@COVID_19_Canada@StatCan_eng Please bear with us as we tinker with display settings. We're in the middle of converting the website to a simpler version, and there are glitches.
I'll do a thread summarizing some of the major insights from the excess mortality data here.
It lists the sections of the Canadian Excess Mortality Dashboard, with links, as well as brief descriptions of methods and (linked) sources.
We still have to translate, add ALT-text and explanatory text more geared to non-scientist readers.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite All estimates we present are directly pulled from @StatCan_eng current estimates of excess and expected weekly mortality rates in each age group. We subtract excess deaths in each age group attributable to toxic drugs, the 2021 BC heat dome and 2020 Portapique NS shootings.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite We're using 7-week, mid-centred rolling averages for all parameters. This helps to reduce the bounce of data in smaller provinces and age groups with smaller numbers of weekly deaths, because the value in any given week is the average of flanking positive and negative values.
As of Jun 6, 2020, there may be ~1K deaths that haven't yet been reported in Canada, but some of this may be deaths in regions with smaller first waves. Specific cause of death reporting is missing for ~2K deaths up to this date.
Specific cause of death reporting is missing for at least 72K deaths for this period.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite Basically, what this means is that we'll likely still be seeing considerably more deaths come in, and increases in excess mortality estimates for the period since Jun 6/20, despite the best efforts of @StatCan_eng to statistically compensate for slow provincial reporting.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite The blue box on at left, middle gives the most recent date to which a province has reported any deaths at all to Canadian Vital Statistics Deaths (CVS-D). For Canada this is a mixture of dates, but as you go through the slides for every province you'll see specific dates.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite The blue text box also emphasizes just how important it is NOT to conclude that excess mortality is declining for more recent dates. This is an artefact of very slow and incomplete death reporting.
I repeat, excess mortality will still increase considerably for later dates.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite Finally, the table at bottom right lists total age-adjusted all-cause excess and COVID deaths/100K (first and second rows), and cumulative % excess mortality since Feb 22/20.
I repeat....these excess mortality estimates are still artefactually low for most provinces.
As we do this, you can apply some of what I'm discussing when you look at the pages for your province. The format for these pages is the same for all regions.
Here are some, in no particular order. I suspect we collectively need to do a lot of reflection about these data.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite First, it's clear that excess ALL AGES mortality in the first Omicron wave in Canada will likely be greater than excess ALL AGES mortality for the first wave of the COVID epidemic in Canada by the time StatsCan releases next month's update.
After Apr/21, despite similar temporal patterns we see a divergence in the scale of reported COVID deaths and excess mortality. Excess mortality increases. Reported COVID deaths decrease.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite Some of this divergence probably represents the beginning of an increase in mortality due to delays in medical diagnosis and care during earlier stages of the epidemic, but especially from the spring onward, which is when hospitals really started to be hit by COVID.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite We even see this divergence in Quebec starting at about the beginning of July/21, even though QC has historically been excellent at detecting and reporting most/all of its COVID deaths.
1. The arrival and takeover of Delta, which had higher mortality rates in most age groups than preceding variants.
2. Most people older than 12 had received at least 1 vaccine dose.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite We won't have specific cause of death data for 2021 from StatsCan until Jan/23, so it's hard to know how much of this divergence is because of challenges accessing life-saving medical care, and how much is due to under-detection/under-attribution/under-reporting of COVID deaths.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite I've started working on a crude way to try to estimate warning "signals" of significant age-adjusted increases in weekly from the top 10 causes of mortality from 2021 and 2022 reported by StatsCan.
If I can figure this out I'll post to our website and notify you here.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite The other striking thing that stands out is just how high % excess mortality has been for people younger than 65 throughout the epidemic, even after adjusting for excess mortality attributable to toxic drugs.
This looks a lot like UK and US COVID AND excess mortality data.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite Frankly, Canada has never reported the number of COVID deaths in people <65 that we'd expect given the number of deaths we've reported in older age groups. We REALLY under-detect/report COVID deaths in younger people, maybe because we have an incorrect mental picture of who dies.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite But.... if we look at specific causes of death that changed significantly in Canada 2020 (slide 15 at the link at the top of the thread), it's clear that for people younger than 65 there was also a significant increase in death from alcoholic liver disease.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite The significant increase in deaths from alcoholic liver disease is still nowhere near the magnitude of increased % excess mortality in these age groups, though. "Subtracting" these deaths would reduce % excess mortality by maybe 10-20%.
QC has lower per capita rates of reported toxic drug deaths than many provinces. And given the quality of death reporting in the province it's less likely that drug deaths have been missed in QC.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite In Quebec, we do see elevated excess mortality in people younger than 65 that is clearly temporally linked to COVID death waves. Cumulative excess mortality to date from periods linked to COVID surges/waves is greater than reported COVID deaths in these age groups, even in QC.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite And yes, even in Quebec the number of reported COVID deaths in younger age groups is lower than would be expected based on OECD age-specific infection fatality rates for corresponding variants and COVID seroprevalence data. So, even QC is likely missing younger COVID deaths.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite But, the magnitude of % excess mortality in younger age groups compared to QC during COVID suggests that there's more going on in these provinces than "just" a hell of a lot of missed COVID deaths.
@COVID_19_Canada@StatCan_eng@DFisman@cdavidnaylor@AshTuite One more note before I change topics. In the 2020 specific cause of death data, you can see that by far the largest % increase in mortality in all age groups combined, and each one individually, was a catch-all category for "we don't know" called non-specific signs and symptoms.
In other words "we don't know" vastly outstrips any other cause of death, including drug- and alcohol-related deaths.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite Among people 0-44y there was a significant 29% increase in deaths attributed to diabetes in 2020. Among people 45-64y there were respectively significant 34%, 10% and 3% increases in kidney diseases, diabetes and major cardiovascular diseases. Many of these may be COVID-linked.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite Anyway, we'll be trying to tease this apart for years. And we know people working age and younger have had higher COVID infection rates than older folk, so it's not surprising if a chunk of that is COVID-linked. But the magnitude of that % increase is very concerning.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite One last clarifying point...even if the % excess mortality in younger groups has often been higher than older age groups, this doesn't mean that there were more deaths among younger people. It just means that observed deaths were higher than expected for the same age group.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite For the most part, % excess mortality for all age groups follows % excess mortality for people 65-84 pretty closely, which is what we'd expect for COVID, given higher baseline rates of death from any cause for people 85 and older.
Canada % excess mortality, all ages: DARKER RED
Canada age-adjusted reported COVID deaths/100K: LIGHTER RED
Compared to Canada, QC excess mortality was worse in 1st wave, then progressively improved.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite Quebec is also much further along for Omicron period death reporting than other provinces, so it's likely Canada will likely shift up relative to QC for the Omicron era too.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite BTW if you're getting kind of tired of this slow-moving thread and want the final take-home numbers, you can go to the Table of Contents on the website and skip to the Summary Tables section.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite The Summary Tables show cumulative totals up to June 6, 2020 (the last fairly complete data for death reporting: middle columns), as well as up to June 4, 2022 (right columns).
The upper table on Slide 13 shows details of COVID and non-COVID death reporting status by province.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite The lower table on Slide 13 shows cumulative excess mortality and the number of unexplained excess deaths that have not been attributed to COVID, toxic drugs, heat dome, Portapique shootings.
FWIW in Quebec, you'll note that this number is zero, even to June 6, 2022.
Don't forget to use the sliders on tables to see all the rows. I tend to be a little over the top showing everything you might want to know. But still there is so much missing info :)
Specific cause of death reporting 81% complete (3.2K outstanding)
~7.3K expected (all-cause) deaths still unreported to CVS-D
Unexplained excess deaths: 0
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite Important: Nova Scotia hasn't reported any deaths at all to CVS-D since October 23, 2021, so the excess mortality estimates are likely very low.
Specific cause of death reporting 94% complete (800 outstanding)
~5K expected (all-cause) deaths still unreported to CVS-D
Unexplained excess deaths: 0
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite Important: NB death reporting to CVS-D, or at least the reporting thereof by StatsCan, appears to have gone back in time compared to last month's update.
No deaths due to any cause are in the CVS-D after Dec 18/21, so the excess mortality estimates still likely very low.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite There also appears to be a data glitch for NB ~Apr 2020. I've double-checked my formulas to make sure I didn't introduce it (although I do make mistakes!).
I'm not sure what's happening with NB now, since excess mortality estimates have plunged compared to last StatsCan release.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite I think fall 2021 excess mortality estimates will likely go back up again for NB, based on what we see for other Atlantic provinces, and suspect there are some background reporting challenges going on.
So, we'll have to hold off on interpreting NB data till next month, I think.
Specific cause of death reporting 91% complete (22.3K outstanding)
~26K expected (all-cause) deaths still unreported to CVS-D
Unexplained excess deaths: 4.1K
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite Note: peak of excess all ages mortality in first Omicron wave has already exceeded peak of first wave of COVID epidemic in Ontario, even with no deaths at all reported by the province since March 12, 2022.
Specific cause of death reporting 61% complete (6.3K outstanding)
~11K expected (all-cause) deaths still unreported to CVS-D
Unexplained excess deaths: 0
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite Note: Manitoba has not reported any deaths at all to the CVS-D since June 26, 2021, so you can imagine how much excess mortality is currently under-estimated for the province.
The scale of the left y-axis is set at the same maximum across all datasets. This is to facilitate visual comparison.
@COVID_19_Canada@DFisman@cdavidnaylor@AshTuite In many provinces the % excess mortality in younger age groups soars pretty high above that 35% max, but it's also important to see the trends for older age groups, which constitute the bulk of excess mortality on a population level during COVID.
So, imperfect compromise...
ALBERTA
Excess mortality, all ages: 8.1K
Reported COVID deaths: 4.6K
Specific cause of death reporting 85% complete (10.4K outstanding)
Min 50 expected (all-cause) deaths still unreported to CVS-D
Unexplained excess deaths: 3.5K
I think Alberta's data speak for themselves.
I will also refrain from further comment on Canada's current champion of unexplained per capita and total excess mortality, British Columbia.
BRITISH COLUMBIA
Excess mortality, all ages: 11.1K
Reported COVID deaths: 3.6K
Specific cause of death reporting 87% complete (12.5K outstanding)
Min 100 expected (all-cause) deaths still unreported to CVS-D
Unexplained excess deaths: 7.1K
On that discouraging note, I'm going to stop this thread so I can respond to questions.
Questions, comments, suggestions all welcome. Today is the first day we're posting this, and we know there's still a lot of information people would like, not to mention "lay" explainers.
We'll post this week's Canadian COVID Hazard index and associated Canadian COVID data tomorrow/Saturday.
I've seen a couple of comments today that the Canadian COVID Hazard Index and data reporting we put out weekly is from @GovCanHealth /supported by PHAC.
A small correction :)
Our COVID and vaccine Zoom discussion series and outreach are generously supported by PHAC.
@GovCanHealth The work we do on the COVID Hazard Index and Canadian COVID data reporting is done entirely on a volunteer basis by me and grad students in my research group.
@GovCanHealth We will gladly share our data and methods with ANY organization, private, public or media who wants to use them to support Canadians' who need current, useable COVID information to make personal decisions.
You can use them for free if you credit us. We will help you find data.
And, as many have said, that reduces the impact of long COVID by two thirds.
Canada is facing labour shortages in many sectors? This is a simple, low cost way for governments to help with labour shortages.
From @ONS UK data, we know that about 8% of people infected with Omicron have symptoms lasting longer than 4 weeks, and that two thirds of these (5% of all infections) have long-lasting symptoms that limit the ability to perform activities of daily life.
@ONS About 50% of people living in Canada have been infected since Dec 2, 2021, about 60% of people younger than 40, and ~40% of people older than 40.
@UofT Knowing this, and knowing how student, postdoc and employee unions would respond to the same information, would you consider reinstating mandatory indoor masking?
If you knew it would reduce disruption of the functioning of the university, and promote continuity of learning?
@UofT Even if legal advisors are advising that you can't enact policies stricter than those recommended by the Ontario CMOH, I would bet pretty strongly that labour and student unions would have strong cases supporting a return to mandatory indoor masking.
It's pretty clear that if 80% of Canadians return to masking indoors we could reduce average daily COVID deaths and hospitalizations to a level similar to an average flu year.
That's in addition to the human rights argument for ensuring accessibility of all essential public places for all Canadians, regardless of their health status.
We're seeing most regions trend up now for several reasons:
1. Increasing waste water signal and estimated daily infections
2. Increasing hospitalizations
A major concern: weekly change in infections in people 40 and older is increasing 10% faster than in people <40.
The big concern is that infections we're seeing now likely include quite a few previously uninfected people older than 40 whose 3rd dose protection against infection from shots in Dec/Jan is now very low.