DeaconBrews
Well-known member
- Joined
- Mar 27, 2011
- Messages
- 45,006
- Reaction score
- 3,181
Yes, possible to have false negative with a blood test. Antibody levels wane over time as well as take time to build. So say you were infected 2 weeks ago with an asymptomatic infection, took the test this week there's a chance that your antibody levels are below the serological cutoff for positive. This is especially true because a lot of these tests for some reason (Its a cost thing) are running a single IgG response to a single viral antigen. Then you have the opposite spectrum of too much time has passed with say you had the, I was really sick in January with something I never felt before confirmation bias thought and it actually was Covid-19 now that 4 months have passed your antibody levels may have shrunk to a level thats below cutoff.
The cutoff is usually based on a reference sample set, but as I said for some unknown reason this doesnt exist. So every company gets to choose there positive and negative control, now say you want to make sure your test really works you pick only samples from RT-PCR positive hospitalized patients that have recovered a month ago to maximize the amount of antibody for your positive test control thats what happened before the FDA finally realized this let everyone have a test was a dumb approach. You also have a shifting cutoff based on do you want specificity or sensitivity, It is often impossible to have both high sensitivity and high specificity, and you must select an appropriate trade-off. For
sero-surveillance of pathogens with low prevalence (<20%), you prioritizing specificity. For example, if the true sero-prevalence of SARS-CoV-2 in a population is 8%; a test with 100% sensitivity and 90% specificity, so catch all true positive but lots of false positives would return an estimate of 1x8% + 0.1x92% = 17.2%, a big difference from true prevalence.
Well.... shit