Methods for evaluating the performance of diagnostic tests in the absence of a gold standard: a latent class model approach.
E. Garrett, W. Eaton, and S. Zeger. Statistics in medicine, 21 (9):
1289-307(May 2002)5560<m:linebreak></m:linebreak>LR: 20071114; CI: Copyright 2002; GR: 5 R01 MH56639-03/MH/NIMH NIH HHS/United States; GR: MH 47447/MH/NIMH NIH HHS/United States; JID: 8215016; ppublish;<m:linebreak></m:linebreak>Proves diagnòstiques.
DOI: 10.1002/sim.1105
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.
%0 Journal Article
%1 Garrett2002
%A Garrett, Elizabeth S
%A Eaton, William W
%A Zeger, Scott
%D 2002
%J Statistics in medicine
%K Baltimore Biological Depression Depression:diagnosis DiagnosticTests Humans MarkovChains Models MonteCarloMethod PredictiveValueofTests Routine Routine:standards SensitivityandSpecificity
%N 9
%P 1289-307
%R 10.1002/sim.1105
%T Methods for evaluating the performance of diagnostic tests in the absence of a gold standard: a latent class model approach.
%U http://www.ncbi.nlm.nih.gov/pubmed/12111879
%V 21
%X 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.
%@ 0277-6715; 0277-6715
@article{Garrett2002,
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.},
added-at = {2023-02-03T11:44:35.000+0100},
author = {Garrett, Elizabeth S and Eaton, William W and Zeger, Scott},
biburl = {https://www.bibsonomy.org/bibtex/2714b997ce811acd45831d75a36adbb89/jepcastel},
city = {Johns Hopkins University School of Medicine, Division of Biostatistics, Oncology Center, Baltimore, MD 21205, USA. esg@jhu.edu},
doi = {10.1002/sim.1105},
interhash = {448e5b903d8c8ae860ec2b5149cde3a5},
intrahash = {714b997ce811acd45831d75a36adbb89},
isbn = {0277-6715; 0277-6715},
issn = {0277-6715},
journal = {Statistics in medicine},
keywords = {Baltimore Biological Depression Depression:diagnosis DiagnosticTests Humans MarkovChains Models MonteCarloMethod PredictiveValueofTests Routine Routine:standards SensitivityandSpecificity},
month = {5},
note = {5560<m:linebreak></m:linebreak>LR: 20071114; CI: Copyright 2002; GR: 5 R01 MH56639-03/MH/NIMH NIH HHS/United States; GR: MH 47447/MH/NIMH NIH HHS/United States; JID: 8215016; ppublish;<m:linebreak></m:linebreak>Proves diagnòstiques},
number = 9,
pages = {1289-307},
pmid = {12111879},
timestamp = {2023-02-03T11:44:35.000+0100},
title = {Methods for evaluating the performance of diagnostic tests in the absence of a gold standard: a latent class model approach.},
url = {http://www.ncbi.nlm.nih.gov/pubmed/12111879},
volume = 21,
year = 2002
}