Safely living with and avoiding the coronavirus is likely to become the major consideration for the next some months until such time as a vaccine is released and available for all (possibly 12 – 18 – 24 months in the future, assuming one is possible), or when a quick/easy/inexpensive (and safe) to administer cure for people with the disease is identified (maybe next month, maybe next year, perhaps never). We discuss these two outcomes in the first part of our two part series.
How to live with the virus threat is already a pressing issue. Now that some countries are starting to perceive they are beating the virus with their social distancing measures, their people are growing restless and eager to be able to move about freely – to return to work, to travel on crowded buses and metro trains, to go to restaurants, bars, concerts and sporting events, send their children back to school, start flying again, and to do everything else which we’re currently limiting or banning outright.
But to relax controls now and return back to normal would be absolutely the wrong thing to do. The virus is still out there, and the minute we give it opportunities to start spreading again, it absolutely will. How tragic it would be if after accepting the sacrifices we’ve made so far, we “give back” all the advantage we’ve struggled to win, and find ourselves right at the terrible square one again, complete with even more depleted medical supplies, exhausted medical services, and an economy less able to come up with another multi-trillion dollar round of very imperfect compensation measures.
The Need for a Nuanced Response
It is time to move from the panicked initial response of shutting down almost everything for almost everyone. We need to develop and implement a more sophisticated approach that balances the desire (and need) for as much of our lives to return back to as close to normal as possible, with the conflicting essential mandate not to allow the virus to start flooding back into our communities again.
It is possible, by selectively controlling/restricting the highest risk behaviors, to allow a substantial return of lower risk behaviors. This also has a positive “feedback loop” – the lower the viral incidence in the community, the safer all activities become.
We see hints of how this can succeed already in a few countries such as South Korea, where their initial rush of cases was staunched and they are now experiencing a mere 75 or so new cases each day with a balanced rather than brutal approach to ongoing social distancing.
South Korea has a population of 51 million, so their daily new case count would be the equivalent of less than 500 new cases a day in the United States (currently we’re suffering over 25,000 known new cases a day). If we assume a fatality rate of 1%, this suggests 5 fatalities a day in the US. In exchange for this, we get the convenience of living a much more normal life. Compare these 5 fatalities to the 115 fatalities a day we accept in exchange for the convenience of automobiles, or the 164 fatalities a day for regular ‘flu and pneumonia, many hundreds a day as a result of social evils such as drinking and smoking, and the number seems like it may be acceptable. Certainly not desirable, but acceptable as one of life’s many trade-offs.
On the other hand, a similar type of gentle response in Japan seems to have become insufficient. After a long period of under 50 new cases a day, there are alarming signs that perhaps Japan was a bit too permissive, with almost a week of new cases greater than 200 a day and last Sunday hitting an all-time high of 515. Japan is now considering further restrictions.
The Japanese example, echoed in other Asian countries, shows a precarious balance between too many and too few social restrictions, and hints at another problem of finding balance – there is a lag of some days between any change in distancing requirements and the impact as measured by new cases, and eventually by new deaths, too. It is necessary to have a “fine grained” system that can quickly be tweaked in small and non-impactful ways, and regionally rather than nationally, and without the need for extended public debate each time a change is made. The public would very quickly tire of major national or state-wide announcements of new restrictions being imposed or lifted, then reimposed, and so on, especially if in largest part such actions did not relate to the situation in the local area.
Our approach to date has understandably been urgent and (sort of) all-encompassing. It seems that in some US states (most notably, Washington, the original “epicenter”) it is already starting to work, as it has internationally too. So we can start to selectively allow the return of some activities, while still standing fast with bans on others.
Let’s Work Smarter, Not Harder, with Our Restrictions
We must adopt a “work smarter, not harder” approach to the virus and avoiding infection. We need to recognize the difference in people and the risks they pose to themselves and others.
At the one extreme, there is no reason why a person who has developed immunity to the virus after having been infected and recovered should have any movement restrictions at all. At the other extreme, a person who is infected needs to be absolutely isolated during the course of their infection.
Then, in the middle, there are a series of more risky or less risky categories of people. The higher the risk a person carries (to themselves and to others) the more restrictive a set of conditions should be applied.
There are a number of ways these people can be divided between the two extremes of immune and infected. While the fewest number of categories might seem simplest, we suggest four groups in the middle so as to allow for easy and subtle shifts in policies and controls as needed to maintain the balance between freedoms in general, and freedom from the virus in particular.
So, from least risk to highest risk we would have, with color coding :
- “Gold” status – Immune, therefore unrestricted
- “Green” status – Very recently tested as uninfected, no risky behaviors subsequently, almost certainly free of infection
- “White” status – Reasonably recently tested as uninfected, only low-risk behaviors subsequently, most likely free of infection
- “Orange” status – Status unknown or expired test result, an accumulation of risky behaviors to the point they may be infected
- “Red” status – Possibly infected (perhaps as a result of close contact with someone who shortly thereafter was found to be infected), needs testing as soon as possible, should be isolated prior to testing
- “Black” status – Definitely infected, must be isolated
This series of categories allows for decisions to be changed from time to time as to what amount of risk is acceptable for a person to be shown as green, then at what point they become white, and at what further degree of risk they become orange, and eventually red. If case counts are growing, people are more rapidly downgraded in status. If case counts are dropping, then greater tolerance for higher risk activities can be allowed.
Note also that while it would seem that a person’s status invariably goes from low risk to higher risk before being “reset” back to green with a fresh test, a person who had been at home with no symptoms and no outside world contacts could see their status move back up the scale towards white or even green.
The magic to making this concept work is knowing, reasonably exactly, and in a real-time manner, which group everyone of us in the US belongs to, being able to clearly differentiate each person accordingly, give them differing sets of movement/socializing privileges, and then real-time track their risks while in public. This is easier than you think.
Personal Identifying Location Tracking Devices
Such an all-encompassing concept of knowing where everyone is might seem impossible, but it is actually quite easy; indeed, it is close to already being in place, for most of us, most of the time.
Everyone would need to have their own personal identifying location tracking device that they’d keep with them at all times. This device would display that person’s current status for others to see, while reporting the person’s location back to a central database.
Now, before you start zoning out and considering this an idiotic and impractical suggestion, let me give another name for this “personal identifying location tracking device”. I’m simply referring to your phone. Because that is exactly what it is. Your phone can tell astonishingly exactly where you are – sometimes to less than one foot (last week an advertising agency executive was boasting to me how he knew where people’s phones are to within an inch). Phone companies already have all this information, and companies like Google and Facebook (as well as, for sure, government-type security agencies) are recording it at present.
How It Would Work
Knowing where you are, and knowing where everyone else is at the same time, means that if someone is subsequently found to be infected, the people they have been close to can be automatically and quickly identified, and all such affected people simply get messages through a special app on their phone, and their status (if not immune) changes to “Possibly infected, needs testing”.
The app, while storing a lot of data, would show a simple color code as well as perhaps offer additional details in text about a person’s status. A main screen would show a picture of the person, perhaps their name, a bar code so the data could be validated and/or read by a scanner or camera, and a color square for their status, making it easy for people to use that information as an ID and to display their status.
The weakness of this type of testing is getting everyone to register and participate. A person without a cell phone would be “electronically invisible”, and could roam at will, without people realizing. Maybe we could create a new social norm that people have their cellphones dangled around their neck, or mounted in special pockets on their clothing, with their screen facing out. That might seem impossible, but think of how many people have ID badges around their necks already. If we are currently making mask wearing somewhere between recommended, desirable, and mandatory, it is clearly something else that could be done if we have the resolve to require it.
We know most phones have their screens auto-time off after a short period of inactivity, but this can easily be overridden by the status display program. Yes, requiring the phone to have its screen on all the time would see battery life shorten. But that’s a simple choice. You want to leave home, you charge your phone first and maybe take a battery recharger with you.
The entrances to major buildings, malls, supermarkets, etc, could be monitored by security or simple automatic systems that require people to show their phone and the status on it before they are allowed to enter – perhaps turnstiles a bit like with some metro stations and amusement park entrances.
Other controls could be associated with a person’s color code. Perhaps, while people were in public, anyone who was not “gold” (ie immune) would be required to wear a mask. Businesses might be allowed to set policies to allow green people also to go maskless, but other people would be required to wear a mask.
The “big data” being collected about everyone and our whereabouts, 24/7 is an enormous computational project. But large cloud computing companies such as Amazon, Google, IBM and Microsoft have that capability already. It just requires the programming to match it all together and calculate risks.
A risk is clearly dependent on how close two people are, and for how long they are close together, and perhaps further modified by if they are in a confined space (a car for example), a moderately confined space (a building) or outdoors. Some risks would cause a person’s status to immediately go to red, which would mean “no movement privileges until tested”. Others might cause a person’s status to go from green to white or orange, and with more low risk exposures, their status would slowly move to red.
The “behind the scenes” calculations for when a person’s color status changes could be varied at any time depending on if the number of virus cases was growing or stable or shrinking, and further changed based on the data received for the relative rates of risk for different activities, and other adjustments – for example, a person who traveled very little in an area with few or no know virus cases, would have their status stay green (after recently being tested) and white (for the reducing accuracy/relevance of a past test) for longer than a person in an area with more virus cases and with higher-risk crowding in with many other people.
The Need for Regular Testing
For this method to work, it is necessary to build up to a universal testing service where everyone is tested and can be retested regularly. This would be facilitated by the existence of abundant testing locations, tests that are not intrusive (ie simply a front of nose or not too far down the throat swab), and results that are reliable and almost immediate.
If people wish to assemble somewhere as a compressed group of many people densely packed together, all people with an orange, maybe even a white status, might be required to be tested before being admitted to the group location. This could be used as a pre-requisite for air travel and cruise embarkation, for example, and as an essential part of allowing people to enter into the country. Visitors would of course also have to load the monitoring app onto their phone and comply with all associated requirements during their stay.
Doing group tests – ten or more people at a time – would speed the testing process and allow it to scale for lots of people in a short time.
The way multiple-person simultaneous testing goes is you collect samples from, say, 20 (or 200 or any other practical number of) people. You mix them together and test. If none of the people are infected, the result will show this, and all the people can have their status reset back to green and be allowed to proceed. If even one of the people is infected, the first test will show that at least one person is infected. So the test would be repeated, with the people now separated into two equal groups. Perhaps the next set of results show one group is all clear, and they can go, and the remaining half of people would then be divided into two smaller groups, and so on.
We do not currently have the ability to test people at the volume required. This implementation would require most of the country to be tested before it could start (the thing would be to roll it out geographically). With 325 million people, and a need to test them all in a short period of time before the passing of time and random risk invalidates the early test results, that is an ambitious – but absolutely feasible – undertaking.
The key to getting this done is the same as we mentioned above. Test multiple people at a time. Particularly when infection rates are low, multiple-person testing can be amazingly efficient. Currently, with the US having a known infection rate of 0.1%, and an unknown “silent” infection rate of people who have the disease with no or very weak symptoms, it seems likely that most of the time, a random group of 100 low-risk people could be tested with all 100 being simultaneously cleared from the one test. A bad case scenario, say two people both infected in the batch of 100 (unlikely but possible), might take as many as 27 tests to find the two people. But that still means in total, 100 people were cleared with 27 tests rather than 100 individual tests – still almost four times more efficient than doing every person’s test individually.
Maybe in total group testing can be six times more efficient (it all depends on group size and the prevalence of positive people in the groups) which would mean if we boost the current test rate of 150,000 tests ten-fold to 1.5 million tests a day, and if each test in effect gives six results, then it would take just over a single month to test the entire population.
In reality, we see many people would probably need/want to be retested once a week or thereabouts, although the less prevalent the virus, the less often people need to be retested. We’d want an ongoing test capability for perhaps 10 – 25 million people a day – under 3 million tests/day with group testing. That’s a lot, but with the cost of our shutdown being trillions of dollars a month, who can say this isn’t worthwhile.
We do not currently have the software to manage this monitoring and realtime alerting/risk calculation, but it could probably be developed in only a few months, first to a rough and approximate level (see our section below about the system not needing to be perfect), and then subsequently to increasingly accurate and more helpful degrees. There are other apps in development or deployment that offer some parts of this functionality, but as far as we know this is the most extensive and integrated concept so far.
We Can’t Rely on Symptoms
Recent research has shown that people are most likely to pass their infection on to other people in the first few days after catching an infection themselves, and before they start to show any observable symptoms at all.
This is a key consideration. You can’t take a person’s temperature and ask them if they are feeling well, and use that to determine if they are safe or not. One of the worst things about this virus is that it is stealthily transferred from person to person while the infected person still feels normal and well, is still going about their lives normally, not taking special precautions, not confined to bed, and while the people around feel relaxed to be in contact with them.
This is not well understood, because many other infections only start to be passed on once symptoms appear. We still read about companies and organisations that are doing an interview/temperature test to decide if people are “safe” or not, and if they decide that a person is “safe”, they then allow risky behavior and close contact with others in the workplace or wherever the person has been admitted to. This is a substantial risk based on a misunderstanding of how and when the virus spreads.
Most of the passing on of the virus is likely to happen in the first three or four days, before any empirical testing makes the infection noticeable.
That is why we need to be able to medically test for viral load with a throat or nose swab – the interesting thing being that it is easiest to detect the virus with that type of medical test because it is when the strongest concentrations of viral particles are being breathed/coughed/sneezed out.
There are two problems with testing as it currently is. The first problem is that a standard test merely detects if a person is currently infected or not. It does not test if the person has been infected in the past and now has a sufficient level of developed immunity to be safe from the danger to be re-infected and to become a disease spreader again.
You might think that such a test is unnecessary – if a person has been detected with an infection, and then is cured, what further testing is necessary? That is a reasonably fair statement (although it too has some limitations) but it ignores the belief most experts hold that there are many people out there who have had mild doses of the Covid-19 infection without even realizing it. Those people might now be immune and qualify for “Gold” level status on the monitoring app.
If it is possible to simultaneously test “normal” people, not just to see if they currently have the disease, but also to see if they have had the disease in the past and are now immune, that would help boost up the number of people who could get the highest level unrestricted Gold level status, and make things easier for everyone, allowing more focus on people who have not yet had the disease.
It is difficult to develop an accurate test for immunity, and even harder to find one that would be quick and easy to perform in large numbers. There are several reasons for this, and they don’t seem to have easy solutions.
Such a test is also limited another way – we don’t yet know if a person’s immunity might fade over time, either “just because it does”, or because the virus is changing and becoming a new type of virus that the body needs to start all over again with (like regular ‘flu – which is a different type of virus).
So that means there are not many people who will get the highest “Gold” status, at least, not initially. That is regrettable, but it is not a problem. It just means some people will need to be tested more often than in theory they truly need. We expect that tests will get better at reporting past exposure and immunity in the future.
The other problem is that some of the currently used tests for infection are unreliable and imperfect. Some tests give high levels of false negative (saying a person does not have the disease when they actually do) and others give a high level of false positive (wrongly saying that a person does have the disease).
We expect that the accuracy of these tests will improve. But, and here’s the possibly surprising thing. Even if they have significant error rates, that is still okay. We don’t need 100% accurate testing. It would be desirable to have perhaps 75% accurate testing, but even 67% accurate testing would still help us wrestle this virus to the ground and keep it there.
The System Does Not Need to be Perfect
One commonly expressed concern is the amount of error that is unavoidably within this concept. The error starts with concerns about the accuracy of test results, and then goes on to more concerns about the accuracy of assessing the possible risk (or lack of risk) of being infected as a person moves about and comes into contact with other people. Some parts of the country have poor or no cell phone service, and even in areas with good coverage, the accuracy of location reporting often drops off a lot inside buildings.
There are partial fixes to some of these challenges. But for the parts that can’t easily be optimized, the good news is that as long as infection rates are kept very low (and even today, almost the entire country is averaging less than 1/10th of one percent infection rates – in other words, only one in every thousand people is infected) then if a few mistakes are made, the increase in risk is still very low. Inherent in seeking a suspension of total lockdown orders is a willingness to accept some small level of infection and death.
That’s the same as with automobiles. The benefit of being able to drive our own vehicles gives us an extraordinary degree of personal and economic freedom and efficiency. The tradeoff is that every day, about 115 people are killed in car crashes in the US, and many thousands more suffer injuries and/or financial loss. We’ve made balancing decisions for where we’ll draw the line with driving freedoms – for example, the acceptable level of intoxication before a person is adjudged to be too impaired to drive, the acceptable level of skill to get a driving license, and the acceptable level of speed on the road before the tradeoff between travel time saved and greater chance of injury/death goes beyond what is thought appropriate.
We must have a similar approach to controlling the virus. We can’t demand 100% safety and a guarantee of never getting infected. We’ll not say what an acceptable percentage is – it is a societal debate and decision for where the tradeoff should exist between the degree of social controls and economic harm on one side, and the amount of acceptable risk of catching the coronavirus on the other side. But whatever that number is agreed to be, this system can then be adjusted to keep ongoing cases at around that level.
As that social equation gradually evolves, the underlying equations that set a person’s status – green/white/orange/red – can be tweaked. Maybe a 10% chance of being infected means green, maybe it means white, or even orange. Plus, the risk elements can be adjusted, too. Maybe society believes that sports games are just “too risky”, but accepts that attending church services is a fair trade-off between freedom and risk. All of these things can be changed within the model’s calculations.
The Results of Even an Imperfect System
The purpose of this system is not to ensure that no-one gets the virus. As long as the virus is “out there”, people will catch it from time to time, and seeking a 100% perfect system of prevention becomes unrealistically expensive – and impossible. As long as someone, somewhere in the world, has the virus, there remains a chance that the person with the virus will start off a new chain of infection that leaks into the US and then runs amok here again, the same as it did in March.
The purpose is to reduce the number of people that a person with the virus can pass it on to, and to be able to quickly respond when a new case is detected. Currently it seems each newly infected person passes the virus on to about 2.5 more people.
If we say that on average there is about 4.5 days from when one person is infected to when they have passed their infection on, then in less than ten weeks, that one person will have created 1 million new infections. If we could reduce the infection rate from 2.5 down to half that – 1.25 new cases per infected person, then instead of one million new cases, we’re looking at 135 new cases in the same time period. That’s an enormous transformation.
Better still, if we could get it down to 0.83 new cases (a two thirds reduction, and now one third the original rate), then instead of one million, instead of 135, we’re now looking at 14 new cases, and no new cases continuing to appear.
So, that in simple terms is our choice. To bemoan an imperfect choice, and either allow each newly infected person to create a million new infections in less than ten weeks, or lock the entire country down each time a new case appears, at a probable cost of some trillions of dollars each time, and with less and less surviving economy after each episode.
Or – and this is our suggestion – to accept an imperfect choice but implement it as soon as possible, while continuing to work on improvements and enhancements to make it better. We could immediately reduce the one million new case count (ie the 2.5 transmission rate) down to perhaps 135 new cases per infected person just by halving the transmission rate. Then we would keep improving things further to get it down to less than a 1:1 transmission rate, at which point it becomes a decaying number that trends down to zero with each passing day, while allowing most of the country to be in a state of cautious freedom.
We need to embrace the positive in this and make it happen as quickly as we can, then continue to improve it. The great thing is that most of the behavioral and risk changes that will be tweaked and changed, probably every day until we get a better appreciation of things, will be “invisible” to people in general. They’ll just know how their color/status changes on their phone app, together with occasional update messages with general statements about high and low risk activities and recommended (or, rarely, mandatory) controls/restrictions.
Do We Truly Lose Privacy?
We dislike the thought of the government knowing where we are 24/7, and most other people probably feel that way too. But, two comments to modify this concern. The first is the government already knows this. As do many private companies, too.
Secondly, while we regret the loss of the imaginary privacy that we could pretend we had, we hate having to be confined at home, unable to work, and – even worse – our friends, colleagues, and customers also being confined in their homes, too.
Would we swap something that, in reality, we’ve already lost, for the ability for our business and personal lives to return safely back to nearly normal? In a heartbeat, you betcha.
The important thing, whether we actually are losing any privacy or not at present, is to ensure that once the virus finally is under control – either via a vaccine or a cure – that we then take the virus tracking app off our phones.
(This was the second part of a two part article series about solving the problem of the Covid-19 pandemic. Please read the first part of the series here.)
Published 8 April
Updated 9 April (new section “We Can’t Rely On Symptoms”