Test positivity, death rates, and why increased testing doesn't account for COVID-19 spike
TORONTO -- As cases ramp up across Canada, and health officials wave warning flags, there are still those who are raising doubts about the severity of COVID-19, questioning the focus on case numbers as an indicator for the virus’s spread.
CTVNews.ca spoke to experts to breakdown some of the common misconceptions about the data surrounding this pandemic.
WE’RE FINDING MORE CASES BECAUSE WE’RE TESTING MORE, RIGHT?
That’s an oversimplification, experts say -- and a dangerous one.
It’s true that regions with limited testing capacities may have an artificially smaller case count than those suffering a similar level of COVID-19 spread with more widespread testing. And since the threshold to receive a test -- whether it’s completely open to the public, or whether you have to have symptoms or close contact with a suspected case to get a test -- differs from place to place, which can also impact the overall case count.
But testing doesn’t pull COVID-19 cases out of thin air. According to experts, the key to understanding how testing reflects the viral spread is to look at the percentage of tests that come back positive -- something called the “test positivity rate.”
“If you're doing a huge number of tests, and if you didn't really have a problem, the more tests you had, the lower your test positivity rate would be,” epidemiologist Cynthia Carr told CTVNews.ca in a phone interview last week.
Since test positivity is a percentage of the tests performed, the population of a country would not increase or decrease it. For example, while America’s large population does play into its large case numbers, it would not explain a high test positivity rate.
“If you had one per cent of the people testing positive of every person tested, if more people are getting tested, it would still be one per cent of all people that should be positive, just as a bigger number,” explained Zain Chagla, an infectious disease specialist and associate professor at McMaster University.
Carr pointed out that even though the U.S. performs so many tests -- more than 111 million so far and counting -- “you shouldn’t have such an alarmingly different positive test rate compared to other countries, your positive test rates should still be, for example, about two, three, four per cent.”
However, while the U.S. has the most COVID-19 cases and deaths in the world, it does not have the highest test positivity rate. Countries such as Mexico and Argentina have test positivity rates from 45 to 65 per cent.
SO, IS CANADA SEEING AN INCREASE IN THE TEST POSITIVITY RATE?
It’s clear just by looking at Canada that the number of tests done does not correlate one-to-one with an increase in test positivity.
Carr gave the example of Prince Edward Island, which has one of the lowest rates of COVID-19 in the country, but one of the highest rates of testing when compared to their population.
“Their positive test rate is 0.2 per cent,” she said. “That's an example where you have a low test positivity rate and a low number of people, even though they’re [the] second highest testing rate.”
On the other end of the spectrum is British Columbia, she said.
“They're the third lowest testing rate in Canada, and yet, right now […] their active case rate is the third highest in Canada,” Carr said.
In Ontario, the highest test positivity rates for the entire pandemic occurred back in April and May, when significantly fewer tests were being performed. According to data from Public Health Ontario, in mid-April, we saw test positivity rates between six and 10 per cent, even though the daily number of tests being processed was less than 10,000 to 15,000.
Case numbers and test positivity dropped during the summer, as lockdown measures continued, but test numbers increased. The lowest test positivity rate in Ontario occurred in August, when daily testing numbers were consistently higher than 20,000.
But as the province enters what Premier Doug Ford has called the second wave, experts say that an increase in tests is not responsible for an increase in the test positivity rate.
And we are seeing an increase.
“The serious part of this is the percentage of positivity is actually going up such that we're seeing more than [one per cent] and creeping into some groups in two to three per cent positivity amongst those being tested, which really suggests that proportion is completely skewed -- that there's actually more people biologically testing positive,” Chagla said.
“If you took 30,000 people today versus 30,000 people in July, there are more people testing positive today than there were in July.”
So are we right back to where we started? Not yet.
At the peak of the first wave of the pandemic in Canada, test positivity rates were much higher than they are now, driven in large part by outbreaks in long-term care homes. But that doesn’t mean we can dismiss the sharp increase in cases.
HOSPITALIZATION AND DEATH COUNTS AREN’T REALLY RISING, SO WHY WOULD THE CASE NUMBERS MATTER IF THEY’RE LARGELY MILD?
It’s true that hospitalization and the number of fatal cases in Canada are not increasing at the same pace as the raw case count. But that's because younger people, who are less susceptible to severe outcomes, make up the largest proportion of new cases.
Carr pointed out that if there were an increase in cases across all age groups and nothing else had changed in terms of lockdown measures or regulations, that might point to the virus having mutated or changed in some significant way.
But since the increase is largely among younger people, this supports a theory that the increased reopening of businesses such as bars and restaurants are allowing a higher circulation of the virus, a conclusion Carr says should inform our public health response.
“We look at differences in patterns to understand better where our high risk population is,” Carr said. “So that right now we're seeing a higher rate of cases in younger people that [are not] currently associated [with an] increase in hospitalizations and deaths.
But that doesn't mean things are good. It means we need to be careful because those younger people could infect vulnerable people and in a week or two or three, then we see a surge in hospitalizations or deaths associated with it.”
Chagla agreed, noting that the trouble with focusing on hospitalization and death rates is that if we wait until those numbers skyrocket, it might be too late to actually respond.
“It is a late marker to start seeing hospitalizations and deaths. It’s often suggestive that that young, low risk group is then passing it into the higher risk group. And there's been historical examples from Australia to France, to Israel, where that transition occurred and that their systems became overwhelmed with people showing up to the hospital,” he said.
“When hospitalizations and death start rising, it’s already almost too late because whatever you do is not going to stop what's happening.”
BUT ISN'T THE DEATH RATE THE MOST IMPORTANT MARKER OF HOW DANGEROUS COVID-19 IS?
There’s more than one way to look at the death rate in any given country. You can look at the per capita death rate or the case mortality rate, which both provide important perspectives on an outbreak within a country.
Per capita means calculating the disease’s toll on the population at large by comparing the deaths to the overall population of a region. In Canada, roughly 25 people have died for every 100,000. Comparatively, in the U.S., around 63 people have died for every 100,000.
“Mortality rate per hundred thousand is really about kind of trying to keep track of how you might compare to other countries, but the disease specific mortality rate says: How dangerous is this virus?” Carr said.
Case mortality rate looks at the percentage of those who have contracted the disease and then died.
This paints a very different picture than assessing by population. In Canada, we currently have an overall mortality rate of six per cent, while the U.S., for example, has a mortality rate just under three per cent, which is why the numbers need to be looked at in context.
“[If] you get outbreaks that caused a significant amount of death in a highly vulnerable population as compared to a population [with a] normally distributed risk, you're going to artificially increase your case fatality rates,” Chagla said.
In Canada, more than 70 per cent of COVID-19 deaths have been in those over the age of 80, many of whom were living in long-term care homes. Those between the ages of 20 and 49 make up only roughly 0.8 per cent of all deaths.
In the U.S., while older people are dying from the virus at a higher rate, there is also a higher proportion of deaths among younger ages. According to the National Center for Health Statistics, those between the ages of 25 and 55 account for roughly eight per cent of all deaths.
Even still, mortality rate can’t give a clear answer to how dangerous the virus truly is because so many factors affect it.
Sometimes mortality rate can reflect outbreaks that ripped through vulnerable sectors of the population, such as long-term care homes in Quebec and Ontario, or as a result of an overtaxed health care system that maybe is not as advanced as other regions.
Carr pointed out that according to the CDC, 94 per cent of people who have died of the virus in the U.S. had at least one or more other condition or risk factor at their time of death -- only six per cent of people had COVID-19 listed as the sole cause of death.