Are case counts limited by testing capacity? Great Q from @rachidelguerrab, but not easy to answer.
Short thread
Short thread
A "confirmed case" of #COVID19 almost always refers to a person who has tested positive for RNA of the SARS-COV-2 virus (usually by nasal swab, but also by oral, throat, or deeper sampling). This means that the # of confirmed cases can't be ># of tests.
Due to the very wide set of symptoms associated w/ #COVID19, many people that aren't infected w/ SARS-COV-2 are tested for it. In addition, the test can miss infection which is more likely later after symptom onset when severe symptoms start (day 10 ~25%).
medrxiv.org/content/10.110…
medrxiv.org/content/10.110…
This means that the fraction of tests that are positive is rarely 100%, but early in epidemic % of tests that were + was amazingly high (35-50% in NY: fig). As testing capacity ramped up %+ dropped everywhere in US, usually to <10%. In NY it's now <3%.
But as infections rise, more people come to hospitals, & more tests are done for these people, making cases & % tests positive increase. As infections increase this has resulted in huge delays in getting results. But have Drs stopped doing tests? Probably for mild cases.
How do we know if we are seeing a drop in cases due to testing capacity limitation? I don't have an exact answer, but data from AZ is worth considering. %+ reached ~30% July 8 & is still ~25%. Given missed cases b/c testing done late (see above), it does seem possible.
These %+ in AZ are near the %+ during peak of epidemic in NYC, so I'm betting we're missing more mild infections in AZ now that Drs simply aren't recommending for testing.
Would love to hear other ways to try to detect testing capacity & see other datasets brought to bear.
Would love to hear other ways to try to detect testing capacity & see other datasets brought to bear.
