testy testaments to tardy tests requests.
Tests: Direct Methods- actual detection of live virus (culture), detection of viral nucleic acids (rtPCR), detection of viral proteins (Antigen detection). Antigen detection requires generation of an antibody that will bind a virus target protein within a bodily fluid/tissue. These will take a bit longer due to how the targeting antibodies are generated.
Indirect Methods- these are surrogate markers for infection- Antibody assays IgM/ IgG/IgA. IgM antibodies are a marker of recent infection and fade (generally) fast with time. IgG antibodies are usually long lasting but often do not distinguish current infection from past infection with the same infectious agent. IgA are mucosal antibodies and are (generally) short lived. Cytokine/host marker profiles- elevated /altered cirulating signaling molecules may define a disease state. Indirect assays can also measure host metabolite changes, volatile organic compounds etc.
That sum things up accurately. One can look at which tests have been grant FDA Emergency Use Authorization (EUA)
here
https://www.fda.gov/medical-devices/...ons#covid19ivd so see which tests have EUA approval. Several where I work should be added to this list by the end of the week.
So far there are 4 antibody tests on the list, 2 are rapid/ what is known as lateral flow assay format and 2 are lab based ELISA format. ELISA is usually more sensitive but takes ~3-4 hours to run with incubation times. Rapid may be completed in as little as 5-15 minutes but lacks sensitivity at lower antibody levels.
Testing- sensitivity (true pos- false negatives) and specificity (true negatives- false positives) of the assay are what drives accuracy of a test.
Now how do we test and who? First, we need to compare a test against known positives by another metric (rtPCR or culture) to determine how sensitive it is within a disease population. We test against known negatives to determine specificity.
But how we select even these populations can skew sensitivity/ specificity. For instance, antibody (again generally) takes 7-10 days to arise post infection or onset of symptoms. rtPCR is more likely to be positive when viral load is higher and begins to decrease around day 10. So antibody tests on people who are between day 0-10 onset of symptoms are more likely to be rtPCR+ and Ab(-) during this time. After day15, most patients will convert to antibody positive but may now be rtPCR negative.
So if we derive sensitivity after day10, the numbers will be higher >95%, than studies when earlier time point samples are included as well.
For specificity, population context is meaningful. Are we testing samples in an area where disease is endemic or non-endemic? Are we testing against other diseases with similar signs and symptoms? Are we testing against related viral infections? The specificity in these populations may be very different.
Serosurveilence requires at least modest sensitivity but very high specificity and is tied to overall prevelance. When prevelence is low say <5%, specificity must be very high and sampling size must be large otherwise the error in the false positive rate becomes large.
Bookmarks