This is a thread on the use of antibody tests (serology) to gauge individual risk for COVID-19 infection. It's triggered by the recent @WHO guidance and the push for "immunity certificates" to get people back into the workplace. @jeremyfarrar @laurie_garrett @mvankerkhove 1/8

The science isn’t there, and it could get us into trouble for two reasons: partial immunity and risk perception. A history of COVID-19 and/or circulating antibodies may not translate to complete protection in all individuals. This is the case for circulating coronaviruses. 2/8

We’ll likely see spectrum of protection, with many people having "partial immunity," meaning that they can be reinfected but with less severe symptoms or no symptoms at all. These individuals may "shed" lower amounts of virus but could still infect others. 3/8

Here's the problem: people with immunity certificates may not take the usual precautions to protect themselves, and they may be asked to go into environments with greater risk of COVID-19 exposure. 4/8

If they develop cold or flu-like symptoms, they're unlikely to be tested for COVID-19 (why would you if you have an immunity certificate?), and the people around them won't have a reason to quarantine and get tested themselves. 5/8

No one will be concerned about COVID-19 because of that immunity certificate. And unbeknownst to anyone, they're contributing to onward transmission of the virus. 6/8

This is about risk for people *around* the immune-certificate-holding person, and it’s tricky risk communication. We can tell "immune" persons that they should act as if they could get COVID-19 again, but what's the point of an immunity certificate that suggests the opposite? 7/8

We need to understand correlates of COVID-19 immunity before we use serology for individual risk-assessment. This doesn't take away from the criticality of these tests to understand epidemiology, the impact of herd immunity, and the basic immunology of the virus. 8/8