BMC Medical Research Methodology

Table 1 Rogan-Gladen adjusted prevalence estimate of the testing dataset for each cutoff method, test, and target specificity

From: Estimating cutoff values for diagnostic tests to achieve target specificity using extreme value theory

	Spike			RBD
	Prevalence	Sensitivity	Specificity	Prevalence	Sensitivity	Specificity
Target specificity=0.995
Empirical	0.64	0.28	0.99	0.26	0.87	0.99
Normal	0.30	0.77	0.97	0.28	0.96	0.99
Log Normal	0.35	0.63	0.98	0.28	0.98	0.98
MAD	0.32	0.99	0.91	0.39	1.00	0.85
Log MAD	0.29	0.93	0.95	0.32	0.99	0.93
Pareto 0.9	0.63	0.27	0.99	0.27	0.91	0.99
Pareto 0.95	0.61	0.31	0.99	0.26	0.87	0.99
Target specificity=0.95
Empirical	0.29	0.93	0.95	0.30	0.99	0.95
Normal	0.30	0.98	0.94	0.28	0.98	0.98
Log Normal	0.31	0.99	0.93	0.31	0.99	0.94
MAD	0.37	0.99	0.85	0.42	1.00	0.80
Log MAD	0.34	0.99	0.90	0.39	1.00	0.84
Pareto 0.9	0.29	0.96	0.94	0.28	0.99	0.97

Empirical and Pareto 0.95 cutoffs are equivalent when the target specificity is 0.95
Abbreviations: MAD mean absolute deviation, RBD receptor-binding domain

Back to article page

ISSN: 1471-2288

Contact us

General enquiries: journalsubmissions@springernature.com