Quick synopsis: San Francisco had 40,000 confirmed covid infections between Dec 16 2021 and Jan 16th 2022. Between Dec 16th and Jan 31st there were 48 covid deaths. Based on the analysis below it appears the unvaccinated are now dying from covid at a rate on par with an average flu season.
As I write this, San Francisco’s Omicron wave has crested. Here’s how it looked on Jan 30th:
We can agree that’s good news. As elsewhere, the fact that SF is highly vaccinated (86% of those age 5+) didn’t mean much for slowing the spread. The chart above shows this; the case rate for vaccinated individuals went from 7 per 100k on December 2nd to 230 per 100k on Jan 8th, an increase of 33.3x. (The unvaccinated rates went from 17/100k to 500/100k in the same period, a 29x increase.)
San Francisco has been in a state of semi-lockdown since March 2020. Offices have shuttered with everyone working from home who could do so, schools were online the entire 2020 academic year, gyms were closed for extended periods of time and indoor mask mandates have been in place throughout, with the brief exception of July-August 2021 and even then I can confirm a large majority continued to ‘mask up’ indoors.
Since August 2021, vaccine passports have been required for bars, restaurants and most indoor venues. No city in the US has been more restrictive vis-a-vis covid than here.
Despite these measures we see the chart at top. None of it slowed Omicron. Yet we still have mask mandates and vaccine passports.
While Omicron made a mockery of slowing the spread, what about deaths and hospitalizations? Did these measures make a difference, and if so, how much? Are we facing a ‘pandemic of the unvaccinated’ in San Francisco?
To find out, let’s see how many unvaccinated people have caught Omicron, how many died, and calculate the case fatality rate (CFR) and from there estimate the ‘true’ death rate, or infection fatality rate (IFR.)
San Francisco’s Department of Public Health (DPH) has a dashboard showing the number of daily reported infections, by vaccination status. Using it, we can see how many vaccinated and unvaccinated people had confirmed infections over a given period.
In this post we’ll look at confirmed infections for the period of Dec 16th through Jan 16th and deaths from Dec 16th through Jan 31st and see how things are going.
The reason for these timeframes is A) there were a large number of infections from 12/16/2021-1/16/2022, giving us a healthy sample size and B) stopping on Jan 16th allows for a two week period for deaths to ‘catch up’— meaning, if someone had a confirmed infection on Jan 16th, if they were going to die of covid they likely would have done so within two weeks.
I’m writing this on Feb 10th, which gives SF DPH ten days to backfill any data lag into the infection and death tallies for the 12/16-1/31 timeframe. That seems reasonable, though there may still be some adjustments.
During the Dec 16th - Jan 16th period, there were 9,464 confirmed unvaccinated infections and 30,219 confirmed vaccinated. We see this from the DPH dataset below (click image to zoom):
As of Feb 10th, There were 48 recorded deaths between Dec 16th and Jan 31st:(**update 2/12/2022—there are revised numbers for Jan. See bottom of post.)
Let’s be generous, in a sense, and assume every death from Dec 16th - Jan 31st was a result of a confirmed infection during that same period. For instance, it’s unlikely someone would be confirmed infected on Dec 16th and die on the 17th but we’ll play it that way.
Being even more generous, if we assume all 48 of these deaths were in the unvaccinated population, that means a CFR of 48/9464 = .5% or a survival rate of 99.5%.
This is the unvaccinated worst-case scenario, where *every* death was in an unvaccinated person. (It also means there were zero deaths in 30k vaccinated cases, which is a stretch.)
Compare this with, say, the flu in 2017-2018, where you had 51646 deaths out of 41 Million cases, or a fatality rate of ~.12% . (The annual flu number has floated in the .1%-.17% range from 2014 to 2019.)
That’s big, .5% vs .12%. But is the actual unvaccinated death rate close to that?
We need to get back to caveats. The flu numbers from the CDC appear to be a range estimate, involving 95% confidence intervals and such, and so appear to be infection fatality rates than case fatality rates. (I could be wrong, corrections welcome.)
Whether that’s true or not we do know the worst scenario .5% CFR for unvaccinated people in San Francisco is a case fatality rate, involving confirmed infections.
But how many actual infections were there over the same period? Maybe double the number of recorded cases?
I would suggest that’s reasonable given the relative mildness of Omicron along with the burden of going out to get tested at a facility —why bother if you’re not feeling that bad— not to mention the large number of asymptomatic cases. If so that would make the infection fatality rate half that of the CFR, or .25%. That in turn makes Omicron twice as ‘bad’ as the 2017 flu season among the unvaccinated.
Another reason for the 50% estimate is based on this paper, in which the authors figure the CFR is 80% of the IFR in Hong Kong and Singapore, places where they are *very* diligent in checking the population for covid and the public health authorities don’t mind some oppressive measures to do so. (The paper also mentioned Australia, a theory that dissolved just as it was published, what with the Omicron explosion there. No way they’re recording 80% of daily infections now Down Under.)
San Francisco takes covid seriously but there aren’t DPH employees going door to door demanding to know if anyone has a sniffle and if so forcing them to test. So a 50% estimate makes sense I think.
The other big caveat: how many of those 48 deaths were in the unvaccinated population? The .5% CFR assumes *all 48* were unvaccinated. SF DPH doesn’t break down covid deaths by vaccination status.
Interestingly I’ve also looked at the Covid / Public Health dashboards for Los Angeles, Houston, Denver, Cincinnati and Detroit and none of them break down deaths by vaccination status either. Of the six only San Francisco even distinguishes cases by vax status.
I’ve written to SF DPH and filed a public records request and have been told they don’t have the information, that covid deaths in San Francisco are not recorded by vaccination status. This seems odd to say the least.
Looking around, I saw that New York City does have this information. On their public health dashboard you can find cases, hospitalizations and deaths broken down by vaccination status.
I’ve downloaded the most recent numbers and created a spreadsheet which you can find here. The takeaway is:
Between 11/13/2021 and 1/29/2022 there were 528427 confirmed vaccinated cases and 1426 deaths, for a CFR of .269%. There were 484187 confirmed unvaccinated cases and 2358 deaths, for a CFR of .487%.
So out of a million confirmed NYC cases (over a ten week span!) there was a 2:1 ratio in the death rate between the vaccinated and unvaccinated. We keep hearing that there’s a 20-1, 30-1 greater chance of the unvaccinated dying vs vaccinated but I’m not sure what sort of adjustments they’re making to get that.
In raw numbers, half a million New Yorkers in each category, there’s a 2:1 ratio in the death rate since mid November.
If we apply this same 2:1 ratio to San Francisco, you would get 32 of the 48 deaths in the unvaccinated group, for a CFR of (32/9464) = .33%, or a survival rate of 99.67%. If we assume an IFR that’s half the CFR, the survival rate is ~99.84%.
This may be assigning too many deaths to the unvaccinated in the case of SF, since the confirmed case rate here is 3:1 in favor of the vaccinated whereas in NYC it’s almost 1:1 in the sample we have, but we’re giving the vaccinated every benefit of the doubt.
So with all those caveats it appears that in January 2022, out of a 9400 person sample size, unvaccinated San Franciscans had a covid death rate on par with the flu.
(**Update 2/12/2022. After I posted this on 2/10, DPH —naturally— revised the January death count from 42 to 60. So this means there were 66 deaths total from 12/15/2021-1/31/2022. There was also a revisal of the case numbers, and the unvaccianted count went from 9464 to 10,580 from 12/16-1/16. Vaccinated case count went to 31,794.
Running the same numbers and using the same assumptions we get an unvaccinated CFR of (66*2/3)/10580= .415, an estimated IFR of (.415/2) = .21, and thus a survival rate of 99.79%)
Does San Francisco have a data dashboard? We do in Massachsuetts. Boston and metro region are really the whole population (add Worcester and Springfield) so it's an easy way to track status. In ages 50 + the vaccination rate is around 90%. Yet the deaths are concentrated still in that age range. Hospital usage rates also in over age 65. So what's going on there? I have this on my stack, but no way to post images here. HoldLLC on Twitter has also covered the data miss handling going on with NYC charts, and you should continue to follow these lines. Like the substack! Will keep checking it out.