But now we should have more data from this pandemic to guide our decisions. We don’t send rockets into space without collecting data to monitor their progress and detect if they are veering off course. And yet we witnessed more than one million Covid-19 deaths in the United States without a clear plan to assess whether we were doing all we could to prevent more.
We should be systematically studying pandemic mitigation efforts in order to learn which interventions are effective and how best to employ them. Just as important: We should do so with the understanding that the absence of evidence of effectiveness is not the same as having evidence of ineffectiveness.
Questions about masking, for example, were recently revived by a Cochrane study reporting that masking (with surgical ones or respirators like N95) makes “little or no difference” in reducing infection at the population level, such as among health care workers or in communities. Some mask opponents claim this validates their assertions that masks don’t work. Some mask supporters are raising questions about the study’s authors and attempting to discredit their conclusions. Which side is right?
As with most things about the Covid pandemic, the answer is most likely somewhere in between.
There is good evidence that masks can protect people who use them correctly and consistently. Laboratory studies clearly show that wearing a mask properly, when in the presence of the virus, will reduce a person’s exposure to it. Other studies show that higher-quality masks, such as N95 respirators, are better able to keep the virus out than less well-fitting surgical masks or cloth masks.
The confusion occurs when we shift from showing that masks work in a laboratory or for individual people to finding evidence that masking works at the population level and what interventions work to encourage it. At the population level, compliance and mask quality may vary, making it difficult to find evidence to review on the effectiveness of masking in reducing the number of respiratory infections. The Cochrane review tried to untangle the evidence in one analysis. And according to that limited evidence, masking at the population level did not have a clear impact on reducing infections. How can this be? Part of the reason has to do with the quality of studies on masking. Though there have been studies observing differences in disease rates between places with masking policies and those without, evidence from these observational studies isn’t of the highest quality because it doesn’t fully account for other differences between masking and nonmasking populations.
To address the quality issue of these studies, the Cochrane review looked only at randomized trials evaluating the effectiveness of masking. Randomized trials are particularly helpful for studying the impact of health interventions within populations because they help to minimize bias and confusion caused by other factors besides the one you are trying to evaluate. For example, if you looked at infections among people who choose to mask versus those who don’t, you may be observing not just the effect of masking but also the effects of other protective decisions that people who are inclined to mask may also take, such as avoiding crowded indoor spaces.
There have been only a few randomized trials specific to masking, and most of the ones included in the Cochrane review were not conducted during the Covid-19 pandemic or in the United States. Many of the studies that the Cochrane review included looked at the spread of influenza.
This is important because while we think there are some similarities between how the novel coronavirus and other respiratory viruses are spread, there also are likely to be important differences. Covid proved to be deadlier than seasonal influenza, which may have influenced how often and well people wore masks. Masking for Covid was also mandated throughout much of the United States, which most likely also influenced masking behavior.
The pertinent question isn’t whether masks work but why masking didn’t prove to be highly effective in the most rigorous studies. Was it because adherence to masking, not the masks themselves, was the problem? Is it because the population studied wore masks when around infected people but then got infected from family members? Maybe people didn’t wear masks properly because they weren’t comfortable or they didn’t fit. Knowing the answers to these questions will help us know how best to use masks and help us better control the spread of infections.
The Cochrane review authors say their examination was limited to whether interventions to promote mask wearing help to slow the spread of respiratory viruses. It’s important to note that masks only work when people wear them, so adherence will always be key.
In early 2020, when we knew little about the virus but saw its toll, masks were a reasonable step because they had few harms. Considering the rapid spread of the virus and its deadly impact, we could not wait until we had all the data to understand how best to use them. And if a new, deadly respiratory disease emerged tomorrow, we’d have few tools besides masks to prevent spread and protect ourselves.
But we should have put into place efforts to rapidly collect and assess high-quality data to understand whether masks were having optimal effectiveness and, if not, how to increase that effectiveness. We should have done this for other mitigations, too, like school and business closings and policies that required exposed contacts of cases to quarantine. Pandemic measures like masking and vaccination have been challenged by political leaders and in the courts. Without clear evidence at the population level that mitigation measures meaningfully change transmission rates, it will be harder to meet challenges that could block effective, lifesaving interventions.
We need to develop clear plans for randomized and other well-designed studies and get them funded. A review of research by investigators affiliated with U.S. governmental public health entities during the pandemic found very few studies that evaluated the impact of measures to control the spread of disease. It is ludicrous to simply hope academic researchers will spontaneously choose and muster the resources necessary to address the most pressing pandemic response questions. Just as we have established research networks and protocols to conduct the highest-quality evaluations of the effectiveness of vaccines, we should have the same for nonpharmaceutical interventions, like masking. We can and must identify the highest-priority research questions and the funding to systematically and rigorously investigate them.