Thread by @MoriartyLab

With the transition of Canadian COVID-19 reporting to the combined respiratory virus dashboard, weekly COVID-19 deaths for Canada are no longer publicly reported by @GovCanHealth .

According to the most recent report, participating provinces reported 971 hospitalizations and 131 deaths for the period from Aug 25-Sep 28/24.

During the same period, Quebec, which accounts for 22% of the Canadian population, reported 4,691 hospitalizations and 256 deaths (current INSPQ data).

If hospitalizations and deaths are similar in Quebec and the rest of Canada (likely), the national report for the Aug 25-Sep 28/24 period should include at least 21,323 hospitalizations and 1,164 deaths.

Based on the most recent @CIHI_ICIS reporting of Quebec in-hospital COVID-19 deaths (Apr 1/22 to Mar 31/23), there were 1.28 times more documented COVID deaths in QC hospitals than were reported by INSPQ for the same period. However, CIHI captured only 84% of INSPQ-reported hospitalizations in QC for this period, meaning QC in-hospital deaths were likely ~1.52 times higher than deaths reported by INSPQ.

About 70% of all deaths in QC occur in hospitals (StatsCan), meaning actual COVID deaths were likely ~2.17 times higher in QC than reported by INSPQ, at least from Apr 1/22 to Mar 31/23.

Is QC reporting only "from" COVID deaths and excluding deaths of people who would have died anyway from other causes?

To test this, we can compare excess mortality estimates from @StatCan_eng for the same period, after correcting them to remove excess mortality attributable to toxic drug deaths and excess cancer deaths (stand-in for potential excess mortality attributable to delays in access to healthcare and MAID).

The StatsCan excess mortality program has been paused for the last 10 months, so estimates for QC are only available up to Sep 2/23, and are still incomplete for weeks before this date.

However, weekly age-adjusted per capita all-cause mortality rates for QC and other provinces are still published every month, meaning weekly excess mortality can still be estimated by comparing weekly mortality rates in one year to mortality rates in the same week of the preceding year, and using these ratios to estimate current excess mortality values compared to previous years when estimates are more complete.

From Apr 1/22 to Mar 31/23, there were 2.19 times more excess deaths in Quebec than reported COVID deaths, after adjusting for excess mortality attributable to toxic drugs and cancer.

This is very close to the CIHI-based estimate of actual COVID-19 deaths (2.17 times higher than reported by INSPQ for Apr 1/22 to Mar 31/23).

Taking the average of both methods for estimating under-reporting in QC, we know that for the year ending Mar 31/23, INSPQ reported 46% of "excess/from" COVID-19 deaths.

If we assume the magnitude of under-reporting of COVID-19 deaths by Quebec was similar after Mar 31/23, then from Aug 25-Sep 28/24 there were likely 558 excess/from COVID-19 deaths in Quebec, and 2,537 COVID-19 deaths in Canada.

Recall, the total number of COVID-19 deaths in Canada from Aug 25-Sep 28/24 reported on the new respiratory virus surveillance dashboard from PHAC: 131

So, it's likely that the new national COVID-19 dashboard is reporting, ~5% of "excess/from" COVID-19 deaths in Canada.

However, reporting of COVID-19 deaths by Quebec has actually worsened since Mar 31/23.

How do we know this? Because from Apr 1/23 to Jun 8/24 (the last date when age-adjusted all cause mortality rates are available from StatsCan), there were 6.37 excess deaths for every COVID-19 death reported by INSPQ, after adjusting for excess mortality attributable to toxic drugs and cancer.

So, from April 1/23 onward, Quebec has likely been reporting only 16% of its COVID-19 deaths.

If the under-reporting rate for Aug 25-Sep 28/24 was the same as for Apr 1/23 to Jun 8/24, then there were likely 1,630 "excess/from" COVID-19 deaths in Quebec from Aug 25-Sep 28/24, and 7,407 "excess/from" COVID-19 deaths in Canada during the same period.

Recall, the total number of COVID-19 deaths in Canada from Aug 25-Sep 28/24 reported on the new respiratory virus surveillance dashboard from PHAC: 131

So, it's likely that the new national COVID-19 dashboard is reporting, fewer than 2% of "excess/from" COVID-19 deaths in Canada for the 2024 respiratory virus season to date.

And no, Canada doesn't have exceptionally high COVID-19 mortality rates.

From December 2021 to March 2023, the estimated median weekly COVID-19 infection fatality rate for Canada was 4% LOWER than the median infection fatality rate for the UK calculated from ONS infection estimates and confirmed COVID-19 deaths, even though Canada is slightly older than the UK and has had lower rates of fresh vaccine dose uptake than the UK from 2022 onward.

What was the median weekly SARS-CoV-2 infection fatality rate for Canada during this period? 0.08747%

What was it in the UK? 0.08824%

Open access ref for UK IFRs: ncbi.nlm.nih.gov/pmc/articles/P…

Why are so few of Canada's COVID-19 deaths reported?

It's likely that at least in part it's because many/most(?) provinces now only report deaths in people infected for the first time, as @sarperotto has confirmed for BC via official communications from BCCDC.

From Dec 4/21 to date, the average person in Canada has had ~3.5 SARS-CoV-2 infections. Reporting only deaths associated with first time infections would reduce reporting by ~70%....meaning only ~30% of deaths would have been reported if this policy was in place in every province.

On top of that, testing in Canada during Omicron has declined ten times more than the decrease in infection and fatality rates, meaning that a substantial number of deaths and hospitalizations simply wouldn't be detected because of under-testing.

So, it's probably a combined issue of under-reporting and under-testing, but either way, it's an enormous problem.

As the Office of Canada's Chief Science Advisor Dr. Mona Nemer @ChiefSciCan has pointed out, Canada currently has gaps in the effectiveness of our communication about COVID-19 vaccines. Fewer than 15% of people in Canada got fresh doses in the fall 2023 campaign.

Perhaps if we started reporting more than 2% of actual COVID-19 deaths on our new national respiratory virus reporting dashboard people would better understand why getting a fresh vaccine is so important, and journalists would have real numbers they can use to help communicate about the issue.

I don't know who else to copy in on this post--it's like shouting into the wind, and I don't know what it will take for Canada and ALL provinces to start improving actual COVID-19 deaths and hospitalizations.

@picardonhealth @NightShiftMD @DrGorfinkel maybe you can help resurrect a national conversation on this topic.

Should also copy in @CBCQueensPark, who's still following official public communication about COVID-19 vaccines, or rather, the near absence of official communication about their importance.

Sorry for the oversight, Mike.

Adding to this post after seeing some comments/questions.

1. Some provinces just may be reporting extremely slowly to PHAC, so it's possible those numbers may still come up.....but they'd have to come up a LOT to match what is expected, if fewer than 2% of actual deaths are being reported.

2. Quebec hospitalizations are almost certainly under-reported too. It's likely it's not under-reporting so much as under-testing.

In the UK, during Omicron, there were 5.16 reported COVID-19 hospitalizations for every reported COVID-19 death. That's a mortality rate for hospitalized cases of 19%, which is about 5 times higher than mortality rates for all non-COVID-19 causes of hospitalization in Canada.

However, UK-reported COVID-19 hospitalizations per COVID-19 death are about half the number of hospitalizations per death reported by France, about half the admissions/death ratio reported by CIHI for Canada from 2022-2023, and about half the number of hospitalizations per death reported by Quebec up to March 31, 2022. Canada and the UK have some of the lowest numbers of hospital beds per capita of any OECD country, so the UK may admit fewer people with COVID, even though the mortality rate is high. Hospitalizations are also defined very tightly to include respiratory codes, which likely excludes COVID cases without respiratory codes in patient charts.

If we take the average hospitalizations per death reported by France during Omicron (France was historically an excellent reporter for COVID hospitalizations), by CIHI for Canada from 2022-2023 and by Quebec for Omicron up to March 31, 2022, then we would expect ~9.4 hospitalizations for every death.

For Aug 25-Sep 28/24, when there were an estimated 1,630 COVID deaths in Quebec, we'd expect 15,322 hospitalizations. However, 4691 were reported for this period. This means that about 31% of expected Quebec COVID hospitalizations are currently being reported.

And we would have expected 69,645 reported hospitalizations for Canada. Recall the number reported during this period on the new dashboard: 971

So, likely fewer than 2% of COVID-19 hospitalizations for Canada are being reported on the new dashboard--similar to the proportion of COVID deaths reported.

Again, this may be under-reporting--and it certainly is for many provinces. For Quebec I don't think it's under-reporting of known hospitalizations so much as under-detection of COVID-19 in people who are hospitalized, because of under-testing. And Quebec still tests considerably more per capita than other provinces.

Finally, it's possible that fewer than 2% of the expected 69,645 COVID-19 hospitalizations in Canada were reported for Aug 25-Sep 28, 2024 not just because of under-reporting and under-testing, but also because some provinces may simply be admitting far fewer people with COVID-19 to hospital.

The number of people who report receiving Paxlovid in Canada is almost negligible, so it's not because we're able to avoid admissions because we're prescribing a helpful drug that's now available to help prevent hospitalization. And most people who need to be hospitalized with COVID need hospitalization more than one week after symptom onset--too late for Paxlovid.

3. Finally, let's compare estimated actual COVID-19 deaths for Canada for Aug 25-Sep 28 to other causes of death.

We'll start by assuming there were about 1,481 "excess/from" COVID-19 deaths and 13,929 hospitalizations per week in Canada from Aug 25-Sep 28, 2024 (or 13,929 people needing hospitalization for COVID-19, although we don't know if all needing hospitalization were actually admitted).

First, in 2024 ~316,000 deaths per week are expected in Canada, from all causes. At this time of year it's lower, because influenza season hasn't really started yet, and influenza/respiratory viruses account for most seasonal variation in Canada's weekly death numbers.

But let's assume 6,084 expected deaths per week from any cause from Aug 25-Sep 28. 2024. 1,481 "excess/from" COVID-19 death per week during the same period is 24% of all expected weekly deaths from any cause. That's a LOT. It also corresponds to an excess mortality rate of 24%. That's a LOT. A quarter more people were dying prematurely than should have from Aug 25-Sep 28.

Of the 6,084 expected deaths each week from any cause, ~1,746 deaths each week were likely from cancer (all cancers combined), and ~1,571 deaths each week were from all major cardiovascular diseases combined. This means COVID-19 is (still) the number 3 cause of death in Canada, as it has been in each of 2022, 2023 and 2024.

In addition, from Aug 25-Sep 28, 2024, COVID-19 was killing about 15% fewer people each week than all cancers combined, and about 6% fewer people each week than all major cardiovascular diseases combined. That's not a helluva lot fewer deaths than either cancer cardiovascular disease.

COVID dwarfs the next leading cause of death in Canada after cancer and cardiovascular disease. The next leading cause of death in Canada is accidents, at ~311 deaths per week. More than half of accidental deaths in Canada are related to toxic drugs, or ~150-200 toxic drug death each week.

We don't have 1,481 COVID deaths every week through the year. It can drop as low as 450 COVID-19 deaths per week from Feb-Apr most years, and we've just had a huge summer wave in Canada.

I've seen reporting out of PEI that a spokesperson for an Atlantic province public health department (not PEI) quite recently said that there weren't as many COVID deaths as cancer deaths. That's quite a statement from a department that should understand the death toll from cancer.

I'm running out of steam, and don't have the energy to assemble hospitalization numbers for any cause and the most common causes of hospitalization in Canada.

However, from 2022-2023 @CIHI_ICIS reported that COVID-19 was the second leading cause of hospitalization in Canada after childbirth.

The 2023-2024 period doesn't look like it has likely changed much, and certainly for Aug 25-Sep 28/24 COVID was at least the 2nd leading cause of hospitalization in Canada.

I don't know how else I can put these numbers in context so that policy makers, decision makers, journalists and the public can understand what is at stake.

Yes, people can make their own decisions about masking and vaccination. However, if we're publicly reporting fewer than 2% of actual COVID-19 deaths and hospitalizations, how can we really expect anyone to make informed decisions and rush out to get their updated vaccines, stay home when sick, test and mask in indoor settings.

If you're a decision maker, legislator and/or contribute to policy making, please help. The emperor has no clothes. Please don't be one of the bystanders cheering for the naked person wearing nothing but a crown.

Please encourage accurate and complete COVID reporting.

Five years ago would you ever have imagined that this kind of reporting problem could happen in Canada? Remember who you were five years ago, and the red lines you couldn't imagine could ever be crossed in your province and in our country.