In 2009, an Australian epidemiologist who had moved to Canada (giving me certain affinity with him), Professor Robert Smith? captured some interest by considering the best methods by which the world could ward off a zombie outbreak. Zombies are, to the best of our knowledge, infected humans who are technically dead but, what is relevant here, are still infectious; passing the problem to others by biting them. The resulting mathematical model showed that the way to deal with a zombie outbreak is to stamp it out quickly in one big, determined push. This yielded some support from Neil Ferguson who would go on to co-author the influential Imperial report for the COVID-19 crisis.
My understanding of zombie biology is that if you manage to decapitate a zombie then it's dead forever. So perhaps they are being a little over-pessimistic when they conclude that zombies might take over a city in three or four days.
My understanding from this is that Ferguson thought that lopping a zombie’s head off was a pretty straightforward approach and that was the end of the matter. Nonetheless, the US Center for Disease Control put up a website to inform the public just in case.
Viral pandemics are similar to zombies in two respects. First, they feast off humans and second, they move from human to human. Where they differ is that viruses can be killed proactively (through anti-viral treatments) or passively by eventually either killing or dying out in a host. (Zombies can only be killed with a headshot.) The problem with viruses, as was already discussed in the previous chapter, is knowledge regarding who is and has been infected. With zombies it is plain as day.
The upshot of this relatively invisible enemy is that the management strategy for a viral pandemic is likely to be on-going and its end hard to measure or be sure of. To be sure, in principle, there are actions that could be taken to suppress COVID-19 in one go. Moreover, suppression requires 100 percent success and, unless knowledge is available very quickly, the type of containment strategy one might use for a zombie situation is not warranted. This means that pandemic management requires an approach that will have to be actively conducted over many months. It will not be comfortably over in the course of a two-hour movie.
In this chapter, I will examine re-emergence strategies following initial isolation actions taken with regard to a pandemic. As has already been noted, isolation is economically and socially costly and, as of the time of writing, it is reasonable to suspect that there will be immense political pressures to plan a path from isolation to re-emergence while managing the pandemic and preventing future un-contained outbreaks. The reason it is important to have a clear strategy is that, in its absence, there may be pressures to let the virus ‘burn’ through the population creating herd immunity even with a massive loss of life.
The cat is out of the bag
Before considering a re-emergence strategy, it is worth reflecting on why that strategy is needed. The basic epidemiological goal is to move the basic reproduction number of the infection (R0) to a point where it is less than one. In that situation, even an unmanaged virus will end up being contained and be on a path to dying out. As already emphasized, the issue is knowledge. If we know whether you HAVE or HAD COVID-19, and presuming that gives on-going immunity, we can use individual specific interventions to achieve that goal.
The benchmark is to consider what we could have done if we had that requisite knowledge from the outset. In that situation, we could continue to test and isolate those who tested positive. To be sure, this would not be fun for those so identified but it is better to isolate justifiably than indiscriminately.
The problem is that such knowledge is not available to even the most alert pandemic response teams. This means that the cat is out of the bag and, thus, the virus is spreading making it even harder to acquire the requisite knowledge. Indeed, it is precisely because of this that most countries (starting with China around Wuhan) had to pursue a widespread isolation or lockdown. This will eventually get the rate of infection down, then question is: then what? After all, so long as there are still infected people somewhere, the basic reproduction number itself will not be below one (as too small a share of the population is immune) and the virus will likely re-emerge.
The answer is that when the cat is out of the bag, first you have to put it back in the bag and then, you are starting from square one. However, this time around the hope is that you have the capability to acquire knowledge to manage the spread of the virus in round two. In other words, all of the initial actions to contain the spread of the virus (including flattening the curve) are about getting to a point where you can have a ‘do over.’ That means evolving to a testing economy as described in the previous chapter. The question for re-emergence is: can we achieve a ‘do over’ while at the same time allowing a relaxation of the policies surrounding initial containment?
Who’s on first?
My starting point is a country that has engaged in a widespread lockdown and has started to see signs that the rate of infection is starting to fall. At this point, it may be possible to predict when the infection rate falls below a level that would cause it to break out again. And before that point, there may be opportunities for a targeted and measured release of people from isolation.
The question is: who first? The lockdown procedures in most places made a distinction between essential and non-essential workers. Essential workers were in health care, key public services and food supply. For the most part, essential was a dictate with regard to the value of their jobs rather than considerations of whether they were in jobs that may facilitate a faster spread of the virus. Indeed, health care workers are a case in point and hospitals, where possible, have put in place measures to stop the spread of the virus within those workplaces and beyond (with those workers being housed away from their homes in some cases).
The harder decisions will be for the non-essential workers. There are two criteria that would guide this choice based on their economic value and their potential to spread the virus. Let’s begin with economic value. According to some studies, about 34 percent of the US workforce are in jobs that permit them to work from home. If you, for instance, work outdoors or work with specialized machinery, it is not likely you can work from home. But if you are doing legal work, computer or mathematical work or management, you do not actually need to be near anyone else to do your job. Thus, on an economic basis, it is reasonable to expect that construction and manufacturing workers will rank highly as candidates for movement back from isolation.
The second criteria is on the basis of what types of jobs would potentially pose a lower risk of spreading the virus. In network theory, the issue is how connected people are to a broader network. In modern societies, you can draw links between almost any two people. Those links are sometimes direct but usually indirect. A virus can potentially spread along those links which is why it is commonplace to see outbreak data given on a country or maybe state basis. There are links between countries but as fewer people travel between them, the ‘distance’ along the network (in terms of number of people along a path) can potentially be greater.
But for re-emergence we are starting from a situation where we have already broken the network. For a ‘stay in place’ lockdown, this is almost to the level of individual households. Those households are components who link with each other. Every household has some members who venture out to obtain food or health care and so, even though they are weaker, there are some links across households by this mechanism. When we take a household and allow members of that household to return to work, we are increasing the number of links between households.
You might think that means that if we allow one member of a household to return to work, we should allow other members to also do so, as that household will be integrated into the main economy with just one connection, so why not have two. However, unless the members are literally going to the same place outside of the household, additional household members being released magnifies the problem. Person A goes to their workplace, comes back potentially infected which infects the household and also Person B who is going somewhere else. Having more links outside embeds the components (in this case, households) more densely in the network which is what you want to avoid. This suggests that where possible at most one member of the household should be, initially, able to return to work.
This logic also explains why it is difficult to, say, remove an entire region from lockdown. In doing this, all of the households in a region become one component and so if there is any incidence of the virus remaining it will spread throughout. By contrast, if you take a more targeted release, even within a region, you can keep the basic reproduction factor low.
The related point is that hubs need to be limited. A hub is a single location (which may, of course, also be a transportation conduit) where many people interact. A candidate here is the central business district of a city but it also applies to schools and commuter colleges. The problem here is that if you open up a type of jobs, say, law firms, then, if they are located in the same place, you create the potential to spread the virus more quickly. This carries over to workplaces in general. To keep the virus from spreading again, you have to limit the number of people in any given location. This means that workplaces need to be open but at a low scale. If they cannot operate in that way, it may be better for those workplaces to move to the back of the queue. This logic certainly applies to schools and colleges that are hubs for interactions and also places where it is hard to use mitigating interventions such as good hygiene practices. It also almost certainly means that public events — especially social ones — like sporting events, concerts, conferences, and elections will not be able to take place as per normal for some time.
This analysis suggests that amongst the first people to be allowed to return to work following isolation will be a subset of those people who cannot work from home. That subset will be determined by how to ensure that if there are connections between what might be otherwise isolated groups (or components) that those connections are sparse (meaning one or a very few connections per person). Of course, this can be modified depending on the ability to use methods (preventative gear and cleaning) to ensure a lack of spread in workplaces and on transportation conduits. At the time of writing, it is difficult to say who that set would be. However, one suspects construction and manufacturing will be high on the list while schools and colleges may well be low. Schools will perhaps be the greatest challenge given their social importance. (Not to mention their role in general parental mental health).
Finally, it is useful to consider what might happen with regard to travel — not locally but regionally and inter-nationally. It is tempting to think that keeping bans on travel in place would be a natural policy. After all, they keep the virus contained within countries. However, as trade and some travel is occurring, while it might be slower, there is potential for the virus to leak through those boundaries. Given this and the economic importance of some travel, there seems no reason to single those jobs out for continued isolation. Instead, I suspect that, at least through airports, there are more opportunities for testing (assuming fast tests are available) and also for the use of preventative gear. Airports are already places where people have experience in dealing with frictions. The additional frictions that might be required may be relatively cheap from that perspective.
The above discussion focuses entirely on reducing the spread of the virus throughout re-emergence. Another possibility is to focus on allowing re-emergence subject to the constraint that those who are more vulnerable remain isolated from others. This is akin to the proposed policy of isolating the elderly and others with identifiably compromised immune systems who are most at risk from hospitalization or worse from COVID-19. As a policy to introduce from scratch when people are not practicing social distancing, targeted isolation appeared difficult to achieve in practice. However, for re-emergence, we may have more confidence that the virus is free in certain places. Therefore, as we allow movement to re-emerge, we can continue to keep people where there are identifiably higher risks isolated, only allowing connections with certain precautions in place. If this were possible, then that would allow a greater number of people outside of those groups to be able to operate more freely. Nonetheless, it is unclear at this stage whether we have the requisite knowledge to confidently pursue this approach of using ‘not at risk’ as a criterion for targeted release.
The Big Rationing
All going well, re-emergence can be achieved and allow a testing economy to develop that allows social and economic life to return to ‘normal.’ Of course, the best way to end the pandemic and its management is to develop and then distribute a vaccine. COVID-19 is novel and, to date, vaccines for coronaviruses have not been developed because, prior to this one, they did not pose a grave health risk. It is for this reason that developing a vaccine that can deal with COVID-19 is a significant challenge. As a result of the protocols governing the development of new vaccines, it will be at least a year before one that is safe to distribute is available.
Even in that somewhat happy instance, there will remain a significant problem: any production process for the vaccine will take time and, so, at any given point, there will be a shortage of vaccines. This means that the available supply will have to be rationed. Absent the use of a market approach to sorting out who gets what when, that requires that a decision on the order of recipients be made.
This is a problem that has been forecast and government authorities have issued guidelines for rationing. For instance, the US CDC has five tiers of recipients for a vaccine. There are two dimensions of priority — occupational groups (reflecting the earlier economic criteria) and high-risk populations (reflecting the at-risk criteria for harm from the virus). In Tier 1, there are the occupational groups who are already priorities for non-isolation during the containment phase today including health care and security services. Tier 2 continues to include essential workers especially with regard to infrastructure services while the remainder of those are part of Tier 3. Using the at-risk criteria, Tier 1 includes pregnant women and babies, Tier 2 expands that to high risk children and people who work with young children, Tier 3 includes the rest of children, Tier 4 is high risk adults and the rest is part of Tier 5. Notice that there are no occupational groups in of Tiers 4 and 5.
For COVID-19, the occupational group ordering makes sense and is in line with current practices. However, the at-risk ordering does not reflect what is currently understood about COVID-19; that is, that the older you are the more at-risk you are. The CDC guidelines take age into account when considering children. However, for COVID-19, it appears that the younger you are the less at-risk you are (although children may be vectors for disease spread). In other words, the guidelines for influenza in general do not reflect the realities of risk with respect to COVID-19. Thus, one would expect those to change. However, this will also generate a decision regarding healthy adults of working age and those who have retired. The latter are far more at-risk while for the former, there are economic criteria that will favor them. My point is to highlight this potential issue and suggest that there will be no easy decision in this regard.
What the criteria also do not reflect is any sense of network theory. For instance, prison populations are potentially risky areas where infections can break out. If testing was not available, there are arguments that they should receive priority for a vaccine.
Even looking beyond the use of guidelines to assign priority, there will be a large pool of people for whom there is a vaccine shortage but no identifiable way of prioritizing them. In that case, a lottery will likely be used (as it was in the movie Contagion). In the end, what we should anticipate is a very uncomfortable process that few will likely forget.