Resources

By mid-October 2020, the list of methods for SARS-COV-2 testing (molecular, antigen, rapid, and serology), broke the 870 mark.  Rather than ask laboratories to evaluate more than 870 possibilities, we've narrowed the candidates down to a short list.

Updated October 14, 2020 [This is out-of-date, in other words, useful mainly for historical purposes]

As of September, only 56 antibody tests have have EUA status from FDA in the US.
(FYI, the FDA have put 147 antibody methods on the Do Not Distribute list)

See even more stories about COVID-19 Laboratory Challenges...

One of the easiest ways to winnow the field is to focus on which manufacturers have submitted and obtained an EUA (Emergency Use Authorization). 

Next, is to narrow that further by focusing on methods that have achieved the FDA goal of 90% sensitivity and 95% specificity.

To further flesh out the list, we've been scouring the pre-print services to locate relevant papers that document the real-world performance of these methods. As we know from all other areas of testing, the first set of data on performance, delivered by the manufacturer itself, can sometimes be overly optimistic. What looks great at launch may not be that good in the field. So we can supplement the data from the manufacturer with reality checks from the front lines of the pandemic fight.

There are further sources of data - for instance, the NHS in the UK has evaluations performed by Public Health England Porton Down (PHE Porton Down). The FDA is conducting independent evaluations of serology methods and is releasing those on the Open FDA website. It's also useful to know which methods have been approved in other countries than the US. France, the UK, Canada have publicly accessible lists of their approved methods.

We've noted the sensitivity and specificity of these tests for a specific window of infection, for patients who are >10 or >14 or >15 days from either positive PCR test or symptoms. Given the immunology response curves, it's fruitless to expect IgG or IgM to have a higher sensitivity in the immediate days after infection.  This points to the continuing need for PCR testing - in order to detect the active disease, you need the diagnostic test. Serology is for surveillance and history and possible, we hope, proof of immunity.

We have noted, where possible, the number of positive and negative samples used to determine sensitivity and specificity of the methods.  More is better.  To see a method where they tested thousands of patient samples to determine specificity is heartening. You can have more confidence in that method than one where they tests only 30, 40, 50 samples.

Also note that even if a method doesn't meet the FDA goals for performance, that doesn't automatically discount them for use in your laboratory - particularly if you can consider an orthogonal approach. If you can combine too less-than-perfect tests in confirmatory protocols, you can get a better combined result than either single test can provide. It may seem like too much effort or expense to run more than one test, but the FDA is strongly advising it. Given the mass scale of testing we will need to perform to truly fight the pandemic, additional testing to make sure our results are as accurate as possible is critical.

Our aim is that when we can combine all these evaluations, we may not see that there is one perfect method, but definitely some methods seem to be performing better than others. This may shorten your list of methods to consider when you choose a serology method.

The data is changing every day, expect that this short list will expand, contract, fluctuate, etc. Stay tuned.