How Ontario’s COVID-19 testing debacle has cost lives speaks with epidemiologist David Fisman about what the province has gotten wrong with its COVID-19 testing so far — and what Premier Doug Ford has gotten right
By John Michael McGrath - Published on Apr 09, 2020
During a press conference yesterday, Premier Doug Ford referred to Ontario’s testing levels as “unacceptable.” (Frank Gunn/CP)



A visibly frustrated Premier Doug Ford said on Wednesday that Ontario’s COVID-19 testing numbers were “absolutely unacceptable.” This came after Ontario’s daily testing levels had fallen to 3,000 or below — about 10,000 fewer than the stated capacity. On Thursday morning, the health ministry announced new expanded testing criteria that will likely see more tests administered.

One of the more prominent critics of the province’s testing regime so far has been David Fisman, a professor of epidemiology at the University of Toronto’s Dalla Lana School of Public Health. Fisman appeared on’s #ONpoli podcast last week and agreed to a follow-up interview on Tuesday, before the premier’s remarks on Wednesday. You’ve been critical on social media about Ontario’s rate of COVID-19 testing so far. What’s wrong with what Ontario is doing?

David Fisman: I think there have been two things wrong. One has been — I don’t know what the inside scoop is, but I think we’re starting to learn — that folks didn’t really anticipate how much testing they’d need to do and didn’t anticipate the need to radically scale up in order to keep up with this epidemic. 

A man filming in The Agenda studio

Our journalism depends on you.

You can count on TVO to cover the stories others don’t—to fill the gaps in the ever-changing media landscape. But we can’t do this without you.

The second issue has been the failure, in my view, to temporize quickly in order to track this disease, given that there were problems with testing. So, you know, the lab couldn’t do the throughput required. Okay, that’s a problem. They gradually solved that, but as has been pointed out by many people, Quebec has been subject to all the same constraints as Ontario, and they’ve moved quite a bit quicker than Ontario did to download testing from central public-health labs to hospitals. There have been solutions: it’s a crisis situation, so moving to alternate test kits when their preferred kits ran out was something they moved to eventually. Again, I think all of this could have been done with more urgency.

If you can’t test for a disease like this, you can also do what we did during SARS, which is using probable case definitions or clinical case definitions. We know what this disease looks like on average, or we think we do anyway, so when testing isn’t working, you can take a page from the book in Hubei, in China. When their testing wasn’t keeping up, they just went straight to clinical case definitions. In the Chinese context, that meant if you have a diagnosis of COVID-19 based on your CT scan — they seem not to fear radiation to the same degree we do; they would just fire people through the CT scanner — then you have COVID-19, and we’ll add you to the count, because we want to manage you so you don’t spread COVID-19 to others. That has nothing to do with lab-test availability. In fact, during the SARS debacle, we didn’t know what caused SARS until May, and the epidemic started in March in Toronto, and we made it through March and April with clinical definitions. It’s just basically a failure to adapt to the battlefield as the battlefield changes that has left us really flat-footed. On this issue of insisting on lab-test results, one thing I’ve been told is that you want to get a proper test confirmation, a lab confirmation, before you do any treatment. But is there any real treatment for COVID-19 at this point?

Fisman: Not that I know of. You can put someone on a ventilator because they aren’t oxygenating anymore. There’s a bunch of experimental therapies, and I think a lot of the old-time Victorian approaches to medicine — like treating people with convalescent serum — look like they’re starting to emerge as possibilities, but, no, it’s not like we have this vast array of therapeutics available that are waiting for lab confirmation. I mean, it’s a public-health crisis, and you want to identify people correctly for a variety of reasons. 

The most painfully stupid part of all of this in Ontario has been this desire to go down with the ship. We do have a probable case definition in Ontario — which, to the best of my knowledge, hasn’t been revised — which still included travel or contact with a known case of COVID-19. That, to me, just demonstrates an absolutely painful failure to change your tools as needed as the epidemiology evolves. 

The [integrated Public Health Information System] case data, they’ve got a public component to that: the three modes of acquisition still say “travel,” “contact with a case,” or “neither.” We don’t even have community-acquired cases in the official definition. You’re a neither! It’s bewildering to me. 

I know a lot people who work at Public Health Ontario, and I think a lot of them don’t like me much anymore, but I know them, and I know they’re very competent. I look at this state of affairs, and I think “gobsmacked” is a good word. I just don’t understand it. You’ve mentioned what Quebec has done. What else should Ontario be doing instead of what we’re doing now?

Fisman: We apparently have a capacity to do 13,000 tests a day, and we’re doing 3,000. The question is, why in the name of sanity are we not using that capacity?

I gave infectious-disease rounds at U of T today, and one of the audience members piped up and said, “You know, there’s a committee looking at changing the testing guidelines,” because, right now, the guidelines are so restrictive that unless you’re basically a health-care worker who’s sick or you’re sick enough to be hospitalized, nobody’s going to test you. Which is a real problem. 

So there are a couple of things. One is we need a meaningful probable case definition; we need to be counting those cases. The other is, if we have that test capacity, let’s get on it. This is a week now the backlog has been worn down, tested down, and it’s been a week now we haven’t been testing at levels we’re capable of. It’s great that they’re testing for clinical care. They should also be casting a very wide net to find cases. If they’re really trying to control this by finding cases and isolating them, you shouldn’t have guidance like what’s on the web tool that says, “You have a new cough, a new fever, but you don’t have known contact with a COVID-19 case, and you haven’t travelled to Wuhan, so just stay in your house for 14 days.” You should want to find that individual. 

There’s some risk created by telling that individual to come out of their house and come to a testing centre, but we’ve also seen innovation for months now from the United Kingdom, from places in Ontario that have set up mobile testing. 

You could imagine, if we have this tremendous test capacity, you don’t want people flooding test centres and giving each other COVID-19. But you could set up mobile testing so people who know what they’re doing and have personal protective items could go house to house and test people. There are lots of ways to scale this. Test everyone who’s being admitted to hospital — whether they’re coming in to give birth, whether they broke their leg, whether they’ve got kidney failure — test them for COVID-19 because you don’t want to miss people. We do this now for people with MRSA: Why on Earth would you not do that for COVID-19 if you have the capacity? Test health-care and chronic-care workers regularly so that we don’t have these awful outbreaks. There’s a lot we could do. You’ve mentioned the high rate of positives we’re getting — I take it because of the low level of testing. Why is that a problem?

Fisman: I think it’s a problem because it’s contrary to [World Health Organization] guidance, and it’s WHO guidance because what they’re saying is, if you’re that high, you’re not testing enough. You’re looking for the disease where you know you’re going to find it. For the people in ICU beds on ventilators, it’s appropriate to know that they do have COVID -19 and not some other disease, but that doesn’t really help tell us what COVID-19 is doing in the community. So you need to be casting a much wider net. Again, it doesn’t have to be all testing, if you have some meaningful clinical case definition, not all those people are going to need treatment, but you should want to find them. You want to find the hot spots, rather than saying you don’t want to find the hot spots, because it’s inconvenient and it creates bother. That’s not how you deal with an outbreak. 

You should be casting your net wide … It’s the old “looking for your dropped earring under the streetlamp thing”: you know you’re going to see it there, so you look there. We’re not looking all over the place, including in the dark, and we need to test more. I’m trying to get a sense of where this should be placed on the public’s list of concerns in terms of the government’s response to COVID-19. Is this, for example, comparable to elected leaders deciding not to order more draconian shutdowns of public businesses?

Fisman: I think this is and is not a problem. I’m actually relatively bearish on how much you can do by identifying cases and isolating them. Even if you’re doing a full-court press, you miss too many cases. 

Could Doug Ford have closed things down a week earlier? Sure, he could have. But, given the advice he was likely getting, and given how challenging it is politically, I think he did really well in terms of the shutdown.

I’ve said it many times: I think the control group for us is New York, and we’re not in the shape New York is in. We should be very grateful, because there’s no reason we shouldn’t be, except that Ford instituted social distancing, physical distancing, and Eileen de Villa did, too, early enough. 

That said, my biggest concern right now in terms of under-testing relates to institutional outbreaks. This is a bug that shows it likes to take off in places where people are gathered together, and we’re seeing that happen in long-term-care homes, in hospitals, and in prisons. 

So, to me, the surveillance and infection-control components to testing are extremely important, and the places you most want to up that testing are any sort of institution where people are crowded together. You want to be testing the workers with some frequency so they’re not giving anything to residents and patients, and you want to be testing both residents and patients liberally, especially on entry to the facility. Given some of the really tragic death tolls we’re seeing out of some long-term-care homes, in particular … I don’t want to be alarmist, but the only way to ask the question is to ask it: Has the lack of testing cost lives?

Fisman: Of course it has. Of course it has. Because how else do you know where you have infected workers? I think the reason this has moved into long-term care like wildfire, and they recognized this quite early in British Columbia, is the economics of it. The people who work in long-term-care facilities — personal-support workers and care workers — are doing very hard work that is underpaid. A lot of them don’t have benefits. A lot of them have multiple jobs to feed their families. So you’ve set up a structure that is guaranteed to move this disease between multiple facilities as efficiently as possible. My understanding of what they’ve done in British Columbia is that they’ve given those people a degree of economic security, they’ve upped their pay by $4 an hour, something like that, and they’ve made people full-time at individual facilities. I’m sure the long-term-care home operators won’t like that, but, if you don’t do that, you’ve created a network the disease can move on. Last Friday, the government published its projections, its modelling for what the next few weeks will look like in terms of COVID-19 cases and deaths. Given the issues with the testing, how confident are you in the projections of any model, including the one you’ve made?

Fisman: I think it’s challenging. We’re doing the best we can with the data we have. We’ve seen some positive signs in terms of the Ontario projections. And it will be interesting if they do scale up testing — it will look a little bit worse for a bit. But, right now, the way it looks is that Ontario’s control is starting to improve, as much as we can say that from the limited data we have. And that’s been a very welcome sign. And we also see some signs of that in the absence of an overflow in the intensive-care units. 

So, again, I think it goes back to what Ford did on March 14, in terms of starting to shut things down. That didn’t depend on testing. That’s a mechanism for epidemic control that doesn’t depend on being able to actually see the epidemic. I think where testing is going to be important is protecting institutions: shelters, long-term-care homes, hospitals, and also in terms of trying to figure out if we’re beating this thing so we feel good about climbing down, because at some point — it’s like the bear chases you up the tree, and now you’re safe in the tree, and that’s great because maybe the bear can’t get you, but at some point, you want to get out of the tree. The threshold we’ve all been talking about is ventilator-bed capacity, in the hopes that we don’t get to the point where there are more people needing ventilators than there are available ventilators. As someone who’s done some of this modelling, you’ve looked at the province’s model published last week —

Fisman: I don’t know if there is such a thing. I made some of those slides. And another fellow I know made some of the other slides. And my model wasn’t a model: it’s basically a forecast. But we made things worse and then we made things better because we didn’t have the data to calibrate to. 

I’m not sure the province actually has a model — I think that was a little bit of political theatre, but it was nice political theatre, because I think it focused people. And I think the qualitative take-home that things could have been worse — they’re now okay but potentially bad, and we want to make it better — you know, I realize people like to have numbers for stuff like that, but, at the end of the day, that was a qualitative exercise. And I agree with it: we do want to do better.

In terms of the ICUs, we’re following it day by day. What you would expect if social distancing is going to work is that this is the time we’d see some flattening, and we’re seeing some flattening. Every day that passes without us having a second surge from people returning from Florida is a good day. So we’ll see what happens this week. As of two days ago, we had 216 people with COVID-19 in ICUs in Ontario. We had thought we had about 400 ventilator beds to work with, so we were at 50 per cent full or so; because of the reductions in elective surgery, actually ICU capacity is up by another 200 beds in Ontario, which is an unexpected silver lining to shutting things down. We still have room; we don’t have to make these horrible decisions they’re making in other places because their ICUs are full.

And, again, I’d circle back — and I know I keep saying it, and I know it’s kind of weird as someone who, probably two months ago, would have been enraged if someone had suggested I was a Doug Ford supporter — but Doug Ford: the guy did it, you know? I have a lot of issues with the public-health response at the provincial level. I don’t have any issues with Doug Ford. He’s not a public-health physician; he’s not an epidemiologist. He’s dependent on advice, and I think some of his advice has been bad. But he’s made the important decisions right, and he deserves a lot of credit. Because you’ve mentioned it a few times now, and because I keep seeing it on social media: I think one of the few real errors so far that people could point to was when Ford said people should still go on their March break trips. How serious do you think that was?

Fisman: I think we’ll see. It seems to have not knocked us over yet. Would one have preferred that the guidance was a bit stronger on March 12? Yeah, I would have. I think we would have. But perhaps we’re going to get away with it. This week will tell us. We’re headed to three weeks out from the return. We’ll see.  

This interview has been condensed and edited for length and clarity.

Related tags:
Thinking of your experience with, how likely are you to recommend to a friend or colleague?
Not at all Likely
Extremely Likely

Most recent in Coronavirus