A probit latent class model with general correlation structures for evaluating accuracy of diagnostic tests

Huiping Xu, Bruce A. Craig

Research output: Contribution to journalArticle

23 Citations (Scopus)

Abstract

Traditional latent class modeling has been widely applied to assess the accuracy of dichotomous diagnostic tests. These models, however, assume that the tests are independent conditional on the true disease status, which is rarely valid in practice. Alternative models using probit analysis have been proposed to incorporate dependence among tests, but these models consider restricted correlation structures. In this article, we propose a probit latent class model that allows a general correlation structure. When combined with some helpful diagnostics, this model provides a more flexible framework from which to evaluate the correlation structure and model fit. Our model encompasses several other PLC models but uses a parameter-expanded Monte Carlo EM algorithm to obtain the maximum-likelihood estimates. The parameter-expanded EM algorithm was designed to accelerate the convergence rate of the EM algorithm by expanding the complete-data model to include a larger set of parameters and it ensures a simple solution in fitting the PLC model. We demonstrate our estimation and model selection methods using a simulation study and two published medical studies.

Original languageEnglish (US)
Pages (from-to)1145-1155
Number of pages11
JournalBiometrics
Volume65
Issue number4
DOIs
StatePublished - Dec 2009
Externally publishedYes

Fingerprint

Latent Class Model
Probit Model
Diagnostic Tests
Correlation Structure
Routine Diagnostic Tests
diagnostic techniques
Likelihood Functions
EM Algorithm
Model
Programmable logic controllers
Monte Carlo EM Algorithm
Latent Class
Maximum Likelihood Estimate
Model Selection
Large Set
Data Model
Accelerate
Diagnostics
Rate of Convergence
probit analysis

Keywords

  • Diagnostic tests
  • Latent class models
  • Monte Carlo EM algorithm
  • Parameter-expanded EM algorithm
  • Probit models
  • Sensitivity
  • Specificity

ASJC Scopus subject areas

  • Applied Mathematics
  • Statistics and Probability
  • Agricultural and Biological Sciences(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Immunology and Microbiology(all)
  • Medicine(all)

Cite this

A probit latent class model with general correlation structures for evaluating accuracy of diagnostic tests. / Xu, Huiping; Craig, Bruce A.

In: Biometrics, Vol. 65, No. 4, 12.2009, p. 1145-1155.

Research output: Contribution to journalArticle

@article{ef7db20ba2cc43cd8a3972becc5af722,
title = "A probit latent class model with general correlation structures for evaluating accuracy of diagnostic tests",
abstract = "Traditional latent class modeling has been widely applied to assess the accuracy of dichotomous diagnostic tests. These models, however, assume that the tests are independent conditional on the true disease status, which is rarely valid in practice. Alternative models using probit analysis have been proposed to incorporate dependence among tests, but these models consider restricted correlation structures. In this article, we propose a probit latent class model that allows a general correlation structure. When combined with some helpful diagnostics, this model provides a more flexible framework from which to evaluate the correlation structure and model fit. Our model encompasses several other PLC models but uses a parameter-expanded Monte Carlo EM algorithm to obtain the maximum-likelihood estimates. The parameter-expanded EM algorithm was designed to accelerate the convergence rate of the EM algorithm by expanding the complete-data model to include a larger set of parameters and it ensures a simple solution in fitting the PLC model. We demonstrate our estimation and model selection methods using a simulation study and two published medical studies.",
keywords = "Diagnostic tests, Latent class models, Monte Carlo EM algorithm, Parameter-expanded EM algorithm, Probit models, Sensitivity, Specificity",
author = "Huiping Xu and Craig, {Bruce A.}",
year = "2009",
month = "12",
doi = "10.1111/j.1541-0420.2008.01194.x",
language = "English (US)",
volume = "65",
pages = "1145--1155",
journal = "Biometrics",
issn = "0006-341X",
publisher = "Wiley-Blackwell",
number = "4",

}

TY - JOUR

T1 - A probit latent class model with general correlation structures for evaluating accuracy of diagnostic tests

AU - Xu, Huiping

AU - Craig, Bruce A.

PY - 2009/12

Y1 - 2009/12

N2 - Traditional latent class modeling has been widely applied to assess the accuracy of dichotomous diagnostic tests. These models, however, assume that the tests are independent conditional on the true disease status, which is rarely valid in practice. Alternative models using probit analysis have been proposed to incorporate dependence among tests, but these models consider restricted correlation structures. In this article, we propose a probit latent class model that allows a general correlation structure. When combined with some helpful diagnostics, this model provides a more flexible framework from which to evaluate the correlation structure and model fit. Our model encompasses several other PLC models but uses a parameter-expanded Monte Carlo EM algorithm to obtain the maximum-likelihood estimates. The parameter-expanded EM algorithm was designed to accelerate the convergence rate of the EM algorithm by expanding the complete-data model to include a larger set of parameters and it ensures a simple solution in fitting the PLC model. We demonstrate our estimation and model selection methods using a simulation study and two published medical studies.

AB - Traditional latent class modeling has been widely applied to assess the accuracy of dichotomous diagnostic tests. These models, however, assume that the tests are independent conditional on the true disease status, which is rarely valid in practice. Alternative models using probit analysis have been proposed to incorporate dependence among tests, but these models consider restricted correlation structures. In this article, we propose a probit latent class model that allows a general correlation structure. When combined with some helpful diagnostics, this model provides a more flexible framework from which to evaluate the correlation structure and model fit. Our model encompasses several other PLC models but uses a parameter-expanded Monte Carlo EM algorithm to obtain the maximum-likelihood estimates. The parameter-expanded EM algorithm was designed to accelerate the convergence rate of the EM algorithm by expanding the complete-data model to include a larger set of parameters and it ensures a simple solution in fitting the PLC model. We demonstrate our estimation and model selection methods using a simulation study and two published medical studies.

KW - Diagnostic tests

KW - Latent class models

KW - Monte Carlo EM algorithm

KW - Parameter-expanded EM algorithm

KW - Probit models

KW - Sensitivity

KW - Specificity

UR - http://www.scopus.com/inward/record.url?scp=70450227267&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70450227267&partnerID=8YFLogxK

U2 - 10.1111/j.1541-0420.2008.01194.x

DO - 10.1111/j.1541-0420.2008.01194.x

M3 - Article

C2 - 19210729

AN - SCOPUS:70450227267

VL - 65

SP - 1145

EP - 1155

JO - Biometrics

JF - Biometrics

SN - 0006-341X

IS - 4

ER -