COVID19: Test, test, test…. But Why?

by Raywat Deonandan, PhD
Epidemiologist & Associate Professor
University of Ottawa
(I add my credentials to these COVID-19 blog posts in case they get shared. I want readers to know that my opinion is supposedly an educated and informed one)

Here we are, about two weeks into isolation. Has it been two weeks? I could look up the actual date of this country’s retreat to housebound safety, but I think it might depress me. I’d rather enjoy the haze of uncertainty, the arguably more natural state of literally not knowing the date. My passage of time is marked now only by inconstant circadian rhythms, the Prime Minister’s ritualistic morning briefing, and my dog’s need to go outside to do his business.

While I’m privileged enough to have a comfortable home and stable income, I’m uncomfortably aware of the widespread discomfort, anxiety, and financial stress being experienced by so many other people. So before I get to today’s topic, I want to ask those of us who can afford it to continue to pay those whose services we’re presently not using. For example, we continue to pay our dog-walker, even though we’ve asked him not to come over. I look at it like we’re paying him not to come. He’s still providing a service, just a different one.

In my last post, I briefly laid out an argument for using testing as a path back to normalcy. I’ve also been making the media rounds of late, pushing that message. Here are some links:

Today I’d like to explain a bit more about the role of testing in preventing and assuaging a pandemic.

Why Do We Test?

A) The Clinical Lens

Testing is done for at least three reasons, from at least three different lenses. When viewed through a clinical lens, we test in order to identify who needs treatment.

This is where most people in the general public begin. I want to know if I have the disease so I can get the cure! Why won’t they test me when I have some of the symptoms?

Now that we are all isolated at home, consider the prognostic outcome of a test. If you think you have COVID19, then you should stay home and isolate. You can use your government’s online self-assessment tool to help make that determination. In Canada, there is a federal tool, and many provinces, like Ontario, have their own too, too.

What if you were to get tested and you tested positive? Well, you would be told to stay home and isolate.

And if you tested negative? Again, you be told to stay home and isolate.

So from a strictly clinical perspective, testing is mostly useful if a patient presents with serious symptoms requiring medical intervention. Then doctors need to accurately know his disease state in order to render effective treatment.

B) The Public Health Lens

Through a public health lens, on the other hand, testing has one primary role: to enable us to identify who has the disease so that we can quarantine them and and trace their contacts. In this sense, testing is a tool used to slow or halt the spread of the outbreak, rather than to aid in treatment.

With most people in housebound isolation, at least here in Ontario, this is slightly less of a concern since those who have the disease are presumably not spreading it to others. But of course that is not completely true. In a long term care facility, for example, shared facilities or a caregiver might prove to be the conduit of transmission.

So in this phase of the pandemic, with most people in isolation, the public health value of testing is in identifying those key members of the community most likely to spread to many people: emergency responders, child care workers, cashiers, drivers, etc. Our finite number of tests is best strategically applied to these people, in addition to the clinical cases noted above.

C) The Health Systems Lens

Third, when viewed through a health systems lens, the value of testing is in planning and preparation. Ideally, health care workers would be tested repeatedly, as they have the most contact with infectious individuals, are therefore more likely to infect multiple people, and are themselves invaluable in the ongoing effort to beat back the pandemic.

Moreover, testing gives us a sense of the population burden of the disease, better enabling us to judge where on the epidemiology incidence curve we presently sit, and how long it will be before a staged return to normalcy is possible.

So Why Can’t I Get Tested?

You’ve probably heard about rich celebrities in the USA getting tested on demand. We don’t have that capacity in Canada, as all of our tests are used sparingly and strategically. Frankly, we just don’t have enough of them…. yet.

As noted above, there is not a lot of clinical utility in testing every dry cough, not when the prescription will not change, and not while we are in such short supply of testing kits.

Here in Ottawa there is a priority in testing those who have traveled recently and are showing symptoms. This is because these people have a much higher probability of being infected, and therefore were recently (and still are) infectious. Quarantine and contact tracing must begin at once.

There is lesser utility in testing those with lower probability of infection, again given the limited number of tests available. It makes sense, though, to lower the symptom threshold for testing in the case of health care workers or others who are likely to become “super spreaders” of the disease.

So if you suspect you are infected, consult the online self-assessment tool and then phone your local public health COVID hotline.

But Wouldn’t Testing Tell Us How Bad Things Are?

Yes.  Absolutely. That’s the plan. Iceland did things right. In addition to deploying their testing for clinical and public purposes, as described above, Iceland did a battery of random testing of the general population.

The key to their strategy was to conduct representative sampling, which means utilizing statistical techniques to sample groups of Icelanders from whose data larger inferences about the greater population could be reliably made. As a result, they have a very good idea about the disease’s incidence, prevalence, geographical spread, and the extent of asymptomatic disease.

Such information allows them to more strategically focus on regions needing more intense public health scrutiny and isolation, and to direct limited resources in a much more efficient manner.

Could We Have Avoided All This Chaos With Testing?

Oh yes. As noted in my previous post, pandemics can be divided into two phases. We’re in the second phase now, that of exponential growth. But in the first phase, which is characterized by a smattering of cases arriving on our shores, testing could have obviated our need for a complete lock-down.

I don’t like second guessing our nation’s public health efforts. But this is 2020, and I must resist the urge to make yet another “hindsight” joke. When China issued their social isolation policies in late January, we should have immediately strengthened our public health infrastructure with special focus on large scale testing capacity.

Imagine a month or two ago, with the pandemic wave approaching, a scenario in which public officials were poised and empowered to respond with the full force of the law. Testing mildly symptomatic individuals would have revealed cases early on. Aggressive contact tracing could have identified additional cases (since community spread would not have occurred). And then quarantine, or limited neighbourhood or regional isolation could have been enacted to prevent national contamination.

What Kinds of Tests Are Available?

I’m a Population Health Epidemiologist, not a laboratory scientist, so please forgive my lack of expert knowledge on this front. But the general categories of tests that are possible include: (1) nasal swabs that are sent to a lab where chemical voodoo is performed to detect the presence of the virus or its DNA; (2) blood tests or cheek swabs that are sent to a lab or an on-site mobile device to detect genetic markers of the virus’s presence; and (3) a serology test, also done on blood, to determine if antibodies to the SARS-CoV2 virus are present, indicating that the person is or was infected.

It’s the first type that is widely available, but the kits are in short supply and the public health labs are overtaxed. In Ontario, the reagents needed to conduct the test are in such short supply that university research labs were being asked to donate their supplies to the effort. Test outcome takes hours. But realistically, due to shipping and high demand, it is now taking days (in some cases more than a week) to get results back.

Abbot Labs just announced the production of tens of thousands of rapid on-site nasal swab tests, though, which should be a total game-changer, for data realization reasons described further below. Presumably it can render an outcome in 5 minutes.

The technology for the DNA test is here. Spartan Biosciences in Ottawa has an on-site reader that uses cheek swabs and that can render a result in 30 minutes.

India has created a very inexpensive test that brings down their lab testing costs and timing from 8.5 hours to 2 hours.

The last category of test, the serology or antibody test, is of enormous interest. If you could know if you’ve already been infected and have recovered, then presumably you could return to society in full confidence. When deployed widely, such a test would tell us the true extent of past infection, as well as how close we actually are to population herd immunity (which I described in this post.)

While the CDC website states that they are looking for such a test, it seems that several are already on the way. The British company Mologic has a prototype that is being rushed into service.

Do keep in mind that the online self-assessment tools are a kind of test, though not a physical one. Data scientists from the University of Toronto are using its parameters to estimate a “heat map” of likely COVID19 cases in Canada. At some point, I would like to see the sensitivity/specificity data for the self-assessment tools. That will tell us how accurate they really are.

What About Pooling of Tests?

Several people have contacted me to explore the idea of pooling tests. Instead of waiting 8 hours for each of 4 tests, why not test them as a batch, like in the famous mathematics riddle of weighing various items?

The challenge there is in not diluting a possible positive test and thus rendering a false negative. But India’s Principal Government Scientist indicated that that country is considering this very path.

How Does The Current Testing Regime Affect Our View?

Here in Canada, as of this date, we are limited to the nasal swab and public health labs. We also don’t have a lot of testing kits (yet), though that might change soon. So remember that the so-called “epidemiology curve” of cumulative cases that you see reported on the news every day represents poor data collected from a very imperfect testing regime.

Consider Ontario’s cumulative incidence curve (extracted from this source).:


Cumulative incidence of COVID19 cases in Ontario as of Mar 29 2020

It’s a little worrying at first glimpse because we went into “social” (now called “physical”) distancing around March 23. And yet the cases continue to grow exponentially since then. Is the distancing not working?

Well, keep in mind that incidence data are a lagging indicator. The incubation period for this disease has a median of 5.1 days.  So the cases being measured now are likely due to exposures that took place just as the isolation took effect, 5-6 days ago.

Moreover, due to delays in getting test results back (remember those backlogged labs with insufficient reagents?) it’s entirely possible that we are seeing in large part the mounting cases of exposures that took place a week or more ago.

In other words, the data will not show the success of physical distancing for another week or two. So expect a continued climb in cases for another two weeks or so, then a bending of that curve as the effects of physical distancing begin to manifest. It will look like magic.

So What Then?

When the curve flattens, then falls, and we proceed to state that I describe as a “simmering boil” of a handful of new cases every few days, we can begin our staged re-entry into society.

Testing will be critical then. The serology test will tell us who among the vulnerable is safe to return. Those who are not vulnerable can join them.

The full deployment of an awesome nation-wide testing regime will follow. Not just once, but multiple times. If you are someone with frequent contact with the public, you will need to be tested frequently, perhaps every day.

Before any medical intervention, a test will be required. It’s conceivable that entry to a crowded event might necessitate a negative test result or proof of immunity.

This will be a strange but interesting new world in our immediate future.

But it will all be made possible by the flotilla of new testing solutions being cranked out by the innovation centres of our civilization’s biosciences sector.

As I wrote earlier, an army of nerds is coming to save us all.