This Article Full Text (PDF) References Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to Saved Citations Download to citation manager Request Permissions Request Reprints Add to My Marked Citations Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (20) Citing Articles via Google Scholar Citing Articles via Scopus Google Scholar Articles by Goetghebeur, E. Articles by Van der Stuyft, P. Search for Related Content PubMed PubMed Citation Articles by Goetghebeur, E. Articles by Van der Stuyft, P. Social Bookmarking                         What's this?

Diagnostic test analyses in search of their gold standard: latent class analyses with random effects

Ghent University, Ghent, Belgium, els.goetghebeur{at}rug.ac.be

Jan Liinev

Ghent University, Ghent, Belgium

Marleen Boelaert

Institute of Tropical Medicine, Antwerp, Belgium

Patrick Van der Stuyft

Institute of Tropical Medicine, Antwerp, Belgium

We review methods for analysing the performance of several diagnostic tests when patients must be classified as having a disease or not, when no gold standard is available. For latent class analysis (LCA) to provide consistent estimates of sensitivity, specificity and prevalence, traditionally `independent errors conditional on disease status' have been assumed. Recent approaches derive estimators under more flexible assumptions. However, all likelihood-based approaches suffer from the sparseness of tables generated by this type of data; an issue which is often ignored. In light of this, we examine the potential and limitations of LCAs of diagnostic tests. We are guided by a data set of visceral leishmaniasis tests. In the example, LCA estimates suggest that the traditional reference test, parasitology, has poor sensitivity and underestimates prevalence. From a technical standpoint, including more test results in one analysis yields increasing degrees of sparseness in the table which are seen to lead to discordant values of asymptotically equivalent test statistics and eventually lack of convergence of the LCA algorithm. We suggest some strategies to cope with this.

Statistical Methods in Medical Research, Vol. 9, No. 3, 231-248 (2000)
DOI: 10.1177/096228020000900304


CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter    What's this?

This article has been cited by other articles:

				R. A Hansen, M. M Kim, L. Song, W. Tu, J. Wu, and M. D Murray Comparison of Methods to Assess Medication Adherence and Classify Nonadherence Ann. Pharmacother., March 1, 2009; 43(3): 413 - 422. [Abstract] [Full Text] [PDF]

				A. L. Baughman, K. M. Bisgard, M. M. Cortese, W. W. Thompson, G. N. Sanden, and P. M. Strebel Utility of Composite Reference Standards and Latent Class Analysis in Evaluating the Clinical Accuracy of Diagnostic Tests for Pertussis Clin. Vaccine Immunol., January 1, 2008; 15(1): 106 - 114. [Abstract] [Full Text] [PDF]

				J. Roy Latent class models and their application to missing-data patterns in longitudinal studies Statistical Methods in Medical Research, October 1, 2007; 16(5): 441 - 456. [Abstract] [PDF]

				P Moayyedi, J Duffy, and B Delaney New approaches to enhance the accuracy of the diagnosis of reflux disease Gut, May 1, 2004; 53(suppl_4): iv55 - iv57. [Abstract] [Full Text] [PDF]

				M. BOELAERT, S. RIJAL, S. REGMI, R. SINGH, B. KARKI, D. JACQUET, F. CHAPPUIS, L. CAMPINO, P. DESJEUX, D. LE RAY, et al. A COMPARATIVE STUDY OF THE EFFECTIVENESS OF DIAGNOSTIC TESTS FOR VISCERAL LEISHMANIASIS Am J Trop Med Hyg, January 1, 2004; 70(1): 72 - 77. [Abstract] [Full Text] [PDF]

				P. Solomon and J. Hutton Editorial Statistical Methods in Medical Research, August 1, 2001; 10(4): 245 - 250. [PDF]