BMC Medical Research Methodology

Table 2 Summary of univariable and multivariable regression analyses for sensitivity/specificity and likelihood ratios of the Papanicolaou smear test (only p-values less than 0.20 are presented)

From: A methodological framework to distinguish spectrum effects from spectrum biases and to assess diagnostic and screening test accuracy for patient populations: Application to the Papanicolaou cervical cancer smear test

	Univariable analyses								Multivariable analyses
	Clinical reading				Optimized interpretation				Clinical reading				Optimized interpretation
	Sensitivity Specificity†		Likelihood ratios‡		Sensitivity Specificity†		Likelihood ratios‡		Sensitivity Specificity†		Likelihood ratios‡		Sensitivity Specificity†		Likelihood ratios‡
Variables	Main effect	Interaction	Main effect	Interaction	Main effect	Interaction	Main effect	Interaction	Main effect	Interaction	Main effect	Interaction	Main effect	Interaction	Main effect	Interaction
HPV test	<0.001	0.03	<0.001	0.033	<0.001	0.18	<0.001	0.18	<0.001	0.09	<0.001	0.011	<0.001	0.15	<0.001	0.17
Study setting	<0.001	0.10	0.001	0.11	<0.001	-	0.023	-	<0.001	0.14	0.015	-	<0.001	-
Age	0.001	0.06	-	0.07	0.04	-	0.12	-	0.026	-			0.045	-	0.047	-
Current smoking	0.03	-	-	-	-	-	-	-
European origin	<0.001	-	-	-	0.17	-	-	-
Educational level	<0.001	0.035	-	0.043	0.014	-	-	-
Postmenopausal	-	-	-	-	-	0.17	-	-
Contraceptive status	-	-	-	-	-	-	-	-

HPV: human papillomavirus.
† For the sensitivity/specificity model: interaction between the variable and the reference standard (see Additional file 1)
‡ For the likelihood ratios model: interaction between the variable and the diagnostic test (see Additional file 1).

Back to article page

ISSN: 1471-2288

Contact us

General enquiries: journalsubmissions@springernature.com