• I thought this was worth noting – The “T-Detect Covid” test uses T cell receptor matching comparisons to population based panels developed with the aid of machine learning to detect previous COVID-19 infection, supposedly with 97.1% sensitivity and 100% specificity. They (Adaptive Biotechnologies in a partnership with Microsoft – I have no financial interest in either) are continuing to refine their models as variants appear and hope to improve those numbers (and, I’m guessing, apply the same methodology to the diagnosis of other infections).
It will be interesting to learn what test results are in patients who have been vaccinated but not had actual disease, how long the positive result persists, and how well such a positive result correlates with immunity in comparison to more standard antibody tests.