PakMediNet Discussion Forum : Biostatistics : sensitivity and specificity when there's no gold standard
sensitivity and specificity when there's no gold standard
I know that to calculate sensitivity and specificity of a given test we must compare it against a golden standard,that is a test with high sensitivity and specificity.
but there are circumstances when there is no gold standard and there are only tests or criteria that are suggestive and not diagnostic (take early disease in some forms of lymphoma as an example)
now is it possible to calculate sensitivity and specificity for a new test ( like polymerase chain reaction) against the existing "suggestive" criteria?
Posted by: bluesdudePosts: 2 :: 25-07-2005 :: | Reply to this Message
Re: sensitivity and specificity when there's no gold standard
In my opinion, it is not possible to calculate sensitivity and specificity of a diagnostic and other test without computing the values in 2x2 table against a 'gold standard'. Because such qulaititative aanalysis you need true positive and true negative subjects as well false positive and false negative subjects/cases. However, for the peculiarity of your research project, I suggest you to consider a type of biostatistical analysis called 'degree of agreement', but I am not sure of the appropriateness this analysis for your data unless I have a clear idea about the nature and type of your data.
[Edited by honesty on 15-08-2005 at 08:45 AM GMT]
Posted by: honestyPosts: 5 :: 15-08-2005 :: | Reply to this Message
Re: sensitivity and specificity when there's no gold standard
There are various methods of assessing diagnostic accuracy of a test in the abscence of a gold standard.
One method is to follow-up the patients, say for a 1-2 years depending on the target disease. Clinical diagnosis based on longitudinal follow-up forms the 'gold standard'.
Some references that might be helpful..
1: Joseph L, Gyorkos TW, Coupal L.
Bayesian estimation of disease prevalence and the parameters of diagnostic
tests in the absence of a gold standard.
Am J Epidemiol. 1995 Feb 1;141(3):263-72.
PMID: 7840100 [PubMed - indexed for MEDLINE]
2: Valenstein PN.
Evaluating diagnostic tests with imperfect standards.
Am J Clin Pathol. 1990 Feb;93(2):252-8. Review.
PMID: 2405632 [PubMed - indexed for MEDLINE]
3: Phelps CE, Hutson A.
Estimating diagnostic test accuracy using a "fuzzy gold standard".
Med Decis Making. 1995 Jan-Mar;15(1):44-57.
PMID: 7898298 [PubMed - indexed for MEDLINE]
Good luck,
Rehman
UK
Posted by: asiddiquiPosts: 26 :: 28-08-2005 :: | Reply to this Message
Re: Re: sensitivity and specificity when there's no gold standardWhat about clinical sensitivity based on a clinical case definition can a test be judged on that criteria for example during a dengue fever outbreak febrile illness with thrombocytopaenia is suggestive of dengue fever how will we test a Dengue IgM ELISA diagnostic sensitiity and if applicable which statistical tests will we apply in the absence of a gold standard diagnostic test
Posted by: mundriPosts: 2 :: 21-03-2007 :: | Reply to this Message