Specificity of antigen vs PCR-tests: data needed!
Many are understandably excited about the new rapid antigen tests. But there are a couple issues I haven't seen robust data on. Maybe you can help?
What is specificity of antigen tests vs PCR?
Thread
@michaelmina_lab
EUA (fda.gov/media/141570/d…) for BinaxNOW specificity (1-false +)=98.5% but:
- sample size is tiny (25 people submitting a total of 102 swabs from SYMPTOMATIC patients)
-95% CI (which should incorporate repeat testing & be analyzed as f(days since symptom onset) is: 92%-99.96%.
This means up to 8% (or more, if analysis done properly; totally unknown for asymptomatic people) of + antigen tests could be false +. If that's the case, these tests would be a disaster for screening people without reason to suspect they are infected (e.g. school kids).
EUAs often have tiny sample sizes, but millions of antigen tests have now been run. What is the real specificity (false +) w/ a reasonable sample size, how does it change vs days since symptom onset, & how does it compare to PCR specificity?
Speaking of which, what is specificity of PCR testing w/ 2 probes (e.g. CDC assay). I've again seen EUAs with tiny sample sizes often showing 100% specificity (on pre-pandemic samples), but haven't seen robust sample sizes. Can anyone help?
Share this Scrolly Tale with your friends.
A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.