Why do we have to lie about the results of the Bangladesh cluster RCT of masks?
Surgical masks won; Cloth Masks lost. Now does it apply to USA?
The Bangladesh cluster RCT of masks was incredibly helpful. It showed that in a rural Bangladesh—where there was essentially no vaccination and essentially no natural immunity—surgical masks slowed the spread of symptomatic sars-cov-2, while cloth masks completely failed on the primary endpoint (symptom driven seroprevalence), being no better than the control arm of no masks.
That’s useful information, and its an important trial, and I wrote about it here: https://www.medpagetoday.com/opinion/vinay-prasad/94399
Now, lets be clear what the trial doesn’t show:
It doesn’t apply to kids in school. It was a study of adults only.
It doesn’t apply to places with high baseline natural immunity & high vaccination rates.
In short, it doesn’t seem very relevant for a place like San Francisco right now (massive vaccination), or for US schools.
But, one takeaway that I thought was indisputable was this: if you recommend a mask—don’t recommend a cloth one.
Now a NYtimes Op-ed is out by the study authors. The authors claim, “ Getting more people to wear [masks]— through mandates or strategies like handing out masks at churches and other public events — could save…. hundreds each day in the United States.”
I think that is speculative. Their well done study does not apply to the USA in this moment with 65% 1st vaccination rate & a seroprevalence that has been poorly studied, but likely ~20%.
Do surgical masks work in this setting? At a minimum, their trial does not help us estimate the effect size.
But, the claim they make that truly baffles me is this one, “If you have the ability to choose between a cloth and a surgical mask, go with surgical. But the best mask is one that a person will actually wear and wear correctly.”
The first part of the claim is true (surgical won), but the second part is just untrue. Their own study failed to show cloth masks work on the primary endpoint.
There is no harm is saying the truth: we recommended the wrong mask.
And that’s why we need cluster RCTs.
Whoops!
Check out my detailed interpretation of their study: https://www.medpagetoday.com/opinion/vinay-prasad/94399
Can you help me understand the clinical significance of slowing the spread of COVID from 0.76% to 0.67% (< 0.1% absolute risk reduction) with surgical masks? It's nice to finally have a cluster RCT on masks. But even in a well-done high quality study, a statistically significant finding may not translate to clinical significance and advisable public policy? This study may show us that cloth masks don't provide benefit. But what degree of benefit are surgical masks providing (which also needs to weighed against the potential clinical and societal "costs/harm" of masks)?
I think it's very telling the only cluster RCT is out of....Bangladesh. We certainly have the resources to conduct more relevant data to our specific situation within the United States. I think taking the whole thing a step further; what should be the endpoint purpose of masks in regard to what many would consider an endemic virus? In other words does the small benefit of slowing the spread in a cloth mask, at a time when a vast majority are either vaccinated or have natural immunity - serve any real medical purpose in light of the fact we will all encounter this virus.