COVID-19 Updates: Part 1
This newsletter is the first of a two-part series on COVID-related stats that have emerged in the past 10 days. (Thanks to Imani Parton and Amy Richardson for suggesting this topic.)
My focus is on vaccine/booster rates and effectiveness; omicron is covered at the end. Next week I'll talk about new COVID-19 vaccines and treatments currently being developed.
There’s a lot of new data, so I've used a FAQ format to keep things organized.
I: Vaccine/Booster Rates
How many Americans are fully vaccinated?
Just over 200 million, as of yesterday. In other words, 60.4% of all Americans. The percentage is higher for people 18 and older, but either way the stats aren’t terribly impressive, as they put us in 18th place among countries for which data are available.
How much variability is there in vaccination rates across states?
A lot. As of two days ago, the lowest rate is 45.6% (Idaho), while the highest is 74.4% (Vermont). In other words, depending on the state, anywhere from about a quarter to just over half of Americans are not yet fully vaccinated.
(States with the highest vaccination rates (Vermont, Rhode Island, Maine, Connecticut, Massachusetts – all above 70%) would rank between 14th and 18th if compared to other countries. The states closest to the world average are Arizona, South Dakota, and Texas, each currently at 55.5%.)
Why aren't more people vaccinated?
They don't want to be. In the past week, a methodologically rigorous Kaiser Family Foundation survey, as well as the slightly messier survey aggregator Morning Consult, have shown that 14% to 18% of American adults are unwilling to be vaccinated. By “unwilling” I mean that they're choosing the most extreme options on the surveys ("definitely not" or "unwilling"). Since more than 18% of adult Americans are currently unvaccinated, you might ask: Who are the folks who aren't vaccinated, but aren't choosing the most extreme survey options? In short, they seem to be the ones who are "uncertain" (Morning Consult) or say they’ll get vaccinated "only if required" (Kaiser).
Why doesn't everyone want to be vaccinated?
It's complicated.
News media often call out demographic categories. For example, the New York Times commented yesterday that those who adamantly refuse vaccinations "include a mix of people but tend to be disproportionately white, rural, evangelical Christian and politically conservative". This is true, statistically speaking, but only because "tend to be" is a vague phrase. Group differences are small, and even additive effects are less than you might expect (e.g., the difference in vaccination rates between white, rural, evangelical Christians versus non-white, urban residents who aren't evangelicals doesn't usually exceed 10%).
In my view, linking vaccine refusal to demographic categories yields diminishing returns, once you know a person's self-identified political affiliation (Democrat, Republican, or Independent). In the Kaiser survey, for instance, 26% of self-identified Republicans indicate that they will "definitely not" get vaccinated, as opposed to only 2% of Democrats.
Surveys that explore specific reasons for vaccine refusal consistently identify safety concerns as the most common reason. In the Kaiser survey, 71% of respondents (mostly Republican) noted that they’re "not too confident" or "not at all confident" that COVID-19 vaccines are safe. Morning Consult data show that safety concerns consist of fears about potential side effects, as well as a more indirect concern that the vaccines have been developed too quickly. In a Times survey conducted back in July, 93% of unvaccinated respondents cited specific safety concerns (50%), but almost as many (43%) indicated a desire to just wait and see whether vaccines are safe.
Survey stats are limited in how much they can tell us about peoples' thought processes. The surveys themselves illustrate these limitations. For example, Kaiser found that among a "wait and see" group, 46% said that they'd be more likely to get vaccinated if a vaccine were available from their personal physician. So, are these folks waiting for more evidence of safety, or more convenience, or both...?
How many Americans have received a booster?
Just over 47 million, as of yesterday. In other words, nearly 24% of already vaccinated Americans. Fears about omicron, and plans for holiday travel, have boosted the booster rate – since last week, almost a million Americans per day have gotten a booster.
Do all fully vaccinated people intend to get a booster?
No. The Kaiser survey found that 18% of respondents who are already fully vaccinated say they "probably" or "definitely" won't get boosted. Many causes of reluctance have been noted. Here, the "wait and see" group is especially large, because folks aren't just waiting on safety data; they also want to know more about the omicron variant. For instance, some people are waiting for a booster that's tailored to omicron, under the assumption that it would offer more protection than current vaccines. (More on omicron at the end of this newsletter.)
II: Vaccine/Booster Effectiveness
Are vaccines still effective at preventing infection?
Yes. Real-world data initially showed that the Pfizer and Moderna vaccines are about 90% effective at preventing infection. That is, receiving a full dose of the Pfizer or Moderna vaccine reduced the chances of COVID-19 infection by about 90%. (The initial effectiveness for Janssen was estimated at just under 70%.)
We know now that vaccine effectiveness diminishes over time. Exactly how much is hard to say. It hasn't been easy for researchers to disentangle natural changes in immune system response (i.e., declining levels of neutralizing antibodies) from the emergence of a more virulent strain of COVID-19 in the summer of 2021 (i.e., delta). In any case, all reputable studies thus far show that the vaccines remain effective, even though estimates of effectiveness vary. The worst news I've seen is a study conducted in Israel showing that Pfizer's effectiveness at preventing infection dropped to 39% after 7 weeks. However, Pfizer’s effectiveness against severe disease remained high (91%). And, the study is somewhat of an outlier. Most research has provided better news. For example, a multinational study of 44,165 Pfizer recipients (and 2264 adults who received a placebo) showed that after 6 months, effectiveness against infection was 92.6% in the U.S., and over 86% for all countries studied.
Why are there inconsistencies in the infection data?
Many reasons. For example, here are four reasons why we expect inconsistencies (and might even feel suspicious if they weren't observed):
—Sampling error is routine. Infection rates for two random samples of 10,000 Texans would almost surely differ (albeit slightly) even if the samples were obtained on the same day.
—Studies have been conducted before, during, and after the ascendance of the delta variant, which is known to be more virulent than earlier strains. Infection rates should thus differ across otherwise identical studies conducted within each time period.
—Studies have differed in how long people have been tracked following vaccination. Any difference in this variable almost guarantees inconsistent rates of infection.
—Studies differ in methodological details (e.g., in some studies delta is confirmed via assay, while in others it's assumed because it was the dominant strain when the study was conducted). Here's a study, updated today, illustrating the role of sampling: During a delta outbreak in a Jamestown, California prison, Moderna's effectiveness at preventing infections was found to be 56.6%, which is considerably less than earlier estimates for the general population. However, one would expect less effectiveness in a prison population, where people live in close quarters (and where COVID-19 is five times more prevalent than in the general population). And, many of those infected were asymptomatic; Moderna's effectiveness against symptomatic infection was 84.2%.
Inconsistencies in effectiveness stats shouldn’t distract from the observation that all studies have documented significant effectiveness. Full vaccination clearly makes you safer. Exactly how much safer is impossible to quantify, owing to the differences among studies mentioned here, and to the fact that your actual risk is determined in part by your behavior.
Are vaccines still effective at preventing severe illness?
Yes. Both prospective and retrospective studies have shown that two doses of an mRNA vaccine (Pfizer or Moderna) still confer 90%+ effectiveness against severe illness (with Janssen closer to 70%), even now that delta is the dominant strain. Specific estimates of effectiveness vary from study to study, but much less so than for infection rates (see here and here).
If I'm already fully vaccinated, what are my chances of a breakthrough infection?
Very small. For example, a Washington State Department of Health report, released on December 1, examined over 4 million COVID-19 cases documented between January 17 through November 20 and found a breakthrough rate among fully vaccinated adults of about 1 in 5,000. Nationally, CDC data show a comparable pattern, with breakthrough rates closer to 1 in 10,000, depending on month of measurement.
In sum, if you're fully vaccinated, your risk of a breakthrough infection is currently low. The data I just cited spans the time period when delta, the most contagious strain of COVID-19, became predominant. (By the end of June, 82.2% of new cases were caused by delta; this number increased to nearly 100% by the end of July and remains at that level today.)
If I'm already fully vaccinated, what are my chances of a severe breakthrough infection?
Extremely small. The Washington State data show that about 1 in 12.5 breakthrough infections led to hospitalization. Since the breakthrough rate was 1 in 5,000 to begin with, we can extrapolate that among vaccinated people, only about 1 in 62,500 developed a COVID-19 infection that required hospitalization. This figure is fairly consistent with CDC data showing that since January, between 1 and 2 out of every 100,000 vaccinated people are hospitalized for COVID-19.
CDC data also show that monthly hospitalization rates for unvaccinated people have always been at least six times higher than for those who are vaccinated. Near the end of July, hospitalization rates began to climb in both groups, but since that time, the rates have never exceeded 5 in 100,000 vaccinated people, while the rates for unvaccinated people have been at least nine times higher.
In short, if you're vaccinated, the risk of contracting a COVID-19 infection that leads to hospitalization is close to zero. If you're unvaccinated, the risk is higher and increasing.
Do boosters confer additional protection against infection and severe illness?
Yes. Although it's early, studies are beginning to show that boosters not only stimulate immune response but also translate into increased protection. As with data on primary vaccinations, results differ from study to study (see Appendix for more details).
The CDC, Dr. Anthony Fauci, and other experts recommend boosters for all adults (6 months after your second mRNA vaccine, or 2 months after Janssen). Previously, the CDC did not recommend boosters for children 17 and under, simply because it was too soon to boost, but as of today the agency recommends boosters for 16- and 17-year-olds, following the FDA's approval for the use of Pfizer boosters with this age group.
Do my boosters need to be the same type as my primary vaccine?
No. Heterologous boosters (those that differ from the primary vaccine - i.e., your first two doses) increase antibodies to levels consistent with >90% effectiveness against delta.
There are two, slightly different "official" views on this topic. One view, represented by the CDC, holds that it doesn't matter whether your booster is heterologous or homologous, because effectiveness and safety are comparable in each case. The other view, represented by the European Union, is that homologous boosters are fine, but heterologous boosting might actually increase your protection slightly if the booster is an mRNA vaccine (Pfizer and Moderna), though not if it’s a vector vaccine (e.g., Janssen).
Neither view should be considered definitive yet, because the research is new, and there haven't been large, systematic comparisons of different mix-and-match options. For instance, one carefully controlled study examining all possible combinations of Pfizer, Moderna, and Janssen showed strong gains in antibody response and virtually no side effects, with some advantage of the mRNA boosters over Janssen, but the sample size was small (458), participants weren’t randomly assigned to booster groups, time from vaccination to boost wasn’t controlled, and participants were only followed for 28 days. (Small sample size is especially problematic for a study like this because there were 9 groups of interest (three primary vaccines x 3 boosters) and therefore barely more than 50 participants per group.) Other studies, including one published last week, also suggest that mRNA boosters are best, but they still indicate that all types available in the US are useful.
III: Omicron
Is omicron more infectious than other variants?
Unknown. Although it spreads "rapidly", as Anthony Fauci and other experts have noted, the speed/ease of transmission relative to other strains is still unclear.
Is omicron more dangerous than other variants?
Unknown, although data released this week have already begun to suggest two themes.
First, compared to earlier strains, omicron may tend to cause less severe symptoms. This was first reported anecdotally by researchers in South Africa. Two days ago, Anthony Fauci indicated a more favorable ratio of infections to hospitalizations for omicron than for delta in several cohorts currently being studied (but for which data are not yet publicly available). That’s encouraging news.
Second, re-infection rates may be higher for omicron than for earlier strains. Dr. Fauci made this suggestion on the basis of unpublished data from several countries. In one of the few publicly available reports on this topic, a South African study suggests that prior infection with an earlier strain provides much less protection against reinfection with omicron, but the study hasn’t been peer-reviewed yet, and the authors note reliability-related limitations (e.g., differences in testing protocols across sites) that I find problematic. In any case, most of us rely on vaccination and safe behavior rather than prior infection to protect us against future infection.
How effective are current vaccines against omicron?
Unknown, although Pfizer reported this week that while a full course of their vaccine is minimally helpful, the addition of a booster produces neutralizing antibodies to omicron at levels comparable to those seen when people who received a full course were exposed to earlier variants. In other words (although Pfizer is cautious not to say this directly), three doses of Pfizer may turn out to be roughly 90% effective at preventing omicron infections. That would be good news.
Next week I should have more information to share about this report. I would caution against making too much of it yet, because (a) it's a summary prepared for the media; Pfizer hasn’t released any raw data yet. (b) As I pointed out in an earlier newsletter, Pfizer aggressively promotes its vaccines, arguably to the point of cutting corners. (c) Although Pfizer's report is consistent with an unpublished study conducted in South Africa, the latter was announced on Twitter this week and as of today is only available on researcher Alex Sigal’s website.
What’s the bottom line for omicron?
As of today, we don’t know how infectious it is, and infection from some earlier strain might not provide much protection against an omicron re-infection. However, omicron infections may produce milder symptoms than earlier strains did, and full vaccination plus a booster may offer as much protection against an omicron infection as it does against infection from delta and other earlier strains.
IV: Conclusions
Get vaccinated, get a booster, be careful, and stay tuned!
Appendix: Booster effectiveness data
Booster effectiveness has been demonstrated primarily through studies of immune system response; the results of these studies indicate protection equal to or greater than that conferred by the primary course of each vaccine.
As for real world data, it's too soon to know anything about long-term effectiveness, but the short-term results thus far are promising. For example, a study on adults over 50 in the UK showed that following a primary course of Pfizer or AstraZeneca, the Pfizer booster was 93% to 94% effective over a two week period. Comparable results were obtained in a study on adults over 60 in Israel. No substantive real-world data are available for Moderna, but its short-term effectiveness as a booster shouldn’t differ much from Pfizer’s, given that these boosters have similar effects on antibody levels, and that the performance of each as a primary course (two doses) has been roughly comparable in real world effectiveness studies.