A pseudo-likelihood approach for estimating diagnostic accuracy of multiple binary medical tests

Wei Liu, Bo Zhang, Zhiwei Zhang, Baojiang Chen, Xiao Hua Zhou

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Latent class models with crossed subject-specific and test(rater)-specific random effects have been proposed to estimate the diagnostic accuracy (sensitivity and specificity) of a group of binary tests or binary ratings. However, the computation of these models are hindered by their complicated Monte Carlo Expectation-Maximization (MCEM) algorithm. In this article, a class of pseudo-likelihood functions is developed for conducting statistical inference with crossed random-effects latent class models in diagnostic medicine. Theoretically, the maximum pseudo-likelihood estimation is still consistent and has asymptotic normality. Numerically, our results show that not only the pseudo-likelihood approach significantly reduces the computational time, but it has comparable efficiency relative to the MCEM algorithm. In addition, dimension-wise likelihood, one of the proposed pseudo-likelihoods, demonstrates its superior performance in estimating sensitivity and specificity.

Original languageEnglish (US)
Pages (from-to)85-98
Number of pages14
JournalComputational Statistics and Data Analysis
Volume84
DOIs
StatePublished - Jan 1 2015

Fingerprint

Diagnostic Accuracy
Pseudo-likelihood
Latent Class Model
Monte Carlo Algorithm
Expectation-maximization Algorithm
Binary
Specificity
Pseudo-maximum Likelihood
Models of Computation
Relative Efficiency
Random Effects Model
Maximum likelihood estimation
Likelihood Function
Random Effects
Statistical Inference
Asymptotic Normality
Medicine
Likelihood
Diagnostics
Estimate

Keywords

  • Composite likelihood
  • Imperfect reference standards
  • Latent class models
  • Random effects
  • Sensitivity and specificity

ASJC Scopus subject areas

  • Statistics and Probability
  • Computational Mathematics
  • Computational Theory and Mathematics
  • Applied Mathematics

Cite this

A pseudo-likelihood approach for estimating diagnostic accuracy of multiple binary medical tests. / Liu, Wei; Zhang, Bo; Zhang, Zhiwei; Chen, Baojiang; Zhou, Xiao Hua.

In: Computational Statistics and Data Analysis, Vol. 84, 01.01.2015, p. 85-98.

Research output: Contribution to journalArticle

Liu, Wei ; Zhang, Bo ; Zhang, Zhiwei ; Chen, Baojiang ; Zhou, Xiao Hua. / A pseudo-likelihood approach for estimating diagnostic accuracy of multiple binary medical tests. In: Computational Statistics and Data Analysis. 2015 ; Vol. 84. pp. 85-98.
@article{1bb030bb459c406792845baa52f54e67,
title = "A pseudo-likelihood approach for estimating diagnostic accuracy of multiple binary medical tests",
abstract = "Latent class models with crossed subject-specific and test(rater)-specific random effects have been proposed to estimate the diagnostic accuracy (sensitivity and specificity) of a group of binary tests or binary ratings. However, the computation of these models are hindered by their complicated Monte Carlo Expectation-Maximization (MCEM) algorithm. In this article, a class of pseudo-likelihood functions is developed for conducting statistical inference with crossed random-effects latent class models in diagnostic medicine. Theoretically, the maximum pseudo-likelihood estimation is still consistent and has asymptotic normality. Numerically, our results show that not only the pseudo-likelihood approach significantly reduces the computational time, but it has comparable efficiency relative to the MCEM algorithm. In addition, dimension-wise likelihood, one of the proposed pseudo-likelihoods, demonstrates its superior performance in estimating sensitivity and specificity.",
keywords = "Composite likelihood, Imperfect reference standards, Latent class models, Random effects, Sensitivity and specificity",
author = "Wei Liu and Bo Zhang and Zhiwei Zhang and Baojiang Chen and Zhou, {Xiao Hua}",
year = "2015",
month = "1",
day = "1",
doi = "10.1016/j.csda.2014.11.006",
language = "English (US)",
volume = "84",
pages = "85--98",
journal = "Computational Statistics and Data Analysis",
issn = "0167-9473",
publisher = "Elsevier",

}

TY - JOUR

T1 - A pseudo-likelihood approach for estimating diagnostic accuracy of multiple binary medical tests

AU - Liu, Wei

AU - Zhang, Bo

AU - Zhang, Zhiwei

AU - Chen, Baojiang

AU - Zhou, Xiao Hua

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Latent class models with crossed subject-specific and test(rater)-specific random effects have been proposed to estimate the diagnostic accuracy (sensitivity and specificity) of a group of binary tests or binary ratings. However, the computation of these models are hindered by their complicated Monte Carlo Expectation-Maximization (MCEM) algorithm. In this article, a class of pseudo-likelihood functions is developed for conducting statistical inference with crossed random-effects latent class models in diagnostic medicine. Theoretically, the maximum pseudo-likelihood estimation is still consistent and has asymptotic normality. Numerically, our results show that not only the pseudo-likelihood approach significantly reduces the computational time, but it has comparable efficiency relative to the MCEM algorithm. In addition, dimension-wise likelihood, one of the proposed pseudo-likelihoods, demonstrates its superior performance in estimating sensitivity and specificity.

AB - Latent class models with crossed subject-specific and test(rater)-specific random effects have been proposed to estimate the diagnostic accuracy (sensitivity and specificity) of a group of binary tests or binary ratings. However, the computation of these models are hindered by their complicated Monte Carlo Expectation-Maximization (MCEM) algorithm. In this article, a class of pseudo-likelihood functions is developed for conducting statistical inference with crossed random-effects latent class models in diagnostic medicine. Theoretically, the maximum pseudo-likelihood estimation is still consistent and has asymptotic normality. Numerically, our results show that not only the pseudo-likelihood approach significantly reduces the computational time, but it has comparable efficiency relative to the MCEM algorithm. In addition, dimension-wise likelihood, one of the proposed pseudo-likelihoods, demonstrates its superior performance in estimating sensitivity and specificity.

KW - Composite likelihood

KW - Imperfect reference standards

KW - Latent class models

KW - Random effects

KW - Sensitivity and specificity

UR - http://www.scopus.com/inward/record.url?scp=84917736404&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84917736404&partnerID=8YFLogxK

U2 - 10.1016/j.csda.2014.11.006

DO - 10.1016/j.csda.2014.11.006

M3 - Article

VL - 84

SP - 85

EP - 98

JO - Computational Statistics and Data Analysis

JF - Computational Statistics and Data Analysis

SN - 0167-9473

ER -