A study goes viral for all the wrong reasons

EVERYONE LOSES TIME to COVID-19. Martin Fischer lost most of a month to masks.

“I’m not getting anything done other than this,” says Fischer, associate research professor in the department of chemistry. “The last three weeks have been this, 100 percent.” By “this” he meant media availabilities, Zoom interviews, and various other responses to his attempt to help out as masks spread through the culture.

Surprise, “especially since my usual line of work is not really involved in topics that get a lot of media attention.” Fischer works in optics, developing techniques in microscopy in fields like materials science and biomedical materials. Then came COVID-19, and he helped out a colleague.

Eric Westman, associate professor of medicine, was working with other physicians on a project called Covering the Triangle that, in the early days of the pandemic, helped sew, acquire, and distribute masks to people in the Triangle who needed them. Westman had a line on some masks he could buy in bulk, Fischer says, but at that point getting masks was a somewhat dodgy business, and Westman wanted to make sure they worked. Sending them to a lab would be expensive and slow and might not even really measure them under real-world conditions. So Westman sent an e-mail asking for help to the physics department. It ended up with Fischer and his optics techniques.

Fischer had seen scientists at the National Institutes of Health use a method of widening a laser beam vertically into a sort of sheet of light; through a computer algorithm, they could then count droplets in the air as they passed through. A subject spoke through a mask into a special box with no air currents. Fischer followed the method and built a quick and cheap mechanism to do the same. To compare the mask Westman was interested in, Fischer expanded the idea, using whatever masks happened to be lying around.

Don’t think about a big complex setup, he points out. He had to get permission from the dean to come on campus to do the work; his daughter, a Duke student, functioned as his assistant because they lived in the same house and so transmission wasn’t an issue. It was a kind of spare-time project. “This project is unfunded,” he said.

They tested fourteen mask types, with one person talking through the mask and the other using the camera. And again: The point was the method, and sharing it so others could quickly test masks hither and yon. “This was not a largescale systematic study of all mask types under all conditions,” Fischer says. “But we developed a very simple method of visualizing the effects of a mask.”

The data looked worthy, so he and his colleagues thought this easy and cheap method of mask measurement should spread. “It’s super easy to set up. Let’s hope people pick this up and do their own demonstrations and quick checks,” Fischer said. “This was never intended to be a certification or an endorsement of any kind of mask.”

Tell that to the media. The study came out. Then came dozens of stories, mostly focusing on the fact that according to the study’s measurements, one mask—a thin neck gaiter—actually resulted in more droplets than no mask at all. Reporters got excited, and stories about some masks being worse than nothing at all spread, were criticized, were responded to, were amended. And always with more phone calls to Fischer.

Mind you, Fischer points out, the study made no claims about the effects of more or fewer droplets, the viral load per particle, or anything like that. This was a quick proof-of-principle test, and it worked. As science, it was something of a counterexample to his daughter, who got her first publication out of it. “I did repeatedly tell her, don’t think this is usual science,” he says of the study, which took a couple of weeks of research and was printed a couple of months later. “Usual science works differently. How often do you go into a lab and a couple days of data-taking you have a paper that has a million-anda- half views?”

That’s a lot of views; it’s spreading quickly. Spreading almost like a…almost like something.