Methods for evaluating the performance of diagnostic tests in the absence of a gold standard: A latent class model approach

Elizabeth S. Garrett, William W Eaton, Scott Zeger

Research output: Contribution to journalArticle

Abstract

In many areas of medical research, 'gold standard' diagnostic tests do not exist and so evaluating the performance of standardized diagnostic criteria or algorithms is problematic. In this paper we propose an approach to evaluating the operating characteristics of diagnoses using a latent class model. By defining 'true disease' as our latent variable, we are able to estimate sensitivity, specificity and negative and positive predictive values of the diagnostic test. These methods are applied to diagnostic criteria for depression using Baltimore's Epidemiologic Catchment Area Study Wave 3 data.

Original languageEnglish (US)
Pages (from-to)1289-1307
Number of pages19
JournalStatistics in Medicine
Volume21
Issue number9
DOIs
StatePublished - May 15 2002

Fingerprint

Latent Class Model
Diagnostic Tests
Routine Diagnostic Tests
Gold
Diagnostics
Predictive Value of Tests
Baltimore
Operating Characteristics
Latent Variables
Specificity
Biomedical Research
Sensitivity and Specificity
Estimate
Standards

Keywords

  • Depression
  • Diagnostic criteria
  • Latent class models
  • Markov chain Monte Carlo
  • Operating characteristics
  • Validation

ASJC Scopus subject areas

  • Epidemiology

Cite this

Methods for evaluating the performance of diagnostic tests in the absence of a gold standard : A latent class model approach. / Garrett, Elizabeth S.; Eaton, William W; Zeger, Scott.

In: Statistics in Medicine, Vol. 21, No. 9, 15.05.2002, p. 1289-1307.

Research output: Contribution to journalArticle

@article{4a028e9bea9a45d89854176ffd9f679e,
title = "Methods for evaluating the performance of diagnostic tests in the absence of a gold standard: A latent class model approach",
abstract = "In many areas of medical research, 'gold standard' diagnostic tests do not exist and so evaluating the performance of standardized diagnostic criteria or algorithms is problematic. In this paper we propose an approach to evaluating the operating characteristics of diagnoses using a latent class model. By defining 'true disease' as our latent variable, we are able to estimate sensitivity, specificity and negative and positive predictive values of the diagnostic test. These methods are applied to diagnostic criteria for depression using Baltimore's Epidemiologic Catchment Area Study Wave 3 data.",
keywords = "Depression, Diagnostic criteria, Latent class models, Markov chain Monte Carlo, Operating characteristics, Validation",
author = "Garrett, {Elizabeth S.} and Eaton, {William W} and Scott Zeger",
year = "2002",
month = "5",
day = "15",
doi = "10.1002/sim.1105",
language = "English (US)",
volume = "21",
pages = "1289--1307",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "9",

}

TY - JOUR

T1 - Methods for evaluating the performance of diagnostic tests in the absence of a gold standard

T2 - A latent class model approach

AU - Garrett, Elizabeth S.

AU - Eaton, William W

AU - Zeger, Scott

PY - 2002/5/15

Y1 - 2002/5/15

N2 - In many areas of medical research, 'gold standard' diagnostic tests do not exist and so evaluating the performance of standardized diagnostic criteria or algorithms is problematic. In this paper we propose an approach to evaluating the operating characteristics of diagnoses using a latent class model. By defining 'true disease' as our latent variable, we are able to estimate sensitivity, specificity and negative and positive predictive values of the diagnostic test. These methods are applied to diagnostic criteria for depression using Baltimore's Epidemiologic Catchment Area Study Wave 3 data.

AB - In many areas of medical research, 'gold standard' diagnostic tests do not exist and so evaluating the performance of standardized diagnostic criteria or algorithms is problematic. In this paper we propose an approach to evaluating the operating characteristics of diagnoses using a latent class model. By defining 'true disease' as our latent variable, we are able to estimate sensitivity, specificity and negative and positive predictive values of the diagnostic test. These methods are applied to diagnostic criteria for depression using Baltimore's Epidemiologic Catchment Area Study Wave 3 data.

KW - Depression

KW - Diagnostic criteria

KW - Latent class models

KW - Markov chain Monte Carlo

KW - Operating characteristics

KW - Validation

UR - http://www.scopus.com/inward/record.url?scp=0037094078&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0037094078&partnerID=8YFLogxK

U2 - 10.1002/sim.1105

DO - 10.1002/sim.1105

M3 - Article

C2 - 12111879

AN - SCOPUS:0037094078

VL - 21

SP - 1289

EP - 1307

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 9

ER -