Nathan Shapiro, MD, (Emergency Medicine, BIDMC) discusses restaurants and the COVID-19 pandemic. Shapiro was one of the authors of a CDC study released in September that found that people who tested positive for COVID-19 were approximately twice as likely to have reported dining at a restaurant than those who tested negative.
MIT Technology Review – October 28, 2020
How to make restaurants safer during the pandemic
It’s a cruel irony that the things that make a restaurant appealing are precisely what currently make it dangerous—the intimacy, the coziness, the groups of people deep in conversation, whiling away the hours over drinks and a meal. Eating in a restaurant is one of the riskiest things you can do during the coronavirus pandemic.
To understand why, you need to think about the latest science around how covid-19 passes from person to person. The official line from the World Health Organization from the start of the pandemic has been that the coronavirus is mostly spread by the droplets we generate as we talk, sneeze, or cough. However, the evidence has been mounting for months now that aerosols—which are smaller than droplets and can hang in the air like smoke—are a significant route for infections, if not the main one. This would explain why virtually every recorded coronavirus outbreak has occurred indoors.
Sadly, the advice to the public still hasn’t caught up. The US Centers for Disease Control and Prevention has only just started to acknowledge the possibility of airborne transmission, and many countries don’t mention it in their official guidance. As a result, many restaurants are still stuck following advice that simply isn’t reflective of the latest science—obsessing over cleaning, wearing visors (which don’t protect you from aerosols), and setting up plastic dividers between tables. Some of these measures might be marginally useful, but they mostly amount to “pandemic theater”—interventions that provide the appearance of safety, but little in the way of real protection.
Why, exactly, are restaurants so risky? First off, they tend to be noisy spaces. People talk loudly, expelling more air than usual—and thus more potentially virus-laden aerosols. Researchers are yet to work out precisely how much virus you have to breathe in, or how long you have to be exposed to someone shedding viral particles, to get infected. The CDC estimates it’s possible to get infected from just 15 minutes of close proximity, but the reported cases of infections in restaurants “all involve an infected and susceptible person sharing the air for a significant amount of time, often 30 minutes up to a few hours,” says Jose-Luis Jimenez, a chemistry professor at the University of Colorado, Boulder, who has studied aerosols for two decades. It’s also possible, theoretically, to catch covid-19 through the aerosols left behind by an infected person who has already left the room—but there aren’t any confirmed cases of this occurring, according to Jimenez. The virus loses infectivity with time, “typically in one to two hours,” he says.
Then there’s the lack of mask-wearing inside restaurants. Diners tend to take them off, because you can’t eat or drink while wearing one. You may have heard that ventilation is pretty important as well—another area in which restaurants typically score poorly. Inadequate ventilation allows tiny virus particles to hang in the air for long periods of time, just waiting to be breathed in.
And of course, for any restaurant to be successful, it needs to be popular enough to attract people from around a neighborhood, city, or even further afield to come and dine under the same roof. It’s hard to imagine a more inviting setting for an airborne pathogen like SARS-CoV-2 to spread (other than perhaps cruise ships). It’s little wonder, then, that restaurants have shown themselves to be the perfect breeding ground for superspreading events, where one person passes the virus to dozens of others. Virtually every documented case of superspreading has taken place in a noisy, poorly ventilated room—many of them restaurants.
At the start of October, Public Health England found that for people who’ve tested positive for the coronavirus in the last two months, “eating out was the most commonly reported activity in the two to seven days prior to symptom onset.” Scotland’s government has consistently found that a quarter of people returning positive tests for covid-19 had been to a restaurant, pub, or cafe in the week before. In September, a CDC study of 802 adults in the US found that people who tested positive for covid-19 were approximately twice as likely to have reported dining at a restaurant than those who tested negative.
“Without a doubt, there’s an association there,” says Nathan Shapiro, a professor of emergency medicine at the Beth Israel Deaconess Medical Center, one of the authors of the CDC study.
With the growing case against dining out, it’s no wonder the pandemic has devastated the restaurant business. While some big-name chain restaurants with drive-through and takeout options have thrived, tens of thousands of dine-in restaurants have been forced to close, potentially taking millions of people’s livelihoods with them.
Making eating out safer
Despite the dire outlook, there are ways to reopen restaurants while minimizing the risk of infection “Any time there are people indoors there is risk,” says William Bahnfleth, a professor of architectural engineering at Pennsylvania State University. But many of the dangers can be mitigated. The crucial thing to remember is that no one measure is enough on its own; increasing safety is about layering as many different efforts on top of one another as possible.
First and foremost, people should eat outdoors whenever possible. “The risk of infection is 20 times higher inside than outside,” says Bjorn Birnir, director of UC Santa Barbara’s Center for Complex and Nonlinear Science. However, some restaurants either can’t get the approval for outdoor seating from their local authorities or don’t have the money for outdoor furniture or the patio heaters that will help make diners comfortable as winter rapidly approaches in the Northern Hemisphere.
If outdoor seating isn’t possible, eateries should focus on simpler stuff. Servers need to wear masks, as should customers while they’re not at their table. Although masks won’t prevent all aerosols from getting through, they will stop some. Tables should be as far away from each other as possible. Again, this isn’t a perfect solution—but the farther away you are from other groups of customers, the less likely you are to inhale a big concentration of their breath. Use the measures you’d take to try to avoid secondhand smoke as an analogy, says Jimenez.
Some adaptations are more inventive. For example, restaurants should turn the music down to discourage customers from talking loudly, says Sam Harrison, who owns a brasserie called Sam’s Riverside in London. And although it might feel unnatural, it’s a good idea for diners to sit diagonally from anyone who isn’t in their household. Simulations generated by the supercomputer Fugaku in Japan found that about 75% fewer droplets will reach you that way than if you sit opposite someone.
It’s difficult to judge how safe a restaurant is just by looking at it. You can’t tell at a glance how many air changes per hour are taking place. Bahnfleth, the architectural engineer, says you want to aim for six full replacements of indoor air with virus-free air per hour. You can do that by opening a window or door, filtering the air to remove particles, or disinfecting the air with germicidal ultraviolet light. It’s tricky to measure the air change rate without hiring expensive air quality consultants, but one shortcut could be to use a carbon dioxide monitor (you can buy these for about $150) as a proxy. If your levels stay below about 800-950 parts per million, ventilation is probably sufficient.
Keeping score
Restaurateurs who want to get an idea of how well they’re addressing risk can use one of the free online risk estimators from places like Setty, an engineering firm, experts at Oregon University, or the University of Colorado, Boulder. These models let you input details about your space—size, ceiling height, average occupancy, and so on—and then produce a score that tells you roughly how safe it is. The risk scores are based on modeling of relative aerosol risk, and they require a good basic grasp of numeracy and science, but they can be a useful tool. “These are the best things we have, but they’re still based on a fairly uncertain degree of knowledge about how much virus an infected person sheds, and how much you need to inhale to get infected,” Bahnfleth says. Although they’re based on peer-reviewed science, they should be taken as guides rather than immutable truths, because they rely on many unknowns (they can’t know, for example, if people are wearing their masks correctly).
If open doors aren’t an option, air purifiers can dispose of more than 99% of aerosols in the air stream that passes through them. Some restaurants may already have these installed as part of their overall heating, ventilation, and air conditioning system. For those that don’t, standalone purifiers cost about $100 apiece and can be placed around the dining area.
Finally, there’s a category of interventions that might be marginally useful but verge on pandemic theater. Temperature checks are widespread and highly visible, and can help to weed out some people with symptoms—but they do nothing to prevent asymptomatic people from entering the premises. Dividers between tables, meanwhile, could stop people from sneezing or coughing on each other, but are useless to stop aerosol transmission.
The sad truth is that as long as there are high levels of virus circulating in a community, people are going to be justifiably nervous about eating out. That’s something restaurant owners can’t control. All they can do is adapt—more takeout meals, more outdoor seating—and try to survive. Harrison, the owner of Sam’s Riverside, doesn’t see a return to pre-pandemic levels of profit for the foreseeable future. “It won’t kill us, but it’ll get pretty damn close,” he says.
