News

How gaps in Canada's COVID data are making it harder to fight the second wave

On too many key metrics, in too many provinces, messy or incomplete information leaves us flying blind

As more Canadians doom-scroll through daily updates of an intensifying second wave, publicly available data on the COVID-19 pandemic is becoming less reliable. Eight months in, public health officials are dealing with a Sisyphean task of collecting and analyzing a never-ending deluge of statistics, often by outdated methods (fax machines, anyone?) that have them struggling to keep the information both current and useful.

Now, they’re having to triage their priorities. “Surveillance systems break down when they are overwhelmed,” explains Jean-Paul Soucy, a Ph.D. epidemiology student at the University of Toronto and co-creator of the authoritative COVID-19 in Canada website.

Alberta’s travails may be a sign of what awaits other provinces. Throughout November, the province’s COVID-19 situation has deteriorated. It started the month with a rate of new cases of 110.2 per million population. Now, that rate is 220.7, second worse only to Manitoba. To make matters worse, though, Alberta has had repeated technical issues with its COVID-19 data—so many that it’s released regional data on just nine out of the first 18 days in November (on weekends, it releases topline, provincewide data.)

Critically, Alberta hasn’t released any data on the number of people tested since Nov. 6, meaning no one can calculate a key metric of the pandemic: positivity rates. As well, on Nov. 12, the province stopped virtually all contact tracing, effectively blinding the public to potentially new danger areas for the spread of the virus.

Across the nation, basic data that used to be available have vanished. Other facts and figures, when released, require time-consuming crunching before they can be compared to provincial requirements for rollbacks or reopenings. Some metrics that could show how well governments are doing in their battles against the pandemic remain frustratingly hard to obtain.

As the second wave intensifies, then, the provinces are like cars speeding ever faster down the COVID-19 highway. On some, the engine-temperature warning lights don’t work; on others, the fuel gauges function only sporadically. On a few, the repair indicators have been red for months.

If Canadians can’t monitor exactly what’s happening in the pandemic, how can we know where and how to fight this virus?

Recently, more alarms have been going off. The daily “data update” note at the top of Soucy’s website used to include brief explanations of historical data revisions by provinces and local public health units. Now, more and more, it details problems with the data itself.

On Wednesday, it noted Alberta’s missing testing data; Quebec’s now frequently late reporting of COVID-19 deaths (the 24 fatalities reported that day included five new ones as well as 19 others that occurred as far back as Nov. 1); and Saskatchewan’s second-wave data for regional health units can’t be compared to that from early months, because when the government expanded its set of health regions on Sept. 14, it didn’t sync the old cases to the new boundaries. That latter issue, means local areas can’t see if they’re doing better or worse than in the first wave.

Soucy’s note also updated an ongoing saga surrounding British Columbia’s data. On Monday, the province posted its case and lab data as usual, then pulled them down and replaced them with Friday’s statistics. So, for a while, the COVID-19 data reported by Dr. Bonnie Henry, B.C.’s provincial health officer, on the provincial interactive dashboard and on government websites all showed different numbers.


The issues may be geeky, but the ramifications are felt by everyone. “It makes it more difficult to discern trends [or the] trajectory of the epidemic, as well as undermining the credibility of the numbers themselves if it seems like public health has lost control of the situation,” says Soucy. For instance, those missing positivity rates from testing in Alberta could indicate trajectory of its current crisis, as rising rates signal “either a growing outbreak or insufficient testing (or more likely, both),” wrote Soucy in a recent post. “Since insufficient testing directly impacts our ability to isolate cases, identify contacts and break chains of transmission, both situations should raise alarm bells.”

In early October, Toronto stopped contact tracing, hindering its ability to help neighbourhoods hardest hit with new cases. On Wednesday, citing an internal presentation to hospital officials, the Toronto Star reported a 14 per cent positivity rate of COVID-19 tests done by Scarborough Health Network, which is in one of the lower-income parts of the city that is home to many essential workers.


Manitoba has had problems from the beginning. “It’s never been particularly great,” says Michael Pereira, a data journalist at the Winnipeg Free Press, of the province’s COVID-19 data. While he doesn’t think it’s getting less reliable in the second wave, lots of vital information simply isn’t regularly available. That may not have sparked public concern when caseloads were low, but it’s creating gaps in understanding now that they are the highest in the nation. Pereira’s data wish list includes hospital occupancy stats (beds, ICU, ventilators, staffing), contact-tracing numbers (cases handled, distribution of contact counts), case counts and the status of congregate-living facilities in outbreak (personal care homes, schools, correctional, hospitals).

In late October, his newspaper reported on the struggles of researchers and journalists trying to “pry out basic information they need to assess how Manitoba is handling the pandemic.” That included testing data from vulnerable sectors, such as long-term care homes and schools, and basic contact-tracing numbers. Moreover, as cases began to climb in late summer ahead of school reopening, Manitoba stopped releasing the age information that used to accompany each new case. “You can’t fix what you don’t measure, and what you don’t assess transparently,” Manitoba NDP health critic Uzoma Asagwara told the Free Press.

Kevin Wilson, an epidemiologist in Halifax, says he has seen “data problems pretty much the whole way through [the pandemic.]” For example, even though case numbers were dropping in Quebec by June, the province’s number of active cases stayed stubbornly high. “I was staring at it. ‘That doesn’t make sense, does no one get better?’ ” he remembered thinking. The next day, Quebec’s number of active cases plummeted from around 25,000 to 1,500. There was no big nefarious reason for the discrepancy, Wilson explained; rather, the province had neglected to build into its system a way to convert active cases to recovered ones.

Outdated data continues to be a concern. That Winnipeg Free Press article explained that reporting stale-dated information on the number of new cases “makes it impossible to accurately assess the growth of cases and the shifting impact on the health-care system.”

Sometimes, what data is available isn’t in a helpful format: detailed information about COVID-19 outbreaks and deaths in Ontario’s long-term care sector, for instance, appears in a horrifyingly long list on its website. As a result, a patchwork of sources have sprung up to collect and present the information in ways that the public can easily digest. “The data also has to be easily understandable and accessible for the public,” says Dr. Jennifer Kwan, a family physician from Burlington, Ont.

When the pandemic started, Kwan didn’t even have a Twitter account. Seeing a need to raise awareness of the situation—even before Canada began shutting its schools, borders and businesses—she began tweeting charts that plotted hospitalizations, deaths, testing and other basic measures. Now, she has 35,000 followers who tweet and retweet her daily updates.

In other instances, the available data doesn’t answer big questions. A key indicator in Ontario’s COVID-19 “response framework” is the number of weekly incidences of COVID-19 per 100,000 population. Yet, though the government recently upgraded its graphs and tables of data, it doesn’t provide that weekly tally of cases; nor does every local health unit. Those obsessed with COVID-19 data know that Ed Tubb, an editor at the Toronto Star, publishes that data almost daily on Twitter. But that doesn’t help the majority of Ontarians hunting for a critical marker on whether their regions will be adding restrictions or moving toward reopening.

Whatever the issues with the current data, everyone interviewed hankers for more. Kwan wants the demographics of those in Ontario’s hospitals and intensive care units. She’s also wondering about “patients who may be suffering from persistent symptoms or potential complications of COVID19”—the so-called “long haulers.”

After all, long after vaccines are available, thousands of Canadians will continue to suffer from the effects of this pandemic. The question is whether the public will have data to tell them how bad things really are—and whether they’re improving.