Skip to main content

‘We have the will, we have the expertise, but we don’t have the data': Nationally, the lack of coronavirus-related health and economic data is stalling efforts to analyze the fast-moving COVID-19 crisis

Open this photo in gallery:

Healthcare workers do testing at a drive-thru COVID-19 assessment centre at Etobicoke General Hospital in Toronto on April 21, 2020. Because of massive data gaps, the coronavirus death toll likely is being underestimated.Nathan Denette/The Canadian Press

Gaps in key health and economic data are hindering Canada’s response to the COVID-19 pandemic, leaving Canadians in the dark about who is being infected or struggling with the devastated economy, say researchers, politicians and scientists.

These blind spots could blunt the federal economic rescue effort, hide inequities in deaths from the disease and slow our emergence from self-isolation in the months ahead. Experts are urging provincial and federal leaders to open up more streams of data immediately, as doing so might save lives and livelihoods.

Canada has a long-standing problem of information gaps, The Globe and Mail found in a year-long series, and that has left us vulnerable during public health crises before. A government audit found that during the 2009 swine flu pandemic, data deficiencies left the Public Health Agency of Canada “unable to answer basic questions such as the rate of spread” of the virus.

In the dark: The cost of Canada’s data deficit

Globe editorial: Hey Canada, mind the data gap

Many experts say Canadians are now similarly vulnerable.

It’s not known how many health care workers in British Columbia have contracted the virus or how many home-care workers in Ontario. How many intensive-care beds are available per hospital in Quebec is not public knowledge. Nor are the ages of those who have been tested for the virus or the specific location of hot spots.

Nationally, the ethnicity of those who have been infected or have died is unknown. Because of data gaps, the death toll likely is being underestimated.

On the economic front, Canadians don’t know how many in each province are applying for employment insurance every week (as the United States does by state). They don’t have up-to-date numbers on bankruptcies, mortgages in arrears, how workers in the gig economy are faring, the extent of layoffs or the degree to which the federal government’s plan for an enhanced wage-subsidy program has spurred rehiring.

Arjumand Siddiqi, the division head of epidemiology at the University of Toronto’s Dalla Lana School of Public Health, said she and her colleagues are eager to help analyze the fast-moving crisis to a greater extent, but have been stalled by a lack of detailed figures on the demographics and locations of confirmed cases, among other things.

“We have the will, we have the expertise, but we don’t have the data,” she said. “It would be good to know what is actually happening.”

Open this photo in gallery:

A patient is brought to the emergency department of the Verdun Hospital in Montreal on April 23, 2020. One of the most pressing gaps, Dr. Arjumand Siddiqi said, is information about the ethnicity of those who have tested positive for COVID-19 or died of the disease.Ryan Remiorz/The Canadian Press

One of the most pressing gaps, Dr. Siddiqi said, is information about the ethnicity of those who have tested positive for COVID-19 or died of the disease. No Canadian province makes this data available, in keeping with a long-standing national aversion to publishing statistics about racial disparities in health. (Toronto’s Medical Officer of Health, Eileen de Villa, has announced that the city is exploring ways to collect race-based coronavirus data on its own.)

But there is reason to suspect race may be a factor in determining who is being infected and dying from the virus, Dr. Siddiqi said, both because of the prevalence of various underlying health conditions in some racialized communities, and their over-representation in low-wage jobs such as nursing, delivery and retail, which make them highly prone to exposure to the virus. Early U.S. data indicate that black Americans are being admitted to hospital and dying from COVID-19 at a disproportionate rate.

“We are very clear that we want to know who is at risk,” Dr. Siddiqi said. “But we’re just very hesitant – and that’s kind of putting it mildly – to add race to the set of dividing factors that we’re willing to entertain.”

This blind spot extends to Indigenous people, whose health care is largely provided by the federal government. NDP MP Charlie Angus would like to change that. In a letter to Health Minister Patty Hajdu last week, he urged the government to start keeping data on COVID-19 cases among Indigenous people, saying, “It would be irresponsible at this time to turn a blind eye to the movement of COVID through vulnerable populations."

“It seems bloody obvious that you would want to track this and make policy based on this information,” he said in an interview. “I think there’s a naive arrogance in the principle of saying: ‘We’re not the United States, we don’t have their problems, we don’t discriminate like that.’ ”

Trudeau announces $1.1-billion for COVID-19 research

Ottawa proposes commercial relief that would cover up to 75 per cent of rent payments for three months

Canada says one million face masks from China failed to meet proper standards, won’t be sent to provinces

Even government-funded groups such as the Canadian Institute for Health Information (CIHI) have begun calling for race-based data around coronavirus cases. The organization now supports the idea of health care providers asking a common question about the race of COVID-19 patients and says it would be willing to compile the data.

“The COVID pandemic is certainly exposing gaps in important data flows within and between health care systems in Canada,” CIHI spokeswoman Alex Maheux said.

Other salient facts about who is contracting COVID-19 remain unknown in most of Canada, including the location of case clusters. In its daily report on the pandemic, the government of Ontario provides a heat map of cases broken down by the province’s 34 public health units, so people can see that Toronto is the regional epicentre of the pandemic. But New York provides a far more detailed map showing the percentage of positive tests in each of the city’s dozens of zip codes, revealing stark geographic divides within a single metropolitan area.

Data frequency comparison, U.S.

unemployment claims vs. Canada

employment insurance claims

In thousands, seasonally adjusted,

data available as of April 23

April 23

7,000

U.S. unemployment

insurance claims

5-day gap

6,000

5,000

4,000

3,000

2,000

1,000

0

Jan. 2020

Feb.

March

April

Canada

employment

insurance

claims

250

200

150

83-day gap

100

50

0

Jan. 2020

Feb.

March

April

*The federal government has recently started releasing aggregated data on CERB applications - but this data wasn't available in the first month of the economic shutdown, and it is not broken down by province.

Age and gender data gap in

confirmed COVID-19 cases

Percent of cases, by province, data as of April 22

Gender

Female

Male

Unreported

0

20

40

60

80

100%

Gender unreported: 100%

N.L.

N.S.

PEI

N.B.

Que.

Ont.

Man.

Sask.

Alta.

B.C.

Age group

80 and older

60 to 79

40 to 59

20 to 39

0 to 19

Unreported

0

20

40

60

80

100%

N.L.

Age unreported: 100%

N.S.

PEI

N.B.

Que.

Ont.

Man.

Sask.

Alta.

B.C.

JEREMY AGIUS, DANIELLE WEBB AND

MURAT YÜKSELIR / THE GLOBE AND MAIL,

SOURCE: PROVINCIAL GOVERNMENTS;

U.S. DEPARTMENT OF LABOUR;

STATISTICS CANADA

Data frequency comparison, U.S.

unemployment claims vs. Canada

employment insurance claims

In thousands, seasonally adjusted,

data available as of April 23

April 23

7,000

U.S. unemployment

insurance claims

5-day gap

6,000

5,000

4,000

3,000

2,000

1,000

0

Jan. 2020

Feb.

March

April

Canada

employment

insurance

claims

250

200

150

83-day gap

100

50

0

Jan. 2020

Feb.

March

April

*The federal government has recently started releasing aggregated data on CERB applications - but this data wasn't available in the first month of the economic shutdown, and it is not broken down by province.

Age and gender data gap in

confirmed COVID-19 cases

Percent of cases, by province, data as of April 22

Gender

Female

Male

Unreported

0

20

40

60

80

100%

Gender unreported: 100%

N.L.

N.S.

PEI

N.B.

Que.

Ont.

Man.

Sask.

Alta.

B.C.

Age group

80 and older

60 to 79

40 to 59

20 to 39

0 to 19

Unreported

0

20

40

60

80

100%

N.L.

Age unreported: 100%

N.S.

PEI

N.B.

Que.

Ont.

Man.

Sask.

Alta.

B.C.

JEREMY AGIUS, DANIELLE WEBB AND

MURAT YÜKSELIR / THE GLOBE AND MAIL,

SOURCE: PROVINCIAL GOVERNMENTS;

U.S. DEPARTMENT OF LABOUR;

STATISTICS CANADA

Data frequency comparison, U.S. unemployment claims vs.

Canada employment insurance claims

In thousands, seasonally adjusted, data available as of April 23

April 23

7,000

U.S. unemployment

insurance claims

5-day gap

6,000

5,000

4,000

3,000

2,000

1,000

0

Jan. 2020

Feb.

March

April

250

Canada

employment

insurance

claims

200

150

83-day gap

100

50

0

Jan. 2020

Feb.

March

April

*The federal government has recently started releasing aggregated data on CERB applications - but this data wasn't available in the first month of the economic shutdown, and it is not broken down by province.

Age and gender data gap in confirmed COVID-19 cases

Percent of cases, by province, data as of April 22

Gender

Age group

Female

80 and older

60 to 79

Male

40 to 59

20 to 39

Unreported

0 to 19

Unreported

0

20

40

60

80

100%

0

20

40

60

80

100%

Gender unreported: 100%

N.L.

Age unreported: 100%

N.S.

PEI

N.B.

Que.

Ont.

Man.

Sask.

Alta.

B.C.

JEREMY AGIUS, DANIELLE WEBB AND MURAT YÜKSELIR / THE GLOBE AND MAIL,

SOURCE: PROVINCIAL GOVERNMENTS; U.S. DEPARTMENT OF LABOUR;

STATISTICS CANADA

(In Ontario, the government has provided data on the first half of postal codes to a small group of researchers, including the University of Toronto’s Laura Rosella. That has allowed her to show that disadvantaged neighbourhoods in Toronto have been disproportionately affected by the virus. She was able to publish her analysis last week, but not the raw data underpinning it.)

The ages of people who are being tested are not made public in most of the country, either, noted University of Toronto PhD students Isha Berry and Jean-Paul Soucy, who have been producing a national comparison of coronavirus data. Having a better understanding of testing ages would give us a clearer picture of who is being counted among Canada’s confirmed cases. But Mr. Soucy calls the subject a “black hole” in Canada. Estonia is one country that publishes such data, he says.

Even our death count is likely skewed because of data gaps. To reflect the likelihood that many patients are dying at home, some countries have looked at total excess deaths during the pandemic to estimate the true toll of the disease. In Britain, researchers using this method believe COVID-19 deaths have been about twice as high as the official count, which only includes people who have died in hospital after testing positive for the virus. In Canada, we likely won’t be able to make that kind of estimate until 2021. Statistics Canada’s regular lag time in releasing national mortality data is 11 months after the end of a year. Statscan says that they’re working with provinces and coroners to provide more timely death statistics.

Some provinces have done better than others. Alberta is widely seen as a national leader in providing COVID-19 data. The province periodically releases testing data by age to academics or journalists who request it, and has published detailed maps of virus hot spots in Calgary.

“From the beginning, Alberta has really done an excellent job of transparency in the COVID case data,” said Ilan Schwartz, an assistant professor of medicine in the infectious-diseases division at the University of Alberta.

Open this photo in gallery:

The wind catches the protective gown of a first responder as she gets ready to transport a patient to the emergency department of the Verdun Hospital, on April 23, 2020 in Montreal. The ages of people who are being tested are not made public in most of the country, either, noted University of Toronto PhD students.Ryan Remiorz/The Canadian Press

Other provinces are becoming less transparent. Some health authorities in the greater Montreal area, for example, recently stopped publishing the death statistics of each long-term care home – places where the pandemic has cut a deadly swath.

In addition to provincial gaps, there is also the difficulty of comparing numbers between provinces. Only about a third of the country’s reported COVID-19 cases say whether the patient was admitted to hospital, according to the federal government’s daily epidemiology update. That leaves us wondering how well the health care system will hold up as cases mount.

Statistics Canada is trying to plug some of the gaps on the run, allowing researchers remote access to stores of data usually guarded at the bunkers known as research data centres, for example.

But Michael Wolfson, a former assistant chief statistician at Statscan, says a lack of comparable national data could slow our ability to transition away from strict physical-distancing rules and restart the economy safely.

“It’s currently impossible to compare health data, including test results, between provinces in real time,” he said.

If Canada had a national database of detailed information about every COVID-19 patient and everyone who had been tested, Mr. Wolfson says, Canadians could more readily see patterns in how different types of people and communities are faring.

“For example, if Moose Jaw started by opening restaurants while another community opened clothing stores instead, we could see more rapidly how these different approaches affected the spread of the disease,” he said. “Canada has many really talented epidemiologists, but they are sidelined because they cannot access these kinds of data.”

Getting Canadians back to work and cushioning the pain of layoffs will also be harder without better economic data, researchers and economists say.

Open this photo in gallery:

Space available signs are shown on storefronts on Queen Street in Toronto on April 16, 2020. Another question without a clear source of data right now is how many businesses have closed permanently or temporarily in Canada.Nathan Denette/The Canadian Press

The country lacks key details about how workers in the gig economy are faring, for example – a long-standing gap, but a crucial one now, because so many gig workers seem to have lost their jobs, said Deloitte Canada chief economist Craig Alexander. “It would be really useful to have the information to know what’s happening with that segment,” he said. “That could shape your programs and policies on the other side of this crisis.”

Another question without a clear source of data right now is how many businesses have closed permanently or temporarily, said Frances Donald, global chief economist at Manulife Investment Management. “There’s limited visibility on that,” she said. It’s also unclear how long companies can remain shut before they are closed permanently and to what extent rehiring is currently under way. All of this would help economists assess a potential recovery – and help businesses plan for a reopening.

In the absence of good government data, some economists are turning to unconventional sources of information such as social-media sentiment, traffic patterns and, until the start of the lockdown, Open Table restaurant bookings, which offer a more current, albeit incomplete, picture.

As with health data, chronic problems of government opacity have been compounded during the pandemic by the situation on the ground, which is evolving faster than our ability to keep track.

“The vast majority of our economic data has been almost unusable to forecast the coming economic recession – so much so that most major central banks, including the Federal Reserve and the Bank of Canada, have thrown up their hands and said, ‘We can’t forecast what will happen next, so we won’t even issue forecasts,'” Ms. Donald said. “It’s extraordinary. These are the most powerful economics departments in the world.”

Meanwhile, many economists have praised Statscan’s efforts to produce more timely information during the crisis. The agency has long opted for accuracy over timeliness. But “the current environment has pushed Statistics Canada outside of its comfort zone,” Mr. Alexander said. “I’m really impressed that they were willing to accelerate some of their releases.”

Open this photo in gallery:

The employment insurance section of the Government of Canada website is shown on a laptop in Toronto on April 4, 2020. Weekly employment-insurance claims weren’t published in the first month of the shutdown.Jesse Johnston/The Canadian Press

Understanding and goodwill only go so far when the fate of the economy is at stake, however. In some cases, as with weekly employment-insurance claims, which weren’t published in the first month of the shutdown, when key decisions were being made, the federal government produces data occasionally – but not as often or with the level of detail economists or businesses need.

“The numbers are there, we’re seeing them being announced, dribsy drabsy,” said Kevin Milligan, an economics professor at the University of British Columbia. "Why can’t they put it out regularly? … This isn’t academics, sitting around wishing they had some data to crunch. This is big businesses and small businesses, making billion-dollar decisions about where they think the economy is going in the next few weeks and they’re trying to do it without even basic labour-market data.”

A big retail company, for example, needs to plan based on how much consumers will be spending over the next six months to formulate a restart plan, and figure out how much and when to rehire, Prof. Milligan said. They need to know “how bad is the income hit going to be to people – how many people have jobs and how many don’t have jobs.”

Some figures, such as the number of applications for the new Canada Emergency Response Benefit, can only be released by the federal government. (And economists would like more detail, such as provincial breakdowns, to those numbers.) In other cases, swaths of administrative data exist – but they are locked away by regulators, provinces or financial institutions, said Don Drummond, adjunct professor at Queen’s University’s School of Policy Studies and a former senior official in the federal Finance Department. He’d like to see more timely or accessible data on credit-card spending, outstanding mortgages, power demand and bankruptcy filings.

Too often, he said, various levels of government hold on to silos of data that fall within their authority, and don’t share them with national institutions such as Statscan that nominally answer to the public. Old habits die hard, even in the midst of a pandemic, when data are more important than ever.

“I can tell you from doing this stuff for about 40 years that right there is a recipe for disaster,” Mr. Drummond said.

With reports from Tu Thanh Ha, Karen Howlett, Stephanie Chambers, Justine Hunter and Marieke Walsh


Canada’s COVID-19 death toll passed the 2,000 mark on Thursday. Here’s how the death toll compares across the provinces.

COVID-19 death rates across Canada,

by province

Seven-day moving average of number of

deaths per 100,000 population, as of April 22

B.C.

Alta.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

March 15

April

Sask.

Man.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

March 15

April

Ont.

Que.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

March 15

April

N.S.

N.L.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

March 15

April

THE GLOBE AND MAIL, SOURCE:

PROVINCIAL GOVERNMENTS

COVID-19 death rates across Canada,

by province

Seven-day moving average of number of

deaths per 100,000 population, as of April 22

B.C.

Alta.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

March 15

April

Sask.

Man.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

March 15

April

Ont.

Que.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

March 15

April

N.S.

N.L.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

March 15

April

THE GLOBE AND MAIL, SOURCE: PROVINCIAL GOVERNMENTS

COVID-19 death rates across Canada, by province

Seven-day moving average of number of deaths per 100,000 population, as of April 22

B.C.

Alta.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

7

14

21

March 15

April

7

14

21

Sask.

Man.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

7

14

21

March 15

April

7

14

21

Ont.

Que.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

7

14

21

March 15

April

7

14

21

N.S.

N.L.

1.00

1.00

0.75

0.75

0.50

0.50

0.25

0.25

0

0

March 15

April

7

14

21

March 15

April

7

14

21

THE GLOBE AND MAIL, SOURCE: PROVINCIAL GOVERNMENTS

Sign up for the Coronavirus Update newsletter to read the day’s essential coronavirus news, features and explainers written by Globe reporters and editors.

Follow related authors and topics

Authors and topics you follow will be added to your personal news feed in Following.

Interact with The Globe

Trending