3 ms·

Here's a piece of data I've been following, for Santa Clara County: https://covid19.sccgov.org/dashboard-case-rates-vaccination-status https://covid19.sccgov.o

by mlyle 5y ago

Here's a piece of data I've been following, for Santa Clara County: https://covid19.sccgov.org/dashboard-case-rates-vaccination-status https://covid19.sccgov.org/dashboard-case-rates-vaccination-... The number --- 871 cases per day per 100,000 unvax'd, for the week preceding 1/10-- is an amazing rate (7 day average). It tells us that 6.1% of the unvaxxed population became a case during a 1 week period. This doesn't count people who were infected and did not test or did not test in a way that got reported. Your point about whether mild infections are slipping through is interesting. Probably the vaccinated have more paucisymptomatic infections. The unvaccinated and vaccinated probably test at different rates (which way? required workplace testing of unvaxxed vs. unvaxxed more likely to be deniers who don't want to test). Etc. So there's some confounds here, for sure. But boy that 871 per 100k number is fascinating. Without reinfection (which we don't suspect is likely for Omicron-following-Omicron), it certainly can't stay at 6% of the population per week for long. I doubt they have spotted more than a third of infections.

derobert 5y ago
A quick warning: be wary of that computation because of how it's likely made, and it's worse than higher the vaccination rate gets. The sources section says: > Sources: California Reportable Disease Information Exchange; California Department of Public Health, California Immunization Registry; the U.S. Census Bureau, 2015-2019 American Community Survey 5-Year Estimates. So the the numerator (number of cases) from the reportable disease database and/or department of health data (which probably includes vaccination status, or alternatively is matched against the registry). The denominator for the rate amongst the total population is the population estimate from the census bureau. The denominator for the rate amongst vaccinated comes from the immunization registry. The rate against the unvaccinated? There is no count of them, so instead you subtract the number vaccinated from the population. The problem is that both of those are estimates, and as they get closer the error on that denominator goes through the roof. So the rate amongst the unvaccinated becomes meaningless. As an example, take a hypothetical place with around 100 people. Might actually be anywhere from 98–102, that is ±2%. If you've got a very high vaccination rate (yeah!), say there are 96 people on the immunization registry. You could have anywhere from 2 to 6 people unvaccinated. You'd report 4, but the error is now ±50%, which is huge. And when you use that as the denominator, you get very different answers. (OTOH, of half the village was unvaccinated, 50±4% is a much smaller error!)
- mlyle 5y ago
  Yes, this is a good point. On the other hand, while S.C.C. has a high vaccination rate, it's not high enough that a small population uncertainty makes such a huge difference.

by mlyle 5y ago

derobert 5y ago

A quick warning: be wary of that computation because of how it's likely made, and it's worse than higher the vaccination rate gets. The sources section says: > Sources: California Reportable Disease Information Exchange; California Department of Public Health, California Immunization Registry; the U.S. Census Bureau, 2015-2019 American Community Survey 5-Year Estimates. So the the numerator (number of cases) from the reportable disease database and/or department of health data (which probably includes vaccination status, or alternatively is matched against the registry). The denominator for the rate amongst the total population is the population estimate from the census bureau. The denominator for the rate amongst vaccinated comes from the immunization registry. The rate against the unvaccinated? There is no count of them, so instead you subtract the number vaccinated from the population. The problem is that both of those are estimates, and as they get closer the error on that denominator goes through the roof. So the rate amongst the unvaccinated becomes meaningless. As an example, take a hypothetical place with around 100 people. Might actually be anywhere from 98–102, that is ±2%. If you've got a very high vaccination rate (yeah!), say there are 96 people on the immunization registry. You could have anywhere from 2 to 6 people unvaccinated. You'd report 4, but the error is now ±50%, which is huge. And when you use that as the denominator, you get very different answers. (OTOH, of half the village was unvaccinated, 50±4% is a much smaller error!)

mlyle 5y ago
Yes, this is a good point. On the other hand, while S.C.C. has a high vaccination rate, it's not high enough that a small population uncertainty makes such a huge difference.