Skip to main content

In a First, Randomized Study Shows That Masks Reduce COVID-19 Infections

A large study co-authored by Yale SOM’s Jason Abaluck and Mushfiq Mobarak tested the effectiveness of a mask-promotion program in Bangladesh in increasing mask use and preventing symptomatic infections. The study found that masks significantly lower symptomatic infections, especially among older people and when surgical masks are used.

A staff member handing a mask to a vegetable seller in Bangladesh

A member of the study's outreach team distributing masks in Bangladesh. Photo: IPA.

In the first weeks of the COVID-19 pandemic, health authorities advised against the use of masks by the public, concerned about diverting supplies from healthcare workers and creating a false sense of security that would reduce compliance with public health recommendations like social distancing and hand washing. By April 2020, the U.S. Centers for Disease Control had reversed itself; in June, the World Health Organization followed suit. Masks were eventually adopted in much of the world. But evidence of the effectiveness of masks in preventing infection has been limited to lab experiments and healthcare settings.

Now, for the first time, a randomized trial has demonstrated the effectiveness of masks in preventing infection in a real-world community setting. A large study in Bangladesh, co-authored by Yale SOM’s Jason Abaluck and Mushfiq Mobarak, found that a campaign to promote mask-wearing reduced symptomatic infections significantly, particularly among older people and those using surgical masks.

The study, which involved more than 340,000 people in 600 villages, was conducted by Mobarak and Abaluck in collaboration with the research and policy group Innovations for Poverty Action and scholars from Stanford University, the University of California Berkeley, and other universities.

In the first part of the study, completed this spring, the researchers found that a “cocktail” of four interventions, including endorsements from community leaders and reminders from roving monitors, increased mask wearing to 42% in the targeted villages, up from 13% in control villages, as recorded by an observation team. That study’s recommendations have now reached more than 100 million people, having been adopted by governments across South Asia and drawn the support of major companies and NGOs.

In the second part of the study, the researchers surveyed people in the targeted and control villages about COVID-19 symptoms. Those who reported symptoms were asked to provide blood samples, which were tested for COVID-19 antibodies. The results showed that there were 9.3% fewer symptomatic infections in villages that were targeted in the mask promotion program.

The effect was even greater in the villages where the team distributed surgical masks rather than cloth masks. In those areas, infections were 11% lower overall, 23% lower among people between 50 and 60, and 35% lower among people over 60.

The researchers emphasize that this reduction in infections was achieved with just 42% of people wearing masks. The effect of near-universal mask-wearing may be several times larger, they estimate. The increased protection from infection for older individuals, who are more at risk, is also significant.

“It would be a big mistake to infer from our paper that masks can only prevent 10% of infections,” says Abaluck. “We think what the results show is that masks are an incredibly powerful tool to reduce symptomatic COVID infections and particularly to reduce symptomatic COVID infections in the people who are most vulnerable to death or very serious illness.”

The study also provided an opportunity to test whether wearing masks does indeed result in “compensating behaviors” like reduced social distancing. That remains a key question for policymakers, including the WHO, which met with the researchers during the design of the study. The observation team reported that the percentage of individuals maintaining distancing was in fact 5 percentage points greater in villages targeted by the mask-promotion program, and the density of crowds gathering in public places did not increase.

“It was exactly the opposite of what the fear was,” Mobarak says. “Our interpretation is that once you start talking about masks, people start taking everything about COVID more seriously, including the idea of distancing.”

“It is feasible for us to cost-effectively increase mask usage and that leads to a quantitatively meaningful reduction in COVID transmission, especially among the elderly, which means that it’s also going to save lives.”

The researchers are planning multiple follow-up studies. One will clarify whether masks are reducing the number of infections or reducing the severity of infections, making more of them asymptomatic. As vaccines become more available in developing countries like Bangladesh, another study will test a variety of methods for encouraging people to be vaccinated.

Mobarak says, “Just as we built a model on how to increase mask usage, the same team is now moving on to thinking about how to build that kind of a model to improve the efficiency of last-mile delivery of vaccines in developing countries.”

But it may be a year or more before vaccines are widely available in the developing world. In the meantime, the study points to an important way to make an impact.

“It is feasible for us to cost-effectively increase mask usage and that leads to a quantitatively meaningful reduction in COVID transmission, especially among the elderly, which means that it’s also going to save lives,” Mobarak says. “And given that it’s not that expensive to run intervention with scale to get this increase in mask wearing, you end up saving lives very cost-effectively.”

Department: Research