.516
	Random forest	Specs	4	1.00	0.04	1.00	0.476
	eXtreme Gradient boosting	Sens	3	0.73	0.61	0.82	0.562
	eXtreme Gradient boosting	Specs	3	0.95	0.77	0.99	0.839
	LightGBM	Sens	3	0.91	0.78	0.96	0.205
	LightGBM	Specs	3	0.92	0.75	0.98	0.748
	Logistic regression	Sens	2	0.80	0.73	0.86	0.938
	Logistic regression	Specs	2	0.95	0.68	0.99	0.866
	Deep neural network	Sens	5	0.86	0.76	0.92	0.418
	Deep neural network	Specs	5	0.89	0.83	0.93	0.436
	Convolutional neural network	Sens	3	0.79	0.63	0.89	[Reference]
	Convolutional neural network	Specs	3	0.94	0.74	0.99	[Reference]
	Ensemble	Sens	21	0.88	0.82	0.92	0.095
	Ensemble	Specs	21	0.95	0.89	0.98	0.620
	No ensemble	Sens	12	0.79	0.71	0.86	[Reference]
	No ensemble	Specs	12	0.93	0.87	0.96	[Reference]
	Cross validation	Sens	21	0.85	0.78	0.90	0.926
	Cross validation	Specs	21	0.96	0.92	0.98	0.032
	No cross validation	Sens	12	0.85	0.78	0.90	[Reference]
	No cross validation	Specs	12	0.87	0.81	0.91	[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

You are viewing the site in preview mode

BMC Medical Informatics and Decision Making

Contact us