Eric Hoskins is a medical doctor, public health specialist and the former health minister for Ontario.
When I was collecting data for my PhD thesis in epidemiology at the University of Oxford, my professor would often pull out a favourite saying of his: “Garbage in, garbage out.” The message? Flaws in data-collection lead to faulty conclusions. He might say the same thing today, about what we’re facing with the COVID-19 numbers.
The Public Health Agency of Canada (PHAC) compiles data on COVID-19 cases and deaths, but it is woefully incomplete. Due mainly to inconsistencies in case-reporting, even the most basic information – age and gender – for half of confirmed cases and two-thirds of the deceased is not known. PHAC is also reporting a 7-per-cent crude case-fatality rate, calculated by dividing the number of deaths by the number of cases. This figure implies that for every 100 COVID-19-positive Canadians, seven will die. That number, which has major implications for how we proceed, is highly misleading for one simple reason: the denominator is deeply flawed.
Coronavirus guide: Updates and essential resources about the COVID-19 pandemic
What is the reopening plan in my province? A guide
How many coronavirus cases are there in Canada, by province, and worldwide? The latest maps and charts
The number of identified cases depends on how many tests are carried out, as well as who gets tested. For much of this epidemic, health care workers, the sickest and the most vulnerable have been prioritized for testing. With limited resources, this approach has been entirely appropriate. But it also means that many COVID-19 cases were missed – even more so since as many as 50 per cent of infections could have presented with mild symptoms or none at all.
The true number of COVID-19 infections in Canada is therefore many multiples higher than what has been identified to date. The only way to ascertain the real number of infections is to conduct population-based serological studies – blood tests that look for COVID-19 antibodies as evidence of past infection. None of this has been done in Canada. The plan Health Canada announced Tuesday – serological tests for one million Canadians over the next two years – will not, on its own, get to the answer fast enough to be useful in ongoing decision-making around reopening schools, businesses and hospitals.
Community serological studies, combined with widespread, easily available testing, offer the best assessment of the true presence and spread of the virus, age and other demographic variances, and how lethal it is. Without this information, provinces will struggle to accurately predict the trajectory of the virus, and risk failing to contain spread and protect the most vulnerable.
Serological studies undertaken in other parts of the world offer lessons for Canada. In the state of New York, the Department of Health conducted a serological prevalence survey of 3,000 residents and estimated that 14 per cent had already contracted COVID-19; in New York City, it was a staggering 21 per cent of residents – 10 times the number previously identified through limited testing of suspected cases. Another study conducted by the University of Southern California and the Los Angeles County Department of Public Health found that 4 per cent of L.A. County residents – not the 0.3 per cent found by active testing of suspected cases – had likely been infected.
Here in Canada, just 0.2 per cent of the population has been confirmed positive for COVID-19. It is not inconceivable that we too have already experienced 10 times that many infections. One per cent or more of us may well have already been infected (a number that could be significantly higher in dense urban areas), which would reduce the fatality rate to one per cent or less. That is still a tremendous number of lives lost. But after disaggregating the catastrophic “crisis within a crisis” of COVID-19 in Canada’s long-term care homes, where around 80 per cent of deaths linked to the virus have occurred – shamefully, one of the highest such rates in the world – the risk of dying for the broader population drops to approximately 0.2 per cent, or 2 per 1,000 persons infected, a fraction of the PHAC rate. As the crisis in long-term care stabilizes, it is likely that the overall Canadian fatality rate will also move lower into that same range.
Even at these levels, COVID-19 represents a threat to many tens of thousands of Canadians. But decisions around our shuttered systems will hinge on our understanding of how lethal this virus is. Will we need ongoing, widespread restrictions to prevent future waves, or can physical distancing and rapid, frequent testing coupled with strategically focused, well-resourced efforts protect those most at risk? Emerging data suggests the latter. Our public health strategy must be grounded in a better analysis of the true community-infection rate and case-fatality rate – because otherwise, decision-makers are just reading tea leaves while blindfolded.
Sign up for the Coronavirus Update newsletter to read the day’s essential coronavirus news, features and explainers written by Globe reporters and editors.