Is it possible to calculate sensitivity, specificity, and accuracy between two serological tests without a gold standard test during the disease prevalence test among a group of patients?
If it is possible, what is the appropriate statistical method and its method?
I want to start by saying I'm not a statistician. However, I have ran into this kind of issue In the past when performing a qualification for an assay. Ideally you will get sensitivity and specificity through the use of known positive and known negative patient samples and diluting your positives in pooled negative samples. An orthogonal test is very helpful here, allowing you to screen and categorize the samples and then test concordance between the assays. But accuracy is more challenging. I think the best you could do is infer it from the specificity, sensitivity, and linearity. Here are some links to the FDA guidances:
http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/default.htm
http://www.fda.gov/BiologicsBloodVaccines/GuidanceComplianceRegulatoryInformation/Guidances/default.htm
http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/default.htm
http://www.fda.gov/RegulatoryInformation/Guidances/default.htm
The guidance that may be most useful here is:
Assay Development and Validation for Immunogenicity Testing of Therapeutic Protein Products
Guidance for Industry
There is an ebook I found called:
Statistical Methods for Immunogenicity Assessment
written by Harry Yang et al.
It goes into some statistics for assay validation.
Hope this helps!
Dave
David Allen Hokey
Thank you for your answer. it is very useful; I will read it carefully. I hope to find my target there.
Regards
Elhaig
- Badges
- Science topic
- Similar topics
- Names