Superspreaders and Public Health Policy

New epidemiological research on SARS-CoV-2 can greatly improve public health policies.

The speed with which the novel coronavirus (SARS-CoV-2) pandemic spread throughout the world has forced us to very quickly establish sweeping, one-size-fits-all policies, just for survival: We shut everything down. Though shelter-in-place orders had been seen as successful in flattening the infection curve, we’ve suffered considerable economic and social consequences from these blunt and sweeping instruments, with the impact on unaffected communities being particularly devastating. Yet, improving these policies has been difficult because we started from knowing virtually nothing about the virus, and progress has been slow.

We may have turned a corner. The latest research by two epidemiologists, who published this NYTimes article, “Just Stop the Superspreading,” provides starkly new insights about the SARS-CoV-2 that presents an opportunity to refine these policies to be more effective, pragmatic, and sustainable. Because the authors did not discuss these ramifications of their findings, only an analysis of their research, I think it’s important to consider the potential social and economic policy changes their findings suggests, which can directly address societal and economic distresses.

The key take-aways and a brief analysis from their research are enumerated below, followed by a larger discussion.

Roughly 8–9% of coronavirus carriers are responsible for 80% of the spread to others. These individuals are called ‘Superspreaders.’

Prior to this current research, the models for the coronavirus’ progression assumed national contagion would be evenly and smoothly distributed throughout the population, which served as the basis for the initial (and mostly still current) social and economic policies. However, the new research shows that the disease is not spreading evenly, but overwhelmingly by specific individuals — Superspreaders — who are very few and far between. This means that infections don’t disperse out uniformly in waves, but in clumpy hot-spots, with disproportionately high rates of infection in some regions, while others are largely unaffected. This is called the dispersion factor, also known as “k” in statistical modeling. The new understanding of the virus’ infectivity may allow for policy revisions that can reduce the social and economic burdens on less-affected regions, and even stimulate regrowth where possible.

SARS-CoV-2 carriers are most infectious within the first few days of onset, long before they are symptomatic. More interestingly, the degree of infectiousness diminishes rapidly after symptoms begin, even among Superspreaders.

This is probably the most important finding, and presents an entirely new paradigm in our understanding of the virus and the implications on public policy. The coronavirus’ behavior runs entirely counter to that of other viruses and bacteria, such as flu, TB, AIDS, etc. For these pathogens, an affected person is unlikely to infect others until they show symptoms, and then becomes more infectious as the disease progresses in their systems. And this is what we’re used to. But the latest research suggests that the Coronavirus May Be a Blood Vessel Disease, Which Explains Everything. While the virus begins by lodging itself into the lungs, throat and nasal cavities — which is when it’s most contagious — it then migrates to the blood, where it actually does all the damage, and is no longer contagious. This suggests that the original modeling for its spread and behavior has been flawed. And since these models shaped the policies we adopted, those should all be re-examined.

Carriers are apparently only passing on the virus to others in contained, crowded spaces, where individuals reside for long periods of time. Importantly, that transmission diminishes rapidly as the amount of space expands, or density between people expands, or the time of exposure is reduced.

It’s not news that the highest risk conditions are those that the Japanese call “The Three Cs”: closed spaces, crowds and close contacts. (Ok, that’s four C’s, but who’s counting.) What’s new in this data is that these high-risk environments account for virtually all infections. As the physical (interior) spaces expand, or the distance between individuals expands, or the time spent within a space is reduced, the risk of transmission diminishes quickly. If two or all three of these factors is mitigated, risk is reduced to nearly inconsequential levels. This presents an opportunity to revise policies from the old model of assuming that all spaces are equally risky to, instead, a model that attributes risk profiles (and hence policies) to different categories of venues. For example, by rating places according to size (airflow and ventilation factors), density (average space between people) and time (how long people typically reside in the space). This can lead to more pragmatic and sustainable (yet safe) policies governing social and business operations.

“More shockingly,” the authors say, “70% of those infected by the Superspreaders do not spread the virus to anyone else.” Of the remaining 30%, they typically infect 2–3 others, who themselves fall under the 30% rule.

Since Superspreaders cause the greatest number of infectious (and quickly), and that they are proportionally very rare within the population, the fact that 2nd degree infections are usually well contained accounts for why the distribution of cases is not evenly spread throughout the population. This means that new policies needn’t be “universal,” but, instead, localized.

In order to beneficially employ these new paradigms, we must migrate away from the older models we’ve been using, an effort that will require considerable communication and management.

Interestingly, Economists studying this have come to the same conclusion, but not from a medical perspective, as the epidemiological study had. They just simply observed maps of where outbreaks occurred and said, “this is what happened there, and thus, we could have done that.” But their analysis is only retrospective and based on observations, so it’s harder to convince others on the case for specific policy changes. The new scientific understanding of how the virus behaves should bolster their arguments.

On screening. While it had already been known that COVID-2 patients are most highly contagious when they are presymptomatic, it was still assumed — incorrectly — that fever indicates a worse infection, and therefore, a greater risk of infecting others. This is why many places screen for fever today, such as airports, shopping centers, etc. The new research shows exactly the opposite is the case for the coronavirus.

While fever is an indicator of infection, it’s not a primary indicator. In fact, of all indicators, it’s the least useful because fever begins long after infected individuals reached their peak infectiousness (and during the period when they are less likely to infect others anyway). True, “less” is not “zero,” and if the goal is to keep out all infected people, regardless of what they are infected with (even non-contagious people), then that’s different. But if the goal is to screen for SARS-CoV-2, testing for fever creates a moral hazard: If people are more likely to attend venues that test, it will increase turnout, and thereby expose more healthy people to a larger number of potential carriers who are most infectious because they got past the screen.

On viral and antibody tests. As a matter of public policy, these are absolutely essential, and everyone should be advised to be tested periodically (or under particular conditions, such as exhibition of symptoms or other indicia of risk to public health). There have been proposals for the government to pay for all viral and antibody tests, administered “periodically” as needed to monitor public health. Initial at-home test kits had unusually high rates of false positives and negatives, or simply didn’t work at all, and the FDA has had to remove many from the market. Policy should further mandate that both publicly-financed and commercially available tests meet the current standards of accuracy.

On analyzing test results. This is where the picture gets muddy, given the state of both the science and the practice of testing. Even with relatively high error rates, it doesn’t mean the results aren’t worthless. The more important issue is what is done with results, which is where this new research should modify people’s understanding about themselves and others. In particular, tests won’t distinguish Superspreaders from others who have the virus. And just because one tests positive for the virus, this doesn’t immediately prescribe a treatment plan. Much more information is needed for that, which is also another topic entirely.

The same is true for antibody tests. There is growing evidence that antibodies do confer some level of immunity, but “some level” is vague, with much still yet to learn. In a recent letter to his readers, Peter Attia MD wrote:

“Pre-existing immunity to SARS-CoV-2 is a topic being explored actively and, should it exist, one hypothesis is that it is mediated by previous exposure to other beta coronaviruses with the potential for cross-reactivity with the coronavirus of interest today. A study published in Cell this week found memory T cells in people who had not been infected by SARS-CoV-2 which were nevertheless (in vitro) capable of recognizing the SARS-CoV-2 virus.”

This suggests that antibody test kits should actually test for each of the entire coronavirus family, so we can learn more about the levels of all antibodies out there. As is speculated, we may learn more about cross-reactivity and partial immunity to SARS-CoV-2. Might it be that kids are less affected because they’ve had more recent exposures (as a general rule) to the common cold, which is among the coronavirus family? (Wouldn’t it be humorously ironic if the actual vaccine treatment for SARS-CoV-2 was just to catch a common cold and stay home for a week? No, please don’t start a rumor about that.)

[Feb 18, 2021 Update: New research shows this is NOT the case. ]

Other probing questions include: How much antibody load is needed to protect against Superspreaders vs. an infected non-superspreader? How long will immunity last? If the virus mutates, will older antibodies provide full or partial immunity (akin to flu vaccines)? Only time and further analysis will answer these and other essential questions.

Perhaps the most important fact to communicate about all test results is that they allow for better disease tracking, progression, transmission vectors, contact tracing, and identification of hot-spots. Yes, even error-prone tests can be valuable, though they will widen the error-bars on statistics a bit. Not a bad trade-off if the entire population were tested with a kit that had an error rate of +/- 5%. That’s incredibly valuable from an epidemiological perspective, even though it’d be useless for determining individual treatment.

On shelter-in-place orders. There’s no question it has worked because the primary goal has been to help people avoid the three-C venues. But again, these policies have been a blunt instrument that has had a profound effect on the economic and institutional infrastructure of the country, as explained in the article cited earlier. In short, environments that meet the three-Cs are the riskiest (schools, most workplaces, prisons, churches, and other indoor venues where people congregate for extended periods of time). But avoiding these venues is not the whole story about shelter-in-place.

Outdoor settings present the least risk of exposure, and this NYTimes article from May 5 (ancient by today’s standards) provides a concise review of how and why outdoor settings are safe (which includes paper citations showing evidence of little risk). Even without that research, mere observation makes the case self-evident: For months, millions of people who have been routinely exercising outdoors, going to beaches and parks, among other activities, are not getting infected, nor are they infecting others. If they were, we’d be seeing this data show up in case reports, and we’re not. The data and contact tracing show that infections have all been linked to those who attended events in three-C venues, who then take it home and infect family members.

Indeed, this week (June 6) will put this theory to the test. Today is the end of a week of daily demonstrations across the country from the death of George Floyd. And these demonstrations are expected to continue for days or weeks to come. As for people’s risk of acquiring the coronavirus, Wired Magazine reports this:

“With so many people yelling, chanting, and coughing in close proximity, the protests will almost certainly set off devastating new chains of contagion in the coming weeks. (Let’s not forget that the virus thrives in jails and prisons, and this week the police arrested more than 10,000 people.)”

The above paragraph makes assertions based on the older model in our understanding of the virus. The key errors are:

  1. Risk of infection is much lower in outdoor settings than was previously assumed. These demonstrations are happening outside, not in three-C venues, such as prisons and the other examples the author cites.
  2. The useful definition of “close proximity” between people is now better understood than it was before. Demonstrators are constantly moving around, not stationary, nor in proximity to one another as in prisons or other three-C venues. A series of aerial photos of demonstrations around the world illustrates that most individuals are actually much farther apart from one another when viewed up close than would appear from a further distance or when shot using a telephoto lens. (People appearing closer together in photographs than they actually are is the result of an optical effect called lens compression.)
  3. Another error in the Wired magazine piece is the failure to take account of duration-of-exposure. Remember, infectivity is a product of proximity and time, combined. Given that most protesters are typically outside the radius of infection from others, and they’re certainly not near each other for extended periods of time as they would be in prisons and other three-C venues, where people are in tight conditions for hours on end.
  4. Most demonstrators are wearing masks, which the author doesn’t mention.
  5. The article states that 10,000 people have been arrested, and the paragraph can be interpreted to imply that those people will go to jails and spread the virus. But among those arrested, most are booked and released.

So, will we see massive infections rise due to these demonstrates? My crystal ball is at the cleaners, so I can’t predict what will happen, but the new “Superspreader model” would suggest there shouldn’t be many outbreaks attributable to these events, but this also happens to be a time when the economy is also starting to open up , and caseloads have already been spiking across the country, and are only now becoming visible.

[ July 4 update: This NYTimes article reports that there have been no new infections that resulted from these demonstrations. It further supports the theories presented here: That outdoor infections are exceedingly rare. ]

In a polarized society, there will be those who will say the new cases are from the demonstrations, and those who will say it’s from states opening up too soon. To properly assess the reality, healthcare workers will need to interview individuals to see if they were involved in protests, whether they were wearing masks, how long were they in attendance, and most importantly, did they congregate inside interior venues before or after a given protest? With whom?

As a matter of daily lifestyle choices, more research on outdoor venues continues to confirm the safety of outdoor settings. This NYTimes article discusses the same three-C venues (it correctly calls it four-Cs), and mentions a study about an outbreak in China at a Buddhist temple in Ningbo, in Zhejiang Province, in January.

According to the article, “Some 300 people were at the service, which lasted two and a half hours and included lunch. It was held outdoors, and most of the worshipers were not infected. And of the 30 people who were infected, most had traveled on the bus to the temple and back with the first person who became ill, about an hour’s drive each way.”

On Public Communication. Communication from the media is essential, but in a polarized political environment, determining what is “true” or “false” is already difficult as new science and epidemiological findings evolve. Indeed, the strength of the authors’ misplaced confidence in this Wired article demonstrates how easy it is to miss new information and advice, who invariably left readers misinformed. And this, despite the fact that the article appeared in a major media publication, and published as I was writing this.

On Face-coverings. No one disputes the efficacy of face-masks in limiting risk in general, and therefore, public policy should encourage their use. And, most people agree that face coverings in three-C environments should remain mandated. It’s the more open spaces where things get complicated.

Remember, setting policies is more than just recommending the “right” thing to do, because the greatest risk is that people won’t comply with them. The challenge in writing policies is that people need to believe in their credibility, and the policies must also be regarded as pragmatic — that they can be complied with with reasonable efforts. If so, compliance rises. But if policies appear arbitrary, capricious, misinformed, or frankly, impractical, compliance drops, and with it, the perceived legitimacy of the authorities that drafted them. This can lead to an even greater risk: social disruption and disorder, which lead to tribal behaviors, as we’re already seeing in many communities.

A good point of discussion on policy-making is San Francisco’s May 28 policy “mandating face coverings when outside and when […] within 30 feet” of other people. The rationale is understandable, as they put it: “to provide simple rules” that are easier for the masses to understand and comply with. But “understanding” and “easy compliance” risks being trumped by its transparently obvious illogic: A 30-foot separation between people is unnecessary for protection. Again, we shall see about how well and how long people comply, and the social disruption may take place as individuals begin to splinter into non-compliance.

As we’ve seen with beaches and parks after a few months of quarantine, people will ultimately behave in ways that serve their instinctual needs and empirical observations. People’s own experiences being outside safely in these open air environments for so long without consequence contributed to the swell of pressure on municipalities to lift these restrictions. And the authorities eventually complied, albeit reluctantly.

It’s not that cities want to be draconian, or that they are uninformed about risks. The real worry about policy-setting is the slippery slope argument: That if people started going out to safe places like parks and beaches, it’ll be hard to keep them from going elsewhere that they shouldn’t, and that is what will get them infected.

On Workspaces and School. There’s no question policy-setting is a challenging task where you can never please everyone, and this will be put to its most difficult test for the two most important places: the workplace and school. The two are intertwined because most people can’t go to work unless school or childcare is available for their kids. So the school problem needs to be addressed first, which is an even more difficult task than workplaces. Schools cannot be converted to outdoor settings, or refactored to be safer spaces the same way officeplaces can. And frankly, even if they could, kids are harder to control and to keep in compliance than adults are.

I don’t particularly envision a solution to the school challenge anytime soon, and neither can anyone else. This NYTimes editorial lays out all of the challenges (and dysfunctions) in detail. As the article points out, the proposed changes schools are considering are largely impractical, and school systems simply don’t have the money to do anything else. Even with all that aside, the primary reason schools can’t open is (sing along with me now): You can’t screen for Superspreaders. As a fellow parent told me, “the moment a kid gets infected, everyone will scatter out of school like roaches on a kitchen floor when the lights are turned on.” This would put us back to where we are now: remote learning.

Almost all of this will apply to the traditional work environment, as well as many other interior spaces (social and otherwise), which comprise 90% of people’s lives.

In summary, until there’s a vaccine, we must continue with research and be comfortable with an iterative approach to solutions, open to continual revisions and refinements, all while maintaining compliance with policies that are designed to optimize health and safety. No policy is perfect, nor is every scientific study. Many questions still remain unanswered, and we’re likely to have a very bumpy ride. Until there’s a vaccine, the best recommendation is what my mother used to tell me when I got bored in the house: Go outside and play.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store