Evaluating accuracy of diagnostic tests with intermediate results in the absence of a gold standard

Huiping Xu, Michael A. Black, Bruce A. Craig

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Intermediate test results often occur with diagnostic tests. When assessing diagnostic accuracy, it is important to properly report and account for these results. In the literature, these results are commonly discarded prior to analysis or treated as either a positive or a negative result. Although such adjustments allow sensitivity and specificity to be computed in the standard way, these forced decisions limit the interpretability and usefulness of the results. Estimation of diagnostic accuracy is further complicated when tests are evaluated without a gold standard. Although traditional latent class modeling can be readily applied to analyze these data and account for intermediate results, these models assume that tests are independent conditional on the true disease status, which is rarely valid in practice. We extend both the log-linear latent class model and the probit latent class model to accommodate the conditional dependence among tests while taking the intermediate results into consideration. We illustrate our methods using a simulation study and a published medical study on the detection of epileptiform activity in the brain.

Original languageEnglish
Pages (from-to)2571-2584
Number of pages14
JournalStatistics in Medicine
Volume32
Issue number15
DOIs
StatePublished - Jul 10 2013

Fingerprint

Diagnostic Tests
Routine Diagnostic Tests
Gold
Diagnostic Accuracy
Latent Class Model
Sensitivity and Specificity
Brain
Latent Class
Probit Model
Log-linear Models
Interpretability
Standards
Specificity
Adjustment
Simulation Study
Valid
Modeling

Keywords

  • Diagnostic tests
  • Intermediate results
  • Latent class analysis
  • Sensitivity
  • Specificity

ASJC Scopus subject areas

  • Epidemiology
  • Statistics and Probability

Cite this

Evaluating accuracy of diagnostic tests with intermediate results in the absence of a gold standard. / Xu, Huiping; Black, Michael A.; Craig, Bruce A.

In: Statistics in Medicine, Vol. 32, No. 15, 10.07.2013, p. 2571-2584.

Research output: Contribution to journalArticle

@article{d88f9e156b654a15afbf47ce330a49a7,
title = "Evaluating accuracy of diagnostic tests with intermediate results in the absence of a gold standard",
abstract = "Intermediate test results often occur with diagnostic tests. When assessing diagnostic accuracy, it is important to properly report and account for these results. In the literature, these results are commonly discarded prior to analysis or treated as either a positive or a negative result. Although such adjustments allow sensitivity and specificity to be computed in the standard way, these forced decisions limit the interpretability and usefulness of the results. Estimation of diagnostic accuracy is further complicated when tests are evaluated without a gold standard. Although traditional latent class modeling can be readily applied to analyze these data and account for intermediate results, these models assume that tests are independent conditional on the true disease status, which is rarely valid in practice. We extend both the log-linear latent class model and the probit latent class model to accommodate the conditional dependence among tests while taking the intermediate results into consideration. We illustrate our methods using a simulation study and a published medical study on the detection of epileptiform activity in the brain.",
keywords = "Diagnostic tests, Intermediate results, Latent class analysis, Sensitivity, Specificity",
author = "Huiping Xu and Black, {Michael A.} and Craig, {Bruce A.}",
year = "2013",
month = "7",
day = "10",
doi = "10.1002/sim.5695",
language = "English",
volume = "32",
pages = "2571--2584",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "15",

}

TY - JOUR

T1 - Evaluating accuracy of diagnostic tests with intermediate results in the absence of a gold standard

AU - Xu, Huiping

AU - Black, Michael A.

AU - Craig, Bruce A.

PY - 2013/7/10

Y1 - 2013/7/10

N2 - Intermediate test results often occur with diagnostic tests. When assessing diagnostic accuracy, it is important to properly report and account for these results. In the literature, these results are commonly discarded prior to analysis or treated as either a positive or a negative result. Although such adjustments allow sensitivity and specificity to be computed in the standard way, these forced decisions limit the interpretability and usefulness of the results. Estimation of diagnostic accuracy is further complicated when tests are evaluated without a gold standard. Although traditional latent class modeling can be readily applied to analyze these data and account for intermediate results, these models assume that tests are independent conditional on the true disease status, which is rarely valid in practice. We extend both the log-linear latent class model and the probit latent class model to accommodate the conditional dependence among tests while taking the intermediate results into consideration. We illustrate our methods using a simulation study and a published medical study on the detection of epileptiform activity in the brain.

AB - Intermediate test results often occur with diagnostic tests. When assessing diagnostic accuracy, it is important to properly report and account for these results. In the literature, these results are commonly discarded prior to analysis or treated as either a positive or a negative result. Although such adjustments allow sensitivity and specificity to be computed in the standard way, these forced decisions limit the interpretability and usefulness of the results. Estimation of diagnostic accuracy is further complicated when tests are evaluated without a gold standard. Although traditional latent class modeling can be readily applied to analyze these data and account for intermediate results, these models assume that tests are independent conditional on the true disease status, which is rarely valid in practice. We extend both the log-linear latent class model and the probit latent class model to accommodate the conditional dependence among tests while taking the intermediate results into consideration. We illustrate our methods using a simulation study and a published medical study on the detection of epileptiform activity in the brain.

KW - Diagnostic tests

KW - Intermediate results

KW - Latent class analysis

KW - Sensitivity

KW - Specificity

UR - http://www.scopus.com/inward/record.url?scp=84879110680&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84879110680&partnerID=8YFLogxK

U2 - 10.1002/sim.5695

DO - 10.1002/sim.5695

M3 - Article

VL - 32

SP - 2571

EP - 2584

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 15

ER -