Antibody testing is probably still the number one topic of conversation at work these days. Although I suspect that this will rapidly be replaced by the impacts of children returning to school.
Based on my completely unscientific anecdotal evidence gleaned from snatches of conversations in the corridors, I have started to wonder a little bit at test performance.
My first observation from my study population of colleagues is that it is fair to say the the number of positive tests seems quite low, even in those clinical areas that were highly exposed. Does this reflect that some people simply don’t get infected or does it mean that we have superior hand washing technique or somehow desist from picking our noses?
We are also seeing a number of doctors and nurses who had a positive swab (antigen test) at the time of their symptoms, but who now have negative antibody tests. This also raises a lot of questions that I hope can be answered by those who actually get to see large numbers of patients and not just the stories in the tea room. Which test (if any) is actually correct? Are these people actually immune? Why are managers seemingly so interested in staff having antibody tests?
We have been focused on the false negatives, i.e., people who have symptoms but in whom the antigen test is negative. Interestingly, many of the patients in ITU fell into this category for reasons which we can again only speculate. Was it because many of the complications were from an over-exuberant reaction after most of the virus had been cleared from the body? Of importance, many healthcare workers had symptoms that were highly suggestive of COVID at the time, so it is difficult to write off these negatives as being true negatives. I am not currently aware of antibody testing in our intensive care patients but am sure data will emerge soon.
A last category are colleagues whose partners definitely had the virus but who themselves did not show any symptoms and who are now testing negative. Obviously this is interesting in the context of Karl Friston and others who are suggesting that the dynamics of the outbreak could be explained by a significant proportion of the population. Certainly an interesting theory, but my views on mathematical modelling as a single tool are familiar to those who have read this blog, i.e., maybe a useful adjunct but certainly not to be relied on.
What is also clear is that my surveys of other clinicians is a poor substitute for proper analysis of whole data sets. However, the high frequency of results that raise eye brows make me wonder what the test performances really are. I await these population level analyses with interest.