Coronavirus August 2021—Part 14 The Delta Variant, Masks, and Vaccines

In this post, we answer your questions about the effectiveness of the vaccines against the Delta variant of SARS-CoV-2 and the implications for our behavior if we’ve been fully vaccinated. As usual, if you’re less interested in how we got to our conclusions than you are in the conclusions themselves, feel free to skip to the BOTTOM LINE in each section and the CONCLUSION at the end.

our thinking is the problem

The problem we’re all having in trying to sort out our approach to the COVID-19 pandemic and in figuring out how much we should worry about it is that our brains are wired to think anecdotally rather than statistically. This leads us to fear things we shouldn’t fear, or to fear them out of proportion to the real risk of danger they pose to us, and to dismiss threats we should fear, or to dismiss them too offhandedly. As a result, we aren’t making choices that maximally reduce our risk.

We may hear that a friend who’s been vaccinated has had a breakthrough infection. Then we read a headline that highlights the rate of breakthrough infections in vaccinated people without mentioning the denominator—the number of vaccinated people who haven’t had a breakthrough infection—and we conclude that breakthrough infections in vaccinated people are far more common than they actually are or, worse, that the vaccine doesn’t work well enough to warrant our getting it. We may also conclude that the risk of becoming significantly ill if we’re immunized is far higher than it actually is. Or that the risk of becoming infected and/or significantly ill if we’re not immunized is lower than it actually is. Or we read about rare complications of vaccination and conclude that the vaccines aren’t safe without comparing the risks of those complications to the risks we take every day without even thinking about it.

So what do the science and statistics tell us about the real risks and benefits of contracting COVID-19 and of being immunized against it?


The Delta variant (B.1.617.2) is now the most widespread variant in the U.S, if not the world. As of this writing, if you become infected with COVID-19 in the U.S. now, you have a 93.4 percent chance of being infected with the Delta variant.

Is the Delta variant more contagious than other variants? It seems to be so. First, the fact that it’s become the dominant strain argues for increased levels of infectiousness. Second, secondary attack rates of infection for the Delta variant (meaning the risk of index cases passing it on to household members) in one study was 53 percent (in which only 9 percent were fully vaccinated), significantly higher than the 17 percent secondary attack rates seen on average for other COVID variants. Third, the viral loads in patients infected with the Delta variant are 1,260 times higher than those in patients infected with non-Delta variants when the virus is first able to be detected.

Is it as contagious as chickenpox, as the CDC has claimed? That’s not so certain. The degree of infectiousness of an infectious disease is reflected in its RØ value (pronounced “r-naught”), a number that indicates how many other people a single infected person is likely to infect in a population with no immunity. (So an RØ of 1 means an infected person is likely to infect only one other person while an RØ of 10 means an infected person is likely to infect 10 other people.) It’s important to note that the RØ value is used mostly to calculate the proportion of people in a population who need to be immune for the population to achieve herd immunity. It doesn’t necessarily tell us about our own particular risk of contracting the disease in question because RØ is affected by the contact rate of the infection, which refers to how many people a person with the disease can be expected to come into contact with. This variable is not specific to a disease but instead is impacted by a number of factors, including location and public health measures in place, such as quarantines or travel bans, as well as the behavior of individuals (social distancing, mask-wearing, and so on). Thus, this factor is considerably modifiable. The CDC argues the Delta variant has an RØ of 5 – 9.5 (similar to chickenpox) because of an outbreak in Provincetown, Massachusetts. However, during that outbreak, people were coming in contact with each other intimately to a degree not currently mimicked in the general population, raising the contact rate, which raised the estimation of RØ in that population. In another study in the U.K., for example, the RØ for the Delta variant (technically the Rt, which is the effective reproduction number in a population at a given time irrespective of the population’s immune status) was calculated to be only 1.19 in a population that had significant immunity due to vaccination.

BOTTOM LINE: RØ is extremely difficult to calculate as it’s not only dependent on the intrinsic infectiousness of an infectious agent but also on the behavior of the population in which it’s being calculated. While the Delta variant is more contagious than other variants, as immunity in the population increases, both through natural infection (which is currently accelerating in the unvaccinated because Delta is more contagious) and immunization, at some point, the prevalence of active cases will begin to go down, reducing the risk of coming in contact with SARS-CoV-2 and therefore of becoming infected with it.

Is the Delta variant more likely to cause serious illness? Again, it may be. An observational study from Scotland suggested that patients infected with the Delta variant were 85 percent more likely to be admitted to the hospital compared to patients infected with other COVID variants.

How Effective are the Vaccines Against the Delta Variant?

Studies show that the vaccines remain extremely effective against the Delta variant. A study out of Singapore looked at the severity of infection with the Delta variant in immunized subjects compared to the severity of infection in non-immunized subjects. All people in Singapore who are diagnosed with COVID-19 are hospitalized for evaluation and isolation—including asymptomatic cases—so this study was able to provide a complete dataset for analysis of how well the vaccines (in this study, the mRNA vaccines) protect against all Delta variant infections, symptomatic or not, as well as the severity of symptomatic infection. In this retrospective study, people who’d been immunized were quite a bit older (average age=56) compared to people who hadn’t been immunized (average age=39.5), so we would have expected more severe cases of COVID-19 in the immunized group solely on the basis of age. However, those subjects who were immunized had a 92.7 percent lower chance of developing severe disease (defined as cases that required supplemental oxygen) than those who weren’t immunized. The other finding of the study was that the initial concentration of viral RNA detected in the noses of vaccinated cases compared to unvaccinated cases were nearly identical, but that the viral RNA concentration dropped off more rapidly in subsequent days in the vaccinated patients than in the unvaccinated patients, suggesting that infections that occur in vaccinated people may have an overall reduced likelihood of being transmitted to others.

Other studies also show that the vaccines provide significant protection against symptomatic and severe illness from the Delta variant, albeit to different degrees. In the study from the U.K. (noted above), vaccine effectiveness was estimated to be 62 percent in preventing symptomatic infection with the Delta variant. In another study, the Pfizer vaccine was found to be 88 percent effective in preventing symptomatic infection with the Delta variant. The observational study from Scotland (noted above) showed that for people vaccinated with the Pfizer vaccine who have breakthrough infections with the Delta variant, the risk of hospitalization was reduced by 62 percent. (Notably, the vaccine decreased the risk of infection in the first place by 92 percent for non-Delta variant COVID and by 79 percent for Delta variant COVID.) Finally, in a study from Qatar, the effectiveness of Pfizer’s vaccine against severe, critical, or fatal disease due to infection with the Delta variant was 97.4 percent (similar to the study from Singapore noted above).

It’s important to note that vaccine effectiveness is a relative measure that compares the risk of infection between vaccinated and unvaccinated people. But the risk for unvaccinated people decreases as the pandemic progresses because as the pandemic progresses the fraction of unvaccinated people who develop natural immunity increases. This then will cause an apparent drop in vaccine effectiveness over time, even if the vaccine keeps working as well as ever. Imagine a vaccine that has an efficacy of 90 percent when no one in the population has natural immunity. If half the unvaccinated population develops natural immunity, the apparent vaccine effectiveness, [(rate among unvaccinated – rate among vaccinated) / rate among the unvaccinated x 100], will drop to 80 percent, even though it’s just as effective at preventing infection as it was in the beginning.

Finally, to remind you just how amazing the protection from the vaccines still is against even the Delta variant, recall that to calculate a person’s absolute risk of getting infected after having been vaccinated, you have to start with the base rate of infection, which varies in different contexts and in different populations. The highest base rate of infection we’ve seen in published contact tracing studies (before the emergence of the Delta variant) was around 40 percent (for the spouses of infected people). If, as some studies argue, the Delta variant is on average 50 percent more infectious than previous variants, that absolute risk of infection from an infected person to his or her spouse would be 60 percent with Delta (interestingly similar to the 53 percent risk of a household contact becoming infected with Delta according to the contact tracing study mentioned above). This means that in the context that puts people at the highest risk for contracting COVID-19 (having an infected spouse), the risk of becoming symptomatically infected for an immunized spouse would be, according to the studies noted above, anywhere from 62 to 88 percent less than 60 percent, or 7.2 to 28.8 percent. This, then, would represent the maximal absolute risk of an immunized person contracting COVID-19 in any one context. (Realize also that this would be the risk only if your spouse is indeed infected. If your spouse isn’t infected, your risk of contracting it from them is zero. What’s the risk of your spouse being infected? That depends on the prevalence of infection in your community, your spouse’s behavior, and your spouse’s immunization status.) But here’s the bottom line: the absolute risk of becoming infected to which vaccinated people are being exposed in most situations in which they find themselves will be far less than 7.2 to 28.8 percent.

This does explain, however, why breakthrough infections with Delta can and do occur. But what we care about most—and what the vaccines were really designed to mitigate—isn’t the risk of catching COVID-19. It’s the risk of being hospitalized and dying from it (as well as the risk of developing long-COVID). Here, the CDC data tells the real story: as of this writing (at a time when, as mentioned above, the Delta variant is the dominant strain infecting people in the U.S.), of 164 million people fully vaccinated (with a mix of the mRNA vaccines and the J&J vaccine), 5,285 people have been hospitalized for COVID-19 (which yields a risk of being hospitalized from severe COVID-19 if you’re immunized of 0.003 percent), and of those 1,191 died (which yields a risk of dying from COVID-19 if you’re immunized of 0.0007 percent). When you consider the risks most of us take every day without worrying about them at all—for example, over the course of a year, the odds of getting into a car accident are 3.7 percent on average and the odds of dying in a car accident are 0.3 percent, making the annual risk of dying from a car accident 0.01 percent, which is 14 times the risk of an immunized person dying from COVID-19—our inability to think statistically clearly has us afraid of the wrong things. (This goes for the decision to be vaccinated as well: our annual risk of dying from a car accident turns out also to be 14 times the risk of the most common serious adverse reaction to the vaccines—blood clots with the J&J vaccines—which occurs at the same rate as the rate of death from COVID-19 if you’re fully immunized, a rate of 0.0007 percent.)

What’s the risk of developing long-COVID if you have a breakthrough infection after being vaccinated? It’s not yet clear. One paper from Israel found that “at 6 weeks after their diagnosis, 19 percent [of patients with breakthrough COVID infections] reported having ‘long Covid-19’ symptoms, which included a prolonged loss of smell, persistent cough, fatigue, weakness, [shortness of breath], or [muscle aches],” but the sample size was quite small (n=39).

BOTTOM LINE: The Delta variant is more infectious than previous variants and may be more likely to cause serious illness. Even though the exact effectiveness of the vaccines in preventing infection and in preventing serious illness, hospitalization, and death isn’t certain, the vaccines still reduce the absolute risk of contracting the Delta variant of COVID-19 significantly. Further and most importantly, they reduce the risk of being hospitalized for COVID-19 and dying from COVID-19 to such a great extent as to be almost zero. Yes, you can still get COVID-19 if you’re vaccinated, but if this happens, you’re overwhelmingly likely to have a mild infection only. More work needs to be done to accurately assess the risk of long-COVID after vaccine breakthrough infections.

Waning Immunity

Is the effectiveness of the COVID-19 vaccines waning with time? With the introduction of the Delta variant, it’s difficult to answer this question. It’s more likely that the mildly reduced effectiveness of the vaccines against Delta is due to Delta itself, not waning immunity in general. The only way to accurately answer this question, however, is to continue to compare rates of infection, hospitalization, and death from COVID-19 between vaccinated and unvaccinated populations. This is the gold standard for determining whether immunity from the vaccines wanes and, if it does, over what period of time.

We have data that shows the mRNA vaccines retain their effectiveness for at least six months. We know antibodies formed by the Moderna vaccine last at least six months, and Moderna just put out a press release that claims their vaccine has retained a 93 percent effectiveness six months out from immunization (study not yet published). Note that these are two different things: antibodies could wane to undetectable levels and the vaccines could still retain their effectiveness because memory B cells will quickly ramp up the production of antibody levels when an immunized person is exposed to the virus. Thus, we can’t look at waning antibody levels as a surrogate for vaccine effectiveness in the real world.

Third Vaccine

Because data from the CDC show that overall rates of death from COVID-19 in vaccinated people are nearly zero as noted above, we don’t think third booster shots make sense at this point for most people. Waning antibody levels, as also noted above, can’t be used as a marker of vaccine effectiveness.

In certain immunocompromised patients, however, initial antibody response seems to be lower—especially in patients who’ve undergone solid organ transplants. (Interestingly, Moderna outperforms Pfizer in both dialysis patients and solid-organ transplant patients. After two shots, antibody generation in dialysis patients for Moderna was 97 percent compared to 88 percent for Pfizer; in solid organ transplant patients, 49 percent for Moderna compared to 26 percent for Pfizer.) Dialysis patients who remained without any antibody response after the second Pfizer shot who were then given a third shot, however, did develop an antibody response at a rate of 88 percent.

BOTTOM LINE: As the presence of high antibody levels—whether persisting from the original vaccination or from a rapidly-deployed memory B cell response when an immunized person is exposed to the virus—does correlate to immunity, in certain immunocompromised individuals, a third booster shot may make sense.


To what degree are fully vaccinated people able to spread the Delta variant? It remains unclear. First, because the rate of COVID-19 infection in vaccinated people is so much lower than in unvaccinated people, the likelihood of encountering an infected vaccinated person is low to begin with. But if a vaccinated person becomes infected with the Delta variant, how likely are they to transmit it to others? In the Singapore study mentioned above, initial viral loads seemed identical in symptomatic vaccinated and unvaccinated people who were infected with the Delta variant. However, in the vaccinated people, the concentration of viral RNA fell off more quickly than in the unvaccinated people. Also, in the U.K. study mentioned above, the viral loads were also found to be lower in vaccinated people compared to unvaccinated people. Thus, at this point, it’s likely that the risk of an infected vaccinated person spreading the virus to others is lower than the risk of an infected unvaccinated person. Yet more definitive studies need to be done.

In the meantime, how should we behave? It’s not clear. A reasonable approach would be for a fully vaccinated person to mask up when in close contact with a vulnerable person (for example, an elderly person with multiple risk factors for dying from COVID-19 or someone who’s immunocompromised), especially if in an area where the prevalence of cases is high. Unfortunately, we don’t have good data to tell us how high the prevalence has to be for masking of asymptomatic fully-vaccinated people to have an impact on community spread. Importantly, if a fully-vaccinated person develops a symptomatic COVID-19 infection, given that we don’t have data on how infectious they may be, they should consider themselves infectious for up to 10 days and wear a mask when around anyone.


Given that airlines and airports are continuing to enforce mask-wearing rules, for fully vaccinated people, travel remains, in our view, reasonably safe.

CONCLUSION: It’s hard to know how to think about immunization, the Delta variant, and how we should behave in different circumstances to keep ourselves and those around us safe. We’re all seeing the science unfold in real time, revealing just how messy, uncertain, and difficult it is to figure out what’s really true. But, though it takes time, science ultimately gives us answers we can rely on. We can all argue about what policies make the most sense based on what the science shows, but it’s the science we should all use to help us guide our own behavior. And, as of this writing, the science says the following:

  1. The Delta variant is more contagious than other variants.
  2. The Delta variant may be more dangerous than other variants.
  3. The vaccines are likely somewhat less effective in preventing infection with the Delta variant, but still offer an enormous amount of protection. Breakthrough infections are occurring, but they are overwhelmingly mild.
  4. Vaccinated people probably can transmit the infection but almost certainly at a lower rate than unvaccinated people.
  5. The vaccines remain unbelievably effective at preventing hospitalization and death from the Delta variant, so much so that vaccinated people can continue to live as they did before the onset of the pandemic, with the possible exception of wearing masks to prevent asymptomatic spread to vulnerable people in areas of high prevalence of disease.
  6. A third booster shot for non-immunocompromised people doesn’t make sense at this point in the pandemic. Some people who are immunocompromised may want to consider a third shot.


  1. Coronavirus February 2020—Part 1 What We Know So Far
  2. Coronavirus March 2020—Part 2 Measures to Protect Yourself
  3. Supporting Employee Health During the Coronavirus Pandemic
  4. Coronavirus March 2020—Part 3 Symptoms and Risks
  5. Coronavirus March 2020—Part 4 The Truth about Hydroxychloroquine
  6. Coronavirus April 2020—Part 5 The Real Risk of Death
  7. Coronavirus April 2020—Part 6 Evaluating Diagnostic Tests
  8. Coronavirus April 2020—Part 7 The Accuracy of Our Antibody Test
  9. Coronavirus May 2020—Part 8 How to Reopen a Business Safely
  10. Coronavirus August 2020—Part 9 Masks, Vaccines, and Rapid Testing
  11. Coronavirus December 2020—Part 10 Should You Get the Pfizer Vaccine?
  12. Coronavirus December 2020—Part 11 Should You Get the Moderna Vaccine?
  13. Coronavirus April 2021—Part 12 Should You Get the Johnson & Johnson Vaccine?
  14. Coronavirus May 2021—Part 13 How Effective are the Vaccines in the Real World?

Sign up to get notified when a new blog post has been published.

Leave a Reply

  • Thanks for this, Alex!

    Whenever I hear people say they’re anti-vax because it’s “too early to tell whether vaccines are safe,” I point them to your analyses. What’s the alternative? Trying to convince them people are, indeed, dying from COVID?

  • I am confused about seemingly changing advice regarding vaccine dose spacing.

    Six months ago, people were explicitly told not to get any other vaccine (flu, Shingrix, whatever) within two weeks of a COVID-19 vaccine dose. Now, publicity is encouraging folks to get a flu shot at the same time that they get their COVID-19 vaccine shot or booster. What is your perspective on timing for a patient planning to both get a flu shot and a COVID-19 vaccine dose?

    Linda: The answer is the original advice was given out of an abundance of caution and lack of data. Since then, we have overwhelming evidence that the COVID-19 vaccines are safe. Other vaccines are known to be safe when given with other vaccines. The updated guidance is based on this and the safety data that’s accumulated with millions of people receiving COVID-19 vaccines.