Coronavirus May 2020—Part 8 How to Reopen a Business Safely
UPDATE: Since the original publication of this post, a review of the literature has caused us to revise our original recommendation on cloth masks. You can find the updated recommendation and the reference on which its based near the end of the post. There has also been a study published showing evidence that speaking can produce droplets and aerosols containing SARS-CoV-2 that may remain in the air for up to 8 minutes, underscoring the importance of mask wearing in closed spaces even for asymptomatic patients. We still don’t know, however, how often the viral load achieved in the air from speaking reaches the level required for an infectious dose.
Businesses around the country are facing ever-increasing pressure to find a way to reopen. If we consider the public health perspective only, maintaining the stay-at-home order still makes the most sense. We still don’t have a vaccine or effective treatment that lowers the mortality rate of COVID-19. Whenever large enough groups of people gather, there will be increases in the number of cases. Flattening the curve by maintaining the stay-at-home order spreads those increases out over time so that the healthcare system isn’t overwhelmed and everyone who needs a ventilator gets one.
But the pressure to restart the economy, to allow some businesses to reopen, is growing. In a perfect world where we could mitigate the economic consequences of the pandemic without having at least some people return to work, we’d strongly urge people to continue to shelter in place given that social distancing is currently the best weapon we have to prevent further spread of the disease. But we recognize that keeping the world locked down has resulted in significant economic damage that has to be taken into consideration—that has to be addressed—as we all plan our future behaviors.
So our purpose with this post isn’t to tell you whether or not reopen your business—that’s a judgment each of you have to make on your own—but rather to provide recommendations that will minimize the risk to your business and your employees if you do decide to reopen. We recognize that people are afraid both of contracting COVID-19 and of worsening economic disaster. Both fears are legitimate. But we think all decisions should be made based on science and statistics, not emotion. Our intent, then, is to provide you with the most accurate, up-to-date evidence-based statistics possible.
Some have been thinking that we can’t fully reopen the economy until we have a vaccine. But how soon might that happen? We don’t know. We might never. We still don’t have one for hepatitis C or HIV. But even if we do develop one (and many are currently in the works), proving it both to be safe and effective will take time. To deliberately infect volunteers who’ve been vaccinated against SARS-CoV-2 to see if they’re immune would be unethical given the fact that the disease is fatal in some cases (you can do this with animals, but animals aren’t humans). Instead, once we have a viable vaccine candidate that’s passed Phase 1 and 2 trials, we immunize a large group of people and follow them through time, comparing their rates of COVID-19 infection to a matched group that hasn’t been immunized, all while observing for complications that didn’t show up in the Phase 1 or 2 trials.
Unfortunately, sometimes the treatment is worse than the disease. Studies show that when given immunizations to some diseases—dengue, respiratory syncytial virus, and SARS, for example—a paradoxical phenomenon occurs: subjects develop more severe disease. There are thought to be two mechanisms that cause this: 1) antibody-dependent enhancement, where a virus leverages the antibodies that we generate in response to vaccination to aid infection, and 2) cell-based enhancement, where immunization leads to severe allergic inflammation that can cause worse disease outcomes.
Unfortunately, vaccines used to induce mice to produce antibodies to the virus that causes SARS—a coronavirus similar to the virus that causes COVID-19—were shown to place the mice at high risk for life-threatening cell-based enhancement. Fortunately, when researchers altered their vaccine strategy and aimed to create a vaccine against only a portion of the spike protein of the SARS virus, cell-based enhancement was blocked. Sadly, funding dried up just as researchers were about to proceed to clinical trials in humans. Now, with the COVID-19 pandemic, that research is restarting.
How many people infected with SARS-CoV-2 develop their own antibodies? In one study, 64 percent of a patients who recovered from COVID-19 developed high concentrations of antibodies to the virus, 30 percent developed low concentrations of antibodies to the virus, and 6 percent didn’t develop antibodies to the virus at all. In that same study, older patients were found to be more likely to generate higher levels of antibodies. This might be one reason older patients are more likely to die than younger patients from COVID-19: one study found that a faster and more robust antibody response is seen in patients who die from SARS compared to those who don’t, and that these antibodies may be involved in causing the cytokine storm that leads to severe lung injury and death in COVID-19 as well.
THE BOTTOM LINE: We still don’t have a vaccine or know when or even if we will have one. We also don’t know if infection produces immunity, for how long that immunity might last, or if infection predisposes to a more severe course from re-infection.
spread of infection
Evidence is accumulating, however, that sustained exposure to infected patients—whether those patients are symptomatic or not—is what largely determines the risk of transmission.
One contact tracing study looked at the first 10 known symptomatic cases in the U.S. and found they had contact with 445 people after they became symptomatic. The transmission rate to these 445 people was only 0.45 percent for all close contacts and 10.5 percent among close household contacts. (Because these were 445 people who had contact with the index patients after the index patients became symptomatic, we don’t know how many other people may have become infected who had close contact with these index patients before they became symptomatic, so the true transmission rate for all close contacts may have been higher.) Importantly, “close contact” in this study was defined according to CDC guidelines as “being within approximately 6 feet (2 meters) of a person with COVID-19 for a prolonged period of time (such as caring for or visiting the patient; or sitting within 6 feet of the patient in a healthcare waiting area or room); or having unprotected direct contact with infectious secretions or excretions of the patient (e.g., being coughed on, touching used tissues with a bare hand).” Four percent of these close contacts (19) were household close contacts and 26 percent (5) of those continued to have exposure while in the household with the index patient. This may account for the higher secondary infection rate among household close contacts compared to close contacts outside of the household. Thus, it seems even when people ignore current social distancing guidelines, the risk of transmission is, in fact, low, and—given the much higher rate among household contacts—largely determined by amount of contact time a person has with an infected patient. This is likely because with more contact time, higher risk scenarios are more likely to occur, i.e., coughing without a mask on, touching shared surfaces and then one’s face, and so on. It seems it’s not quite as easy as many think to encounter an infectious dose of the SARS-CoV-2 virus.
Another contact tracing study from China found a transmission rate of 9.7 percent among close contacts of index cases. In contrast to the U.S. study above, close contacts were identified beginning two days before the index cases became symptomatic, perhaps partly explaining the higher transmission rate (9.7 percent vs. 0.45 percent in the U.S. study above) as infected patients are now thought to be infectious for as many as three days before they become symptomatic. Interestingly, this study found a transmission rate close to that of the U.S. study above among household contacts of index cases at 14.9 percent (vs. 10.5 percent in the U.S. study). This study also showed that 80 percent of infections were caused by only 9 percent of index cases, suggesting that only the minority of people are highly infectious and that most infections come from them, arguing that “super spreader” events may be responsible for much of the spread of the disease, either because of an index case’s behavior, intrinsic infectiousness (so-called “superemitters”), or some combination of both. A third study, also from China, used contact tracing data to model the transmission rate among household contacts, which suggested it would be 19.3 percent when identified cases aren’t isolated and close contacts are exposed to the full time period during which an infected person is infectious—that is, in a worst-case scenario for household close contact transmission. A fourth Chinese study found the transmission rate among household contacts was 16.3 percent. The risk of transmission in that study when the household contact was a spouse was 27.8 percent. Finally, another contact tracing study, one out of Taiwan, included close contacts of index cases starting from four days before the index case became symptomatic to the date at which the index case was diagnosed (after symptoms had appeared). It found that overall only 0.8 percent of close contacts of infected patients acquired the infection and 4.6 percent of close household contacts acquired the infection (similar rates to the study above of the first 10 known symptomatic cases in the U.S.).
What does all this mean? First, transmission rates from contact tracing studies seem to cluster under 1 percent or around 10-20 percent. The difference may be due to behavior differences in study subjects, to differences in average contact time of close contacts with index cases, or to methodological differences in the studies. Whatever the reason for the differences in attack rates, all the studies suggest that when you have close contact with an infected person, the risk of becoming infected is, in fact, relatively low. Consider also that what defined “close contact” behavior in the studies above is behavior that most of us are now doing our best to avoid (e.g., being closer than six feet to others for a prolonged period of time or having unprotected direct contact with infected secretions).
This suggests that if we follow the current guidelines—remain further apart than 6 feet from one another, refrain from touching our faces after contacting potentially contaminated surfaces, wash our hands frequently, and wear masks when around others all in the service of avoiding “close contact”—we can probably reduce the rates of transmission when we encounter an infected patient down to the lowest rates seen in the studies above, i.e., to 0.45 to 0.8 percent.
What’s more, even these low rates of infection will apply only if we come in contact with an index case. If we don’t come in contact with an index case, our risk of becoming infected drops, of course, to zero.
What this means, among other things, is that everyone must become strict about self-quarantining once they show any infectious symptoms whatsoever (including: sinus congestion, runny nose, sore throat, ear pain, cough, vomiting, diarrhea, fever, chills, body aches, loss of smell or taste, discolored toes). In America, we’ve endured a culture of going to work while sick that must now end.
If it does end, then besides our strict compliance with measures that prevent “close contact,” our real risk of becoming infected hinges on our risk of encountering an asymptomatic person with COVID-19. So what is that risk?
The only way to know that risk with precision would be to know the total number of asymptomatic cases in the population, and the only way to know that would be to test every single person with a test that has a specificity and sensitivity near 100 percent. There are, however, multiple ways to estimate the prevalence of asymptomatic cases. Perhaps the most accurate way of estimating the prevalence of asymptomatic cases currently would be to look at contact tracing studies to see what proportion of infections that developed from index cases were asymptomatic. Both the study out of Taiwan and the study out of Shenzhen, China found this rate to be 20 percent. As of this writing, then, that would be 905K cases (calculation not shown), leading to a prevalence of asymptomatic infection in the U.S. of 0.27 percent. Another way would be to do what we did in an earlier post and use the data from the Diamond Princess study that showed in a confined population about 50 percent of all cases were asymptomatic (this is likely less accurate as the population aboard the Diamond Princess didn’t represent the general population; the average age, for example, was a bit higher than in the general population). As of this writing, that would be 3.6M (calculation not shown). That would equal a prevalence in the U.S. of 1.1 percent. Thus, the cumulative prevalence of asymptomatic cases in the U.S as of this writing is likely somewhere between 0.27 and 1.1 percent.
How likely would you be to catch COVID-19 from an asymptomatic person working in the same office as you? Research suggests that for close contact with asymptomatic cases, the overall risk of catching COVID-19 is only 0.33 percent (compared to a risk of 3.3 percent for sustained close contact with mildly symptomatic cases, a risk of 6.2 percent for sustained close contact with severely symptomatic cases, and a risk of 13.6 percent for sustained close contact with someone who’s coughing and expectorating, meaning bringing up phlegm).
Certainly, we’ve seen much higher attack rates in other settings that mimic households: in nursing homes and skilled nursing facilities, for example. But in a study of a skilled nursing facility in which 64 percent of residents tested positive for SARS-CoV-2, a full eight days passed between the time the index case was identified in a health care worker and the facility implemented transmission-based precautions on the unit in which the health care worker worked. Thus, the explanation for the explosive spread of disease in these settings likely lies in both symptomatic and asymptomatic transmission due to inadequate precautions and the continued relative immobility of the residents, who were old and sick and therefore also more susceptible to infection. Similarly, a study of an outbreak in an office setting in South Korea, in a call center, showed a transmission rate of 43.5 percent. Importantly, though, 91.7 percent of cases were symptomatic. The significantly increased likelihood of symptomatic patients transmitting the disease (3.3 to 13.6 percent from the study above) combined with the long exposure time of an 8-hour workday over many days likely explains the high rate of transmission in this setting. Another study of a cluster at a Chinese restaurant caused by a pre-symptomatic patient sitting in a strongly air-conditioned room with other families who subsequently became infected highlights the importance of prolonged contact, airflow dynamics, and the likelihood of respiratory droplet and aerosol transmission from index patients not wearing masks.
To reopen or not to reopen
Given the data above around risk of transmission, could some businesses reopen while still keeping the curve flat—that is, without increasing the likelihood that the healthcare system will become overwhelmed? It’s not clear. To enjoy the low rates of transmission described above, certain gatherings simply can’t happen yet—large conventions, religious services, concerts, and so on—where large numbers of people come together in close proximity for prolonged periods of time. But could some businesses mitigate the factors that contribute to the spread of COVID-19 enough to reopen? If every single person in every single business behaved consistently with social distancing guidelines every day, the answer might be yes.
It’s much more difficult to control the behavior of large groups of people, i.e., all of society, than it is small groups, i.e., individual businesses. For that reason, much blunter measures are required for larger groups. This is the reason for the states’ stay-at-home orders and why the mayor of Chicago closed the lakefront. But in smaller groups, like a business, we can envision successfully implementing more precise measures that might allow for more latitude. In smaller groups, it becomes possible to ensure extremely high levels of compliance with safety measures. There can be a single person or small group of people responsible for enforcing safety measures that keep everyone safe when everyone follows them. To emphasize this last point, the biggest challenge any employer will face in reopening lies in finding effective strategies that change the company’s culture into one that prioritizes behaviors that keep everyone safe. That is, every business must find a way to promote a culture in which its employees think about their colleagues’ safety as much as their own.
Below, then, is a list of measures that businesses could adopt to try to achieve the low rates of transmission described above. We offer this list not as an endorsement for reopening but rather in recognition that some businesses feel they have to reopen and are going to reopen and need guidance about how to do so as safely as possible.
- Stay more than 6 feet away from each other as much as possible. This may mean reconfiguring your work space. This is one of the most important interventions.
- Other than that, the most important thing is that workers self-screen for infectious symptoms. If an employee has any symptoms of infection (fever, body aches, chills, runny or stuffy nose, ear pain, sore throat, cough, vomiting, diarrhea, loss of taste or smell, discoloration of toes, the so-called “COVID toe”), they should stay home as soon as the symptoms appear and remain home for 7-10 days, which is the point at which infectivity seems to significantly drop off according to studies.
- You could institute temperature checks for all employees every day, but it’s not likely to be of much value because many employees show symptoms, and are therefore infectious, 5 days before developing fevers—if they develop a fever at all.
- There’s no good reason to think that screening asymptomatic employees with PCR tests at some interval will reduce the likelihood of infection entering your workplace. Also, the PCR tests are quite expensive.
- Allow any employees who can work from home to still do so.
- High risk individuals should consider working at home.
- To whatever degree you can, stagger the employees who come to work to minimize the total number of people in the office on any one day.
- Encourage frequent hand washing with soap and water.
- Encourage the avoidance of face touching.
- Have hand sanitizer with at least a 60 percent alcohol content widely available.
- Have plenty of moisturizer available as hand washing with soap and water is best.
- Wear masks at all times while at work. You should not remove your mask even if you work inside a closed-door office by yourself, because your HVAC may spread virus particles from a non-mask-covered cough throughout the office. Ideally, masks should be surgical or medical masks, not cloth masks. Some household materials have been suggested by studies to filter aerosols and respiratory droplets to some degree better than wearing no mask at all, but they’re inferior to surgical masks. For example, one study of cloth masks compared to surgical masks found that healthcare workers wearing cloth masks were 13 times more likely to experience influenza-like illness than those wearing surgical masks. On the other hand, surgical masks may not be widely available to the general public. Employers should make every effort to provide new surgical masks to employees at the moment employees enter the office on a daily basis, but if surgical masks aren’t available, cloth masks are better than no mask at all. Also, it should go without saying—but we need to say it because we’ve seen it—make sure your employees know not to pull their masks down to cough or sneeze.
- One study showed that disinfecting high-touch surfaces (soap dispensers, door handles—especially to the main entrance—faucet handles, countertops, sinks, light switches, computer keyboards, mice, keys, cell phones, cell phone cases, etc.) once midday resulted in an 85 percent reduction in transmissible virus on employees themselves. More frequent cleaning might further reduce the risk of spread.
- Communicate the data above—the actual risks of spread—to reassure employees who must come into the office.
- Reinforce your policies on a daily basis (over communicate).
- Make someone responsible for conducting daily rounds to ensure compliance with your policies.
- Create a plan for what you’re going to do if you get a COVID-19 positive case.
- Create a plan now for what you’re going to do if there is a second shelter-in-place order.
We hope if you’re considering reopening your business that the recommendations we’ve listed here, as well as the science and statistics on which they’re based, give you confidence that you’ll be able to take steps to minimize the risk to your business, your employees, and your customers.
For previous posts related to COVID-19, see:
- Coronavirus February 2020—Part 1 What We Know So Far
- Coronavirus March 2020—Part 2 Measures to Protect Yourself
- Supporting Employee Health During the Coronavirus Pandemic
- Coronavirus March 2020—Part 3 Symptoms and Risks
- Coronavirus March 2020—Part 4 The Truth about Hydroxychloroquine
- Coronavirus April 2020—Part 5 The Real Risk of Death
- Coronavirus April 2020—Part 6 Evaluating Diagnostic Tests
- Coronavirus April 2020—Part 7 The Accuracy of Our Antibody Test