As COVID-19 fades from epidemic to endemic, we each have to make our own decisions about returning to normalcy.
My wife and I went back to the gym yesterday, and plan to go to the gym regularly from now on.
You can see Post #1163 (June 23, 2021) for a writeup of the last time we did that. This time was not all that different. The mental issues were about the same. You’ve trained yourself to look at some activity as risky. You tend to treat the risks of those activities as black-and-white: Either it’s OK to do something, or it’s not. And for a long time, going to the gym was Not OK. So this is akin to breaking a long-standing taboo. Rational thought only takes you so far.
To each his or her own. Some people never took any precautions during the pandemic. Others may have done so at one time, maybe even gotten a vaccine shot or two, but aren’t taking any other precautions now.
And then there are people like me, just trying to balance benefits and risks.
Give me a free, effective, and near-risk-free vaccine and I’ll take it. Find me a $1 mask that reduces my exposure by 95% and I’ll wear it. Show me a crowded indoor situation I don’t have to be in and I’ll avoid it. Unless there’s some good reason to be in it.
At any rate, as of yesterday, Fairfax County VA was down to 57 new cases per 100K per day. That’s a bit above my cutoff of 40 cases — the point at which the health risks from Omicron, for a vaccinated-and-boostered person, appear no larger than those from flu, in a typical week of flu season, using a very conservative estimate of risk.
That’s close enough, all things considered. You can see the calculations in Post #1163. Best guess, given my age and gender, the benefits of getting regular exercise once again vastly outweigh the COVID-related risk involved in going back to the gym. So back I go.
I have just three more things to talk about:
- Some perspective on the entire pandemic to date.
- A quick recap of comparing risks from Omicron and flu
- A some more in-depth look at COVID-19 vaccine protection against hospitalization and death.
Some perspective
Source: COVID cases and deaths: CDC COVID data tracker. COVID hospitalizations: Calculated from US DHHS unified hospital dataset Hospitalization data are missing prior to mid-2020.. Flu: Twice the median of values in Table 1, CDC Disease Burden of Flu.
Looking back over the entire pandemic, there’s no doubt that COVID-19 posed a far more serious problem than flu. As shown above, in just under two years of the pandemic, the U.S. had perhaps 30 percent more more formally-diagnosed COVID-19 cases than the number of symptomatic cases you would expect from two years of flu. But those cases generated at least five times as many hospitalizations, and about 12 times as many deaths. (I say “at least”, because nobody tracked hospitalizations for the first few months of the pandemic.)
That scorecard for the pandemic as a whole (so far) includes the effects of a lot of proactive measures. Not just the COVID-19 hygiene rules regarding masks and public gatherings, but also the elimination of a year of in-person schooling. And, for about half that period, the use of use of vaccines that were far more effective than the typical flu vaccine.
In other words, the table above reflects not just the virulence of the disease, but also those things done to minimize the impact of the disease. The disease itself was far more virulent, compared to flu, than the table above suggests.
Lest we forget, the original wave — the one that largely caught us unaware and unprepared — reflected the underlying severity of the disease, before we got proactive about it. Here’s how the case mortality rates appeared at the peak of each of the major waves:
Source: Calculated as ratio of peak deaths to peak cases, data from CDC COVID data tracker.
I’m supposed to say that maybe that first number (on the left) is exaggerated by the lack of testing. (Recall that, among other things, the CDC botched the first DNA tests and had to recall them.) But my recollection is that mortality rates in some European countries exceeded that, at the time. That’s also below the initial mortality rate reported for Wuhan. So maybe that reflects a lack of testing in the U.S. And maybe that’s just how bad COVID-19 was before anyone had come to grips with how to deal with it.
By the time we got to the third wave, we had the Delta variant, which was far more virulent than its predecessors. But by that time, more-or-less every adult who wanted to be vaccinated had been. And the net result was a reduction in the case mortality rate. But that’s not a uniform reduction. The average case mortality rate fell because the vast reduction for the vaccinated more-than-offset the high rate for the un-vaccinated.
If we look at the CDC’s new analysis of cases and deaths by vaccination status (age-adjusted, age 12 and older only), we can see that the un-vaccinated have had about a 15-fold greater chance of dying from COVID-19. That’s a result of a four-fold greater chance of being infected, and then roughly another four-fold greater case mortality rate.
Source: CDC COVID data tracker.
If I divide that by the fraction of the population (over age 12) that was vaccinated and not, at that time, I end up with the following graph of fraction of COVID-19 deaths at the peak of the Delta wave, by vaccination status:
So, just to be clear, it’s not so much that COVID-19 had gotten tamer, up to the Delta variant. It’s that we’d gotten a lot better in dealing with it.
Once vaccines came into use, this became mostly a pandemic of the un-vaccinated.
In reality, those proportions shown above are due to more than just the pure effect of vaccination. I would judge that, after the CDC age adjustment (accounting for differences in age between the vaccinated and un-vaccinated groups), what you’re looking at above is mostly the impact of vaccination. But it’s amped up by all the other differences in behavior — such as COVID-19 hygiene and willingness to take risks — between the vaccinated and un-vaccinated populations.
To round this out, let me show the CDC’s estimate of the impact of the booster dose. This is from the end of last year, so this is showing effectiveness against Delta. As with the numbers above, these are age-adjusted for persons age 12 and older.
Source: CDC COVID data tracker.
As with the first chart, that likely somewhat exaggerates the true impact of the vaccine. Plausibly, those getting the booster dose also behave differently from the rest of the population. But also as plausibly, some part of the difference between this chart and the last CDC chart is the impact of the booster itself.
The CDC’s data don’t extend into the Omicron period, but Virginia continues to show the vaccinated/unvaccinated comparison on a current basis.
Undoubtedly some portion of that is the impact of vaccines, some portion is behavioral, (and in this case, some portion may be due to differences in age and other demographics). But no matter how you slice it, even with a far-less-virulent Omicron variant, this remains mostly a pandemic of the unvaccinated.
A quick recap, or the tyranny of big numbers.
This section assumes that the reader is vaccinated and boosted.
Let me just briefly recap the calculations of Post #1400, part 4.
In a typical flu season, new flu infections occur at a rate of about 49 / 100K / day. Beyond infections, flu has reasonably well-known case hospitalization and mortality rate (that is, estimated hospitalizations and deaths per symptomatic case). Starting from that benchmark, what incidence rate of new Omicron cases generates the same level of health risks as flu?
You have to factor in two offsetting effects, for the vaccinated and boostered population, to compare Omicron to flu. Vaccine and booster are more effective against Omicron than vaccine is against flu. Almost twice as effective. But Omicron is more likely to hospitalize or kill you, if you get infected. Just over twice as likely.
When I work through the math, the vaccinated and boostered person faces equal risk from Omicron and from seasonal flu once the case rate in his or her area (for all persons) gets down to 40 new Omicron cases / 100K / day.
That result — 40 new cases / 100K / day — looks like a large number, by historical standards. E.g., when I last returned to the gym, Virginia’s new case counts were in the single digits.
But there’s a reason that the numbers have shifted. And you need to shift your perception of them. What looks like a high new case rate, by historical standards, no longer constitutes a high risk rate, for endemic Omicron. At least, not for the vaccinated-and-boostered.
The first reason is the virulence and incidence of Omicron. Omicron is much less virulent than prior strains, and the un-vaccinated account for most of the new cases. So for the vaccinated-and-boostered person, the health risks faced at 40 Omicron cases / 100K / day are not hugely different from (e.g.) high single digit rates of Delta.
Second, once you understand that Omicron is going to be endemic, you have to get your mind around some unavoidable risk, if you are going to get on with your life. And so, at 40 cases / 100K / day, you are taking some risk. In this case, the vaccinated-and-boostered person is taking roughly the same risk of hospitalization and death from Omicron as he or she would face from flu, during a typical week of flu season.
My point is, don’t let the big numbers fool you. You can’t directly compare Omicron daily case counts to prior strains. And, psychologically, you need to get out of the mind-set of avoiding risks, and into the mindset of determining what’s a reasonable risk.
Reduced risk of hospitalization and death.
In the calculations above, I made one extremely conservative assumption about the effectiveness of vaccine and booster against Omicron. I gave vaccination and booster no credit whatsoever for reducing case hospitalization rates and case mortality rates. That is, I assume that they reduced the rate of infection, only. And that they had no further effect on reducing the odds of hospitalization or death, once infected.
To be clear, that is not the mainstream consensus. Pretty much from Day One, health authorities have said that vaccines work better at avoiding serious illness than they do at merely avoiding any infection. Further, that is typical of vaccines for other diseases (e.g., diseases of childhood, flu, and so on). Breakthrough cases tend to be milder than cases in the unvaccinated.
The problem is that, to me, the data on hospitalization and death appeared confusing. First, hospitalization and (particularly) death are such rare outcomes that the clinical trials data often don’t have enough cases to give a precise answer in that area. So we lack the hard numbers. And then, reliance on observational data means that you end up looking not just at the effect of vaccination, but also at any other differences between the vaccinated and un-vaccinated populations. As as result, the estimated impact on (e.g.) deaths seemed exaggerated.
In this section, I just want to emphasize how much I may have grossly overstated actual risks by making that assumption. In other words, at 40 Omicron cases / 100K / day, the actual risk of serious illness faced by the vaccinated and boostered individual may be vastly less than the same risks imposed by seasonal flu.
First, let me return to the CDC data on death rates. Whatever is causing the vastly lower death rate (under Delta, for the fully-vaccinated population), it’s pretty consistent across age groups. These mortality curve look pretty much the same for all age groups.
Source: CDC COVID data tracker.
That suggests that the mortality effect is not primarily driven by differences in behavior. I doubt, for example, that the 65+ population is hitting the singles bars to the same extent as the age 18-29 population. What’s constant across these graphs is the disease and the vaccination, but not the behavior.
Second, other countries show the same huge impact on mortality rates in their own observational data. A reader pointed me to a recent weekly report out of Great Britain, where they compare the boostered population to the un-vaccinated population, by age:
Source: UK Health Security Agency, COVID-19 vaccine surveillance report, Week 3, 20 January 2022.
Their observational data show on-order-of ten-fold differences in the mortality rates, across all age groups, despite only trivial differences in reported infection rates. In other words, they show vastly different case mortality rates based on vaccination status.
Finally, there are good first-principles reason to think that vaccination would reduce the number of severe cases. Even if Omicron is able to avoid existing anti-COVID antibodies, other parts of the immune system would remain primed to fight COVID. The (slower) action of these other immune reactions might not prevent any infection, but plausibly would prevent the most severe infections.
The upshot is that at 40 new Omicron cases per 100K per day, the risk of severe disease from Omicron may be substantially less than the same risk from typical seasonal flu, for the vaccinated and boostered population.
This just reinforces the main point, though. If you’re vaccinated and boostered, and if you don’t worry about the risks of being out-and-about during a typical flu season, then you really shouldn’t give Omicron a second thought, once the new case rates drop below 40 per 100K per day, or so.
At that level, if you are vaccinated and boostered, your risk of any infection with Omicron is under half the risk of picking up a case of flu (in a typical year). And your risk of a severe infection — risk of being hospitalized or dying — is no higher than that from flu, and might be as low as one-tenth the risk you face from flu, depending on whose data you happen to believe.
I’ve reached the point where I’ve fully grasped this, rationally. It’s still going to take a while to shed some habits and reactions developed during the pandemic.
