Oncologist Dr. Vinay Prasad is at it yet again in a recent article published in City Journal, promoting himself as the scientist-physician who shines a light on government incompetence and corruption as he leads us to a post-COVID promised land. City Journal, I discovered after writing this post, appears to be a conservative publication which “holds itself to the highest intellectual, journalistic, and literary standards, aiming to produce absorbing reading for intelligent and discerning readers.” I urge them to keep trying.

Prasad’s article (which is representative of his other articles found if you Google “Vinay Prasad masks” or “Vinay Prasad vaccine”) is titled “Public Health’s Truth Problem”, and the first paragraph outlines his over-arching theme that public health organizations are continually spreading “falsehoods” and “omitting uncomfortable truths” with long-term consequences to these institutions. He then starts several paragraphs with an italicized statement made by the CDC or FDA which he presents as a falsehood. Let’s review his assertions and whether they hold up.

Any mask is better than no mask

Last week, CDC director Rochelle Walensky asserted that any mask is better than no masks at all. This statement was factually incorrect when she said it. The only published cluster randomized trial of community cloth masking during Covid-19—performed in rural Bangladesh—found that surgical masks reduced the spread of Covid-19 among villages assigned to wear them, while cloth masks were no better than no masks at all regarding the primary endpoint of blood-test-confirmed Covid-19. In an umbrella review of masking that I coauthored, we found no good evidence to support cloth masking. Two days after Walensky’s statement, the CDC conceded that cloth masking was inferior to other masks. Notably, however, this is still misleading because cloth masking is not just less effective—it is entirely ineffective.

Prasad appears to suggest that “any mask is better than no mask” is logically inconsistent with “cloth masking was inferior to other masks,” or that the CDC flip-flopped by “conceding” the latter. However, the two conditions can co-exist.

Furthermore, Prasad’s interpretation of the Bangladesh trial’s results (“cloth masks were no better than no masks at all”) is incorrect. To interpret the results we look at the confidence interval for the adjusted prevalence ratio (aPR) for diagnosis of COVID-19. From the original article:

Although the point estimates for cloth masks suggests that they reduce risk, the confidence limits include both an effect size similar to surgical masks and no effect at all. (aPR = 0.94 [0.78,1.10]; control: 0.67%; treatment: 0.61%).

The shorthand way of interpreting the confidence interval (CI) is “we are 95% confident that the true adjusted prevalence ratio is within this range”, here from 0.78 to 1.10. The CI given allows for the possibility that cloth masks are better than (<1.0), the same as (equal to 1.0), or worse (>1.0) than wearing no mask at all. So it would be more accurate to say the trial was uninformative with respect to cloth masks vs. no masks. You cannot say “this trial demonstrates that cloth masks are no better than not wearing a mask.”

To expand on this further, consider that randomized trials ask specific questions, statistically and otherwise. The statistical plan and results of the study indicate that the investigators asked a “two-sided” question: Is there a difference between masking and not masking with respect to seropositivity? The null hypothesis (default position) is “the difference is zero” and the statistical methods are specific for this default position. By stating, “cloth masks were no better than no masks at all” Prasad is stating a one-sided question similar to that of non-inferiority (of no masks). With non-inferiority, you identify a small difference below which you think is unimportant with respect to differentiating the effects of cloth masks vs. no masking, then the null hypothesis or default position is “a difference of at least that much exists.” Different statistical methods are used to answer this one-sided question than are used for two-sided questions. In essence, you don’t get to change the hypothesis after the statistical analysis is done.

The beneficial effect of surgical masks was larger than it was for cloth masks, which is unsurprising. In the Bangladesh trial surgical masks had a statistically significant effect while the effect for cloth masks was not (although the point estimate was in the desired direction). Readers should be aware that, for whatever reason, the control (non-mask) positive seroprevalence rate for COVID-19 was higher in the surgical mask villages (0.81%) than in the cloth mask villages (0.67%). It is easier to show a statistically significant improvement when the baseline situation (SARS-CoV-2 infection) is more common or more severe than when it is rare or less severe, given the same number of people or villages. In addition, there were also fewer villages that received cloth masks than received surgical masks, and a smaller sample size also makes it more difficult to demonstrate statistical significance if there is a real but small effect. Therefore, the cloth masking comparison was somewhat handicapped compared to the surgical mask comparison and is another reason “cloth masks don’t work” cannot be accepted as “the truth”.

Also consider that in the Bangladesh trial, mask wearing was 13% in control villages and 42% in intervention villages. The issue may be that the intervention villages did not reach a threshold needed for cloth masks to be effective. This distinction – do masks work or is it that we can’t get people to wear them consistently and properly? – is important. If they don’t work, we abandon the idea. If they do, our efforts go towards facilitating proper mask wearing.

Another important consideration for the Bangladesh trial is the type of cloth masks provided:

We used high-quality surgical masks that had had a filtration efficiency of 95% (standard deviation [SD] = 1%); this is substantially higher than the filtration efficiency of the cloth masks we designed, which had a filtration efficiency of 37% (SD = 6%). These cloth masks had substantially higher filtration than common commercial 3-ply cotton masks, but lower than hybrid masks that use materials not commonly available for community members in low-resource settings.

Another article found that masks using cotton cloths of hybrid fabrics have higher filtration efficiency, i.e., over 90%. The authors of the Bangladesh trial have a reasonable rationale for testing the type of cloth mask that can be used in low resource settings. The problem is that Prasad lumped all cloth masks together with his “no better than no mask at all” remark – this is clearly an oversimplification. (I will defer a discussion as to whether we can extrapolate results obtained from rural villages in Bangladesh to cities and towns in high resource countries.)

Prasad mentions his review of masking studies where he stated that the quality of most studies was poor. If he was so hot on getting results from a high-quality randomized trial of mask wearing, why didn’t he design, fund and conduct one? Given his numerous commentaries and media appearances, he seems to have plenty of time for grant writing – except that any results from a trial of his would be compromised by his competing interests.

You should wear an N95 mask

Now the CDC has endorsed the use of N95 or equivalent masks in community settings, which it presents as the superior choice. Here, too, the evidence is misleading. First, a masking policy involves more than just the filtration properties of the material; it should consider both filtration and human behavior. Will people wear the mask appropriately?

Since he can’t argue that N95 masks are ineffective, he switches tactics to say that no one is going to wear them properly anyway, so there’s still no point to wearing a mask. If the real problem is we can never get people to wear masks properly, why did he spend time talking about how surgical masks work and cloth masks don’t?

Prasad doesn’t provide a reference for where CDC made this statement. Here is the statement in context. The CDC’s position early in the pandemic was that it was important to reserve the supply of N95 masks for health care personnel. More recently, the concern about the supply of N95 masks has decreased to the point where the CDC feels they can be used by people outside of healthcare. This is neither an arbitrary nor illogical change in position.

Prasad is misleading when he states that CDC says, “You should wear an N95 mask”. The actual CDC position is, and has been, “CDC continues to recommend that you wear the most protective mask you can, that fits well and that you will wear consistently“.

The virus changes, but our policies remain the same. Masking—even if it works—is not a permanent solution. It cannot work when you stop doing it. Recently, in a striking admission, Anthony Fauci confirmed not only that the virus will not be eliminated, but also that it will eventually infect us all. Even vaccination is not enough to entirely halt omicron breakthroughs. Thus, even if N95 masking delays the time to infection, we will eventually be infected. The question becomes: Is it worth it? We aren’t getting any younger, and at some point we will have to trust our immune systems (helped by vaccination) to fight off the virus. Is it worth it for a young person to delay exposure with an inconvenient and intrusive mask?

This appears to be Prasad’s attempt to dazzle people with gibberish:

  • We can’t wear masks forever. So forgettaboutit! Also, no one has said we will have to wear masks forever.
  • In the previous paragraph Prasad said that CDC changed their policy. Now he’s saying CDC policy doesn’t change.
  • Anthony Fauci predicted that the virus will not be eliminated. Why is this a “striking admission”? Because at one time, if everyone had collaborated to reduce viral transmission, it might have been possible to essentially stamp out the virus?
  • We are all doomed to be infected. Anthony Fauci said so. Even though he is a major spokesperson for a public health organization that regularly misleads us.
  • “We will have to trust our immune systems.” What? Prasad must have read that book that said all you need to do is send your wish into the universe.

“Is it worth it for a young person to delay exposure [by wearing a %$#[email protected] mask]?” I find this philosophy of “let’s intentionally contract the disease” incomprehensible. Once again, Prasad demonstrates his lack of understanding that viruses can’t be penned like cattle. Second, as Dr. Howard has explained on SBM, some young people will become very ill with COVID-19. Even if they don’t die, they may have sequelae that last a long time. In a move I find baffling given his assertion that young people don’t suffer bad consequences from COVID-19, Prasad provides a link in a subsequent paragraph to an article where Paul Offit describes the heart wrenching suffering of children with severe COVID-19 illness.

Kids suffer no harm from not seeing faces

On August 12, 2021, after more than a year in which daycare providers routinely wore masks while caring for infants, the American Academy of Pediatrics tweeted: “Babies and young children study faces, so you may worry that having masked caregivers would harm children’s language development. There are no studies to support this concern”…The truth is, we don’t know. We are running an unprecedented experiment on our youth. We have never concealed faces from children in daycare for so many hours a day for so many years. Thus, we cannot be certain of the full effects.

Sometimes when I hear about the things we wring our hands over, it makes me appreciate how far humankind has come (at least for some fortunate countries). Generations ago, we feared the enemy would come to our village and hack our baby to death. Now, we worry about the possibility of a delay in language development.

My flippant comment aside, we should consider whether there are downsides to children not seeing adults’ faces, and how severe these effects might be. Then they must be compared to the problems of allowing COVID-19 to run through the community. First, if we don’t reduce transmission there will be staffing shortages of daycare workers. Second, since children spend time at home every day, it’s highly unlikely that young children are never seeing other people’s faces. Many parents have had to take their children out of daycare due to the pandemic, where kids are most likely seeing their parents’ faces.

If we want to stop wearing masks so that young children can see faces, this should motivate us to take actions that end the pandemic (masking, social distancing, vaccination) rather than offer up young people as a sort of “viral farm” whose crop eats us instead of the other way around.

Vaccine policy must be one-size-fits-all

First, consider boosters. The case for population-wide boosters, including for young, healthy adults, is tenuous and was contentious even among senior scientists. Marion Gruber and Phil Krause—the director and deputy director of the FDA—reportedly resigned over White House pressure to approve boosters for all…But substantial uncertainty persists that boosting a 20-year-old man will redound to his net health benefit. After two doses of mRNA vaccination, he will have a markedly reduced chance of hospitalization or death. He will also face a nonzero risk of myocarditis from a dose three. While a third dose may provide short-term protection against symptomatic disease, his disease would likely be mild anyway. We do not know with confidence that such a person should receive a booster. Recently, in light of these concerns, Paul Offit, director of the Vaccine Communication Center of Children’s Hospital of Philadelphia, advised his own son not to receive a booster.

Prasad directly acknowledges that “the case for population-wide boosters … is tenuous and was contentious even among senior scientists.” “Contentious” and “tenuous” represent uncertainty and a case for both sides of an issue rather than “falsehoods” or “omitting uncomfortable truths.”

“…uncertainty persists that boosting a 20-year-old man will redound to his net health benefit.” Public Health 101: public health policies are aimed at maximizing the health of the population, not necessarily bringing the most benefit to specific individuals within that population. The latter is the role of medicine.

I love (not) the ominous connotation of “nonzero risk of myocarditis”. Others have already demonstrated that vaccine-induced myocarditis is very rare, treatable, less severe than myocarditis caused by SARS-CoV-2 infection, and the risk of vaccine-induced myocarditis does not justify foregoing vaccination. “Nonzero risk” is also not a useful concept and if you can only accept a “zero” risk, you have other issues (eating and drinking present a nonzero risk of choking).

Recently, the government made an unusual regulatory change. It had initially recommended boosters six months after the initial vaccination series for Moderna, but it has accelerated this timetable to five months. The FDA admitted that this decision was based on data submitted for the Pfizer vaccine after the initial Pfizer series. Thus the FDA used data from a different manufacturer, at a different dose, to make vaccine policy.

I’m not sure how worrisome this particular example is, but I can’t object to Prasad’s concern here. The FDA might not want to make a habit of this. However, this isn’t a “falsehood” or “omission of an uncomfortable truth”.

A theme that resonates throughout Prasad’s commentary is the idea that public health organizations should hold off on recommendations until multiple rigorous randomized trials are done or we otherwise have accumulated a lot of solid evidence behind them. Unfortunately, during an epidemic they simply don’t have that luxury. I can’t directly comment on the job that the CDC and FDA have done, mostly because once I got the message of “masking, social distancing, and vaccination” I didn’t micromanage it beyond that (except for wanting to get an mRNA vaccine). I wouldn’t be surprised if they made some mistakes or didn’t always communicate clearly. That’s not the same thing as systematically spreading falsehoods and omitting important facts. One should also keep in mind that public health organizations, whether national, state, or local, operate under the direction of politicians. It’s not unheard of for a public health organization to recommend “A” but it doesn’t happen because “A” is not politically expedient.  In those situations, blaming the public health organization misses the point.

Recently, the CDC published an alarming study suggesting that children who get Covid-19 may develop diabetes at higher rates than normal. Yet the study was grievously flawed. It did not adjust for risk factors—being poor or overweight—that predispose to both diabetes and Covid-19. It also failed to note that the absolute risks were astonishingly low. One can’t help but wonder if the CDC promulgated a highly imperfect study to push vaccination in this age group.

It’s interesting to compare Prasad’s provocative language (“alarming”, “grievously flawed”, “astonishingly low”) with the dry language used by government employees. He seems to lean heavily on pushing people’s emotional hot buttons.

The CDC’s article is found in Morbidity and Mortality Weekly Report (MMWR) and discusses both type I and type II diabetes. True, CDC did not control for BMI or other important confounders, most likely because they couldn’t with the data they had. The evidence presented is interesting, and the results do not tip the scales to accepting as settled the idea that COVID-19 causes diabetes in some pediatric patients. The study’s authors acknowledge this. Furthermore, Prasad omitted the fact that other investigators in different populations are making this same observation of new cases of diabetes following COVID-19, both in adults and children (the MMWR article provides the references for this).

The CDC’s recommendations are reasonable: “Monitoring for long-term consequences, including signs of new diabetes, following SARS-CoV-2 infection is important in this age group.” Although the absolute risk is low, it applies to millions of children and, if this association is causal, can affect people over most of their lives.

Conclusion: Prasad fails to demonstrate that public health is the source of misinformation

To recap:

  • “CDC flip-flopped and a RCT shows cloth masks don’t work” – wrong
  • “Now the CDC endorses N95 masks for the community but no one will wear them properly anyway” – not a falsehood or omission of an “uncomfortable fact,” but a misleading statement by Prasad
  • “The virus is changing but our policies don’t, don’t listen to Anthony Fauci but listen to him, we’re not getting any younger, we need to trust our immune systems” – gibberish
  • “Masks impair child language development” – vague, scary assertion made without supporting evidence by Prasad and not the CDC or FDA
  • “The case for boosters is contentious and tenuous” – perhaps, but this is not a falsehood or omission of uncomfortable truths
  • “The FDA made a recommendation for Moderna based on Pfizer data” – true, but not a falsehood or omission of uncomfortable truth
  • “The CDC published a poorly done study suggesting an association between incident diabetes and previous COVID-19 infection in children in order to promote vaccination in children” – CDC truthfully presented the data and its context, so essentially wrong; Prasad omitted an important piece of information and made an ominous-sounding assertion without supporting evidence
  • “Falsehoods and half-truths have consequences.” I hope they will for Vinay Prasad. Public health’s real truth problem is the mistruths he continues to spew.

And finally, “Yet, repeatedly, federal agencies and respected organizations push recommendations that are deeply uncertain, rely on fearmongering, or provide hollow reassurances.” I hear the word “fearmongering” a lot. It’s always followed by a different form of fearmongering. Prasad’s commentary gives the impression of a powerful entity either implementing a hidden agenda designed to control our every move or bungling every decision. You know what I’m afraid of? Choking to death over days or weeks in the hospital without getting a chance to say goodbye to my family. Did that fear originate from the government or big Pharma? No, it came from working in a hospital and talking with physicians who take care of patients with COVID-19.

Author

  • Lynn Shaffer, PhD, is an epidemiologist and statistical analyst with the Mount Carmel Research Institute where she supports investigator-initiated clinical research. Dr. Shaffer has no competing interests to disclose, financial or otherwise. Her views with respect to this work are her own and not associated with or endorsed by Mount Carmel Research Institute or Mount Carmel Health System.

Posted by Lynn Shaffer

Lynn Shaffer, PhD, is an epidemiologist and statistical analyst with the Mount Carmel Research Institute where she supports investigator-initiated clinical research. Dr. Shaffer has no competing interests to disclose, financial or otherwise. Her views with respect to this work are her own and not associated with or endorsed by Mount Carmel Research Institute or Mount Carmel Health System.