Re-opening across the United States has been an exercise in irrational optimism about the infectiousness of COVID-19. COVID-19 remains a highly infectious virus capable of expontial growth, it is just as infectious as it was 6 months ago. Early in the outbreak in New York State both new infections and deaths were doubling every 3 to 7 days. In Florida during a June and early July, cases and deaths doubling every 5 to 10 days.
When we reduce contact between individuals through economic shutdowns, social distancing, and mask wearing, new infections decline, when we re-open the economy and resume various forms of human interactions, infections increase. This law is pretty immutable as long as we continue to live in a country where we neither have COVID-19 under tight control nor do we have herd immunity.
A handful of countries have achieved tight control of the virus using the containment strategy of aggressive testing, containment and quarantine. Taiwan, New Zealand, Iceland, South Korea and a few others have been able reduce infections to a low enough level that health officials can closely monitor those who are infected and the people they came into contact with for signs of the virus. By closely monitoring a small number of people, the larger population is able to live with fewer restrictions. The United States failed to adopt an effective containment strategy at the outbreak here in the US. The country also failed to attempt containment during the brief nadir of infections in June. Lacking a national strategy individual states could have attempted tight control of statewide infections when they had very low infections rates. But both states and the country as a whole, skipped that step in a rush to re-open the economy.
How much re-opening can we safely afford?
Many metrics have emerged since the pandemic began that attempt to quantify the level of risk from the virus in a country or region. I find the information provided by Covid Act Now to be useful – see figure 1. This rates states based on 5 metrics and provides a color coded overall rating for each state.
For the purposes of calculating if we can “afford” to further open the social and economic activity in a region, the single most important measure is the effective reproduction rate of the virus sometimes referred to at R(t). It estimates how many secondary infections are likely to occur from a single infection in a specific area. Values over 1.0 mean we should expect more cases in that area, values under 1.0 mean we should expect fewer. This measure R(t) is the real world version of the theoretical R-naught R(0). R-naught is the theoretical reproduction rate for a virus. The R(0) for COVID-19 has been estimated to be in the range of 2.5 to 5.0. Which means that each new case of COVID-19 will, on average infect 2.5 to 5.0 more people. R(0) does not take into account things like social customs, efforts to contain the virus nor any accumulated immunity. The website in figure 1 includes as one of its 5 metrics an “Infection Rate” which is a measure of R(t). Figure 2 shows a screen shot from another website that only estimates R(t).
Figure 2 shows the R(t) for each of the states. As you can see they all hover around 1. A R(t) of one means on average each person infected with COVID-19 will infect 1 new person with COVID-19. At an R(t) of 1 the number of infections will remain more or less level, as each new case will on average infect one new case. Measurement of R(t) is not straight forward. There are differences between the values presented by the site in Figure 1 and Figure 2. Since we do not have a true measure of the number of new infections occurring in the community, these sites rely on the number of tests conducted, the number of positive tests and the test positivity rate to estimate R(t).
We should consider the degree to which R(t) is below 1 as our possible budget for re-opening. Figure 2 shows 19 states with an R(t) at or below 1. It is important to remember, R(t) tells us nothing about the current number of infections, hospitalizations or deaths. It just tells us if infections are currently expanding or contracting. So Figure 1 shows three of the hardest hit states in recent weeks: Texas, Florida and Arizona as having an R(t) at or below one. With the number of infections high and hospitals at or near capacity in these states, it is probably not a good time to consider further increases in social and economic activities that result in increase human contact. But if R(t) remains below 1, these other problems should diminish over the coming few weeks.
A point in time measure of R(t) is helpful, but looking at R(t) trends is even more helpful. Figure 3 shows a graph of Florida’s R(t) over time. Like all states, during the early phase of the outbreak the R(t) was quite high. Once public health steps were put in place and people became more cautious, the R(t) declined rapidly, even before shelter in place orders in Florida. R(t) was rising in Florida even before the shelter in place orders were ended, probably due to social distancing fatigue and visitors and segments of the population not feeling the virus was as dangerous as initially represented.
By June 1st, R(t) in Florida had risen to 1.36. Nothing like the theoretical R-naught of COVID which is in the 2-5 range. But enough to trigger a doubling of new cases every few days between early June and early July.
How limited is our opening up “Budget”?
Massachusetts has gone through 3 phases of a 4 phase re-opening plan which is summarized below in figure 4. Figure 5 shows the Massachusetts estimated R(t) over time. During shelter in place period shown on figure 5, the R(t) in my state declined significantly from well above 1 to well below 1. The decline continued for a while after phase I was carried out on May 18th. The Mass R(t) was already rising back toward 1 when phase 2 was executed on June 9th. By the time phase 3 was put into place, Massachusetts had an estimated R(t) above 1.
|Phase||Date||Example of allowed activities|
|1||May 18||Worship, Construction, Hair Salons and Pet Grooming (by appt only), Curbside retail.|
|2||June 9||Retail stores – limited capacity, Restaurants – outside dinning only, hotels, beer gardens, golf|
|3||July 6||Gyms and fitness centers (40% capacity), casinos, wedding venues (no bars), museums|
It seems pretty clear that the cost Massachusetts Phase 3 reopening was beyond Massachusetts “budget” for reopening. Its also likely that even Phase 2 was too expensive, as the trend was already upward, and had exceeded 1.0 well before phase 3 was started. It appears that R(t) is sensitive to very small changes is the amount of social and economic activity. As a Massachusetts resident, I noticed that that there was a considerable relaxation of social distancing in my area late May and early July. People had started having small back yard gatherings with friends. In my neighborhood groups of teenagers we meeting in the parking lot and play ground of a local school to just hang out. So the changes in R(t) are not only effected by official policy, but also by small changes in individual attitude toward social distancing.
What can be buy based on our R(t) budget?
Of course it is impossible to precisely quantify the impact of each type of social and economic activity in terms of R(t). But its clear, that even small changes have an impact. It is also clear that there is a delay between official changes in policy, and its eventual impact of R(t).
Shortly after ending the Shelter in Place guidance in Massachusetts, the R(t) in that state reached a low of .74. The end of Shelter in Place, and the limited re-opening activities in Phase one were enough to change the trajectory of R(t) from declining to inceasing. Combined with a general decline in the public’s level of social distancing in private activities, represented the full amount of Massachusetts’ re-opening budget of .26 R(t).
Similarly, in Florida’s data (Figure 3) we see that the R(t) decline actually ended during Florida’s sheltering period, likely due to decreased social distances among visitors and residents, even before formal re-opening steps were initiated.
Even for people who question the threat level posed by COVID-19, R(t) is an important measure. Given the US failure to contain the virus earlier in the pandemic I have wondered if a Herd immunity strategy like Sweden appears to be pursuing might make sense even here in the US. The essence of the strategy is that we may find it less disruptive to allow the infection to progress through the segments of the population less likely to have adverse outcomes so that we can achieve a level of immunity to protect the entire community. See my blog Lack of a coherent plan endangers social fabric for more details on this strategy.
Advocates of letting the infection run its course point out that hospital care has improved, we have some treatments, and that the overall death rate has therefor declined. So any sane approach to herd immunity would need to ensure that the health care system did not get overwhelmed by the volume of sick people needing hospitalization. Remember that R(t) does not indicate the volume of new infections, but rather the trend in new infections. Those who want to minimize risk while waiting for vaccines or new treatments, want to have a low level of new infections and maintain that level by keeping R(t) near 1. Those who want to get it over quickly, want a high level of new infections, near or just below the sustainable capacity of the health care system. The goal would be to keep new infections at that level, which means keeping R(t) as close very close to 1. A sudden increase of R(t) while running near health system capacity could result in a rapid doubling in cases as seen in NY early on and Florida and Texas in June, which could quickly result in a situation were people are dying in Hospital parking lots for lack of the ICU care needed to save lives.
How to increase you Re-opening Budget?
We can see a strong association between increase social economic human interactions and an increase is COVID-19 infection rate. Are there things that can be done to offset some of these increases? Public health experts are constantly calling for actions that all potentially impact R(t) and hence our ability to re-open our society.
- More testing, with rapid test results
- Contact Tracking
- Higher percentage of population wearing masks
- Ventilation in public spaces
- UV Light
- Hand washing
Testing all by itself does not work to control the infection rate, but rapidly getting testing results back to infected citizens can be an important part of keeping infection rates down. There are two assumptions built into the recommendation. The first is that most infection from one person to another happens during the first few days to one week of becoming infected. This is pretty common with most viral infections and limited research on COVID-19 indicate it is true with this virus. Ideally we would have instant tests, so that patient’s could be told immediately that they should go home and remain isolated. Test results that take multiple days to be receive greatly reduce the value of testing in terms of reducing R(t), because the infected person may have already spread the infection to others by the time they learn that they tested positive. The second assumption is that if people know they are infected, they are much more likely to follow social distancing and isolation recommendations.
Contact Tracing and Isolation are a key element in the classic control of Epidemics. City officials in Medieval Europe, with limited understanding of disease, realized that the isolation of people who were possibly infected was an effective way of protecting their community. With COVID-19 contact tracing and isolation can be built on top of a timely and effective testing system. By quickly identifying people who are infected, we can then quickly identify people who were in contact with the infected person and ask them to self isolate. Doing so contributes to a reductions in R(t). Adequate testing, with fast results combined with rapid and effective contact tracing and isolation could greatly add to our R(t) budget for increased social and economic activity.
Mask wearing is now widely understood to reduce the risk of transmitted an infection to other people. It is generally not thought of as an effective way of protections one’s self from infection. The exact amount of protection provided by mask wearing is not yet known and has become partisan issue all by itself in the United States. But undoubtedly it contributes to the spread of infection via a number of ways. First there is the physical barrier that will reduce the number of droplets with infection projected out of an infected person’s mouth and nose. Secondly, there is the social signalling that wearing a mask sends to other people. When we see someone wearing a mask, we assume they are worried about getting infected. Unless we have a political agenda to promote, we generally respect that person’s views, and likely keep a greater distance when speaking or passing in the street or store aisle.
Recently there has been an increase concern with aerosolization of COVID-19. Aerosolization represents smaller particles given off by infected people that can hang in the air for some time. Initial control of COVID-19 infections focused more on droplets from infected people, these larger particles from infected people are more likely to be caught by masks, and are heavier and therefor hang in the air for a shorter period of time. Control of smaller particles can be addressed through increased ventilation, hepa air filters and a type of UV light that can destroy the virus. These mitigation efforts deployed in indoor public spaces would likely decrease the cost in R(t) for some types of indoor activities and thereby allow us to buy more safe social and economic human interaction.
And of course no discussion of infection control is complete without mentioning hand washing. Routine washing of hands, especially if your have been in contact with other people can contribute to a reduction in R(t), potentially keeping ourselves safe, and buying society more safe social and economic human interaction.
Who decides what we spend it on?
One of the problems with deciding how to spend our R(t) budgets is that its not an individual decision. It has to be made on a community wide basis. I may want schools and libraries to open and not care at all about malls and pet grooming services, but not everyone else would agree with me. I could imagine a clever version of Aristophanes’ play Acharnaian where individuals get to make their own private deal with COVID-19, but in our current reality we must make these decisions across individuals and apply them to towns, states or the entire nation. This means that elected and appointed public figures will be making the decisions and not everyone will agree with them. But in this public health emergency we must follow the guidelines provided by our elected officials and appointed public health professions for the short term. Disagreements should be addressed via elections, not by armed men on state house steps.
Regardless how you perceive the COVID-19 outbreak, R(t) is an important measure for knowing if we can open up more, or need to tighten social and economic activity. For the foreseeable future we will need to keep an eye on it!