Welcome to Plugging the Gap (my email newsletter about Covid-19 and its economics). In case you don’t know me, I’m an economist and professor at the University of Toronto. I have written lots of books including, most recently, on Covid-19. You can follow me on Twitter (@joshgans) or subscribe to this email newsletter here. (I am also part of the CDL Rapid Screening Consortium. The views expressed here are my own and should not be taken as representing organisations I work for.)
For 18 months, Canada had a stockpile of rapid antigen tests that was hard to get out of the door. Part of this was restrictions on their use but they relaxed over time. The other part is there wasn’t demand. Between vaccinations and Covid rates falling, people weren’t demanding them. A few thousand businesses and some schools were testing people regularly but that wasn’t enough to use up supply. But given the forecasts of potential outbreaks, we knew that the stockpile was going to disappear overnight if there was a big outbreak. In many countries across the world, governments failed to order enough tests and the free market lagged behind.
We can lament that but rapid tests are either going to be expensive or in short supply for the foreseeable future. This raises the question: when is a situation “test-worthy”? (Reference to Seinfeld intended). I get asked this question all of the time these days and so I’ll offer some thoughts. These are not thoughts about the use of tests by health institutions or by governments. Instead, it is about the use of tests for managing your own lives.
A broad principle
Tests are information. They tell you whether you are infected/infectious or not. With current levels of Omicron, if a rapid test result is positive, you should treat yourself as infected — if there were concerns about false positives, current prevalence means that those concerns are minimised. If a rapid test result is negative, you are less likely to be infected but, frankly, if you have any other indications that you might have Covid (symptoms, exposure to known cases, cases in the household), you have to take this with a grain of salt.
The question is: how useful is that information? Suppose, as the vast majority of people are, that you have been vaccinated. It is unlikely that you, yourself will need treatment in a hospital. The data seems to indicate that the issues are at flu-like levels of concern for most people. The older you are the more concerned you should be but that is the same with the flu. In other words, unless you are older, unvaccinated, or have some known issue that puts you at higher risk, knowing whether you are infected or not doesn’t offer you any options in managing your own health. In other words, for most people, knowing whether you have Covid or not is, at the moment, not useful for treatment (something that could change with other variants or if there are treatments worth taking).
However, the information could be useful in terms of managing the risk you pose to others. If you know you have Covid, then you will want to isolate to stop harming other people. But in that case what you want to know is not whether your are infected but whether you are infectious.
Here’s our go-to diagram from The Pandemic Information Solution.
After almost two years of struggling to explain the difference between being infected and being infectious, health professions at least seem to get it now. The above diagram shows the difference. You are more likely to test positive on any sort of Covid test if the viral load at a point in time is higher. For the PCR test, you don’t need much of the virus at all to come up positive. It is so sensitive, it can detect dead virus in you which is the case for about two-thirds of the time you would test positive in that way. In other words, if you want to know you are infectious, PCR tests have a very high false-positive rate.
For rapid antigen tests, they have a higher threshold. The good news is that means that a positive result on an antigen test means that you are more likely to be infectious than a positive result on a PCR test. That makes rapid tests the go-to test to inform whether you are potentially riskier to other people which is the key information needed for an isolation decision.
But notice that typically the antigen threshold is still below the threshold for infectiousness. That means that it too has a false positive rate for infectiousness. And that may be significant. There is talk about people testing positive on antigen tests 10-12 days after the onset of symptoms and many days past the symptoms going away. With the whole struggle to get medical experts to see rapid tests as the right way to go, this issue was de-emphasised. But it is there. Even rapid tests are, in a sense, “too good.”
Well, I need to qualify that statement. We don’t really know. Back in the day (2020) when rapid tests were struggling for regulatory approval they were being pitted against PCR tests. So the incentives were to approve rapid tests with a threshold closer to PCR tests without any recourse to the threshold for infectiousness. Actually, we don’t really know precisely what that latter threshold is. It is reasonable to put some weight on the notion that rapid tests are “too good” but there is a chance they may not be. More studies are needed.
In summary, the broad principle is to use rapid tests to manage your risk towards others.
Criteria 1: Are you making an isolation decision?
If you are like me and you are working from home and have no good reason or desire to go out, then knowing whether you have Covid or not offers no assistance in protecting others. You are already doing that.
If you need to go out for work or other matters and that is what you would do, then a rapid test with a positive result can tell you not to do that.
In other words, if you are not currently effectively isolating, then a test is useful. Otherwise, it is not causing you to change what you do so you don’t need a test.
Criteria 2: If possible, concentrate repeated tests in one person
As I anticipated the supply shortages, I had a stock of about 30 rapid tests at home. That has diminished as I have handed them out to others who were worried they had Covid. But we have four people currently in our household. Those could disappear quickly.
But as Omicron is so infectious, there is really no great way to prevent within household spread and, in our situation, we aren’t really going to try. I suspect many others are in the same boat. In that case, it is all for one and one for all. That means that if we know one person is infected or infectious, we can presume we all are. There are a few days margin of error but that is it.
In other similar situations, there are strategies for pooled testing whereby tests are rationing by testing up to 25 people at the same time. Something like this might be possible to rapid tests but you need swabs for each person and those are in the same short supply as the tests. So this isn’t really going to be practical.
So here is another potential rationing strategy based on a similar principle. Take the first person with symptoms and test them daily for a week. Chances are they will be negative for a few days and then turn positive. But they may also be negative throughout. If they continue to be negative switch to another symptomatic person and repeat.
What you want to do is test someone repeatedly so that you can monitor their viral load. The tests themselves don’t do that although if you look really carefully at a T-line you can see variation that does seem to relate to viral load. But that is not really a reliable thing to do without proper measurement. Instead, you can monitor that person’s current status. That will help inform you of other people’s status without them taking tests.
Criteria 3: Who comes into contact with vulnerable people?
Above I suggested doing this on the first person with symptoms. But you may want to select your household test subject based on other criteria. For instance, you might want to test someone who comes into contact with vulnerable people. If they are negative on the day they have to do that, it will help minimise risk. That’s a good strategy regardless but if you are rationing tests and suspect someone else in the household has Covid, it becomes even more useful to do so then.
Criteria 4: Who is the vector in your household?
A final criterion for your test subject is the person you suspect is most likely to bring Covid into the house. For us, that will be the person going to school every day. Interestingly that person will also be the person most likely to harm others. So testing them is a win-win.
How many tests?
So how many tests do you need to have on hand? Well if you need to monitor a person who might otherwise infect others outside of your household you will need 7 at least for them.
But you can also expect to want some more for false alarms and because you just can’t help yourself and will want to know if the headache followed by scratchy throat followed by runny nose followed by dry cough followed by sore back is Covid rather than the “it’s just a cold” either you are or someone else in your household is claiming even though you’d really like to know if you have Covid so you can write a newsletter about it and not be left wondering if you really had Covid last month or not and have your spouse roll their eyes every time you claim you did because you didn’t go anyway other than the drug store and you just sit on your butt all day so how would you pick it up when she didn’t get sick or at least that is a situation I’m familiar with.