Thread by @MoriartyLab

Expected Omicron deaths in Canada: provincial estimates, as of Jan 21

I'll show individual panels as I go through the thread, if this image is too small for those like me who need increasingly strong reading glasses.

🧵

At the end of this thread I'll link to previous threads with explainers, data sources etc.

As always, you can find the spreadsheet I use to estimate FUTURE Omicron mortality rates in Canada here.

It shows expected annual flu deaths, for scale comparison.
docs.google.com/spreadsheets/d…

I don't have estimated infection numbers for multiple provinces, so I can't show all provinces.

The expected future infection fatality rate for PEI is about the same as Nova Scotia. Newfoundland Labrador and New Brunswick are currently about 50% higher than NS.

Expected future infection fatality rates for Nunavut, the Yukon and Northwest Territories is about 50-60% lower than the IFR for Alberta.

The infection fatality rates I use to estimate deaths to Feb 9 from infections contracted up to Jan 12 are calculated from age-specific infection fatality rates (Omicron 50% lower than Delta), 2021 population age structures, age-specific vaccination rates,

cont'd

the Cdn age-adjusted odds ratio for death in vaxxed vs unvaxxed as of Jan 1 (@GovCanHealth 0.2) and the Jan 21 estimate of vaccine protection vs any infection in people with at least 2 doses (@COVIDSciOntario: 46.4%).

@GovCanHealth @COVIDSciOntario For those out there who might be doing estimates in your own jurisdictions, the global IFRs for each province used in these calculations, derived from all of the above info, are as follows:

NL 0.214
PEI 0.126
NS 0.138
NB 0.214
QC 0.126
ON 0.159
MB 0.104
SK 0.107
AB 0.104

@GovCanHealth @COVIDSciOntario BC 0.144
YK 0.067
NWT 0.041
NUV 0.061
CAN 0.1405

@GovCanHealth @COVIDSciOntario I know most of you aren't interested in these numbers (sorry).

@GovCanHealth @COVIDSciOntario hang on....just realized I should have started the thread after supper and that I underestimated how much time I had left.

will be back in half an hour

@GovCanHealth @COVIDSciOntario Good supper!

Here's the link to the thread with the sources, explainer etc

https://twitter.com/MoriartyLab/status/1482783048617836546?s=20

@GovCanHealth @COVIDSciOntario OK, so as I've explained previously, COVID-19 deaths are typically under-detected/under-reported outside Quebec.

You can see that in the graph here.

@GovCanHealth @COVIDSciOntario The blue dashed line shows the actual expected new daily deaths to Feb 9.

The red dashed line is the actual expected deaths divided by the typical under-reporting ratio for Canada ("Expected reported).

@GovCanHealth @COVIDSciOntario The black solid line is the number of death reported since Dec 15.

It matches the expected REPORTED deaths almost perfectly.

So, the under-detection/under-reporting issue in Canada appears to be continuing.

@GovCanHealth @COVIDSciOntario Let's look at this province by province.

Here's Quebec. Reported new daily deaths match expected new daily deaths well.

I suspect downward slope after peak might be steeper, but am waiting to be sure deaths have peaked before adjusting IFR to reflect deaths during up-slope.

@GovCanHealth @COVIDSciOntario Now here's Quebec compared to Ontario.

You can see that reported deaths in ON are considerably lower than expected deaths.

In the past, ON has reported ~50% fewer C19 deaths than estimated actual C19 deaths.

We'll see if that holds true over time, since ON reports slowly.

@GovCanHealth @COVIDSciOntario Here's Ontario by itself.

It doesn't look like new daily deaths will peak before Feb 9, as we'd expect because cases in Ontario seemed to start peaking 2-3 weeks after Quebec.

ON death curves typically have a very long right tail compared to QC. BC does too FYI.

@GovCanHealth @COVIDSciOntario Here are expected vs reported deaths for BC.

Reported deaths are substantially lower than expected.

Part of this is likely reporting delays--you can see the uptick in new daily reported deaths starting around Jan 16, about 7-10 days later than uptick in expected deaths.

@GovCanHealth @COVIDSciOntario Here is Alberta.

So far reported deaths are matching expected deaths quite well, although we'll know better in the coming week.

@GovCanHealth @COVIDSciOntario Saskatchewan

Expected new daily deaths are small (less than 2/day up till now), so there's a fair bit of bounce. We'll know better if reporting is as expected in the next week.

@GovCanHealth @COVIDSciOntario Manitoba

Reported deaths are definitely lower than expected deaths. I suspect we'll see a big increase this coming week.

@GovCanHealth @COVIDSciOntario Finally, Nova Scotia.

Reported deaths still lower than expected deaths. In the past about 50% of NS's likely COVID-19 deaths have been reported, although this estimate is less certain than for provinces that have had large numbers of deaths in previous waves.

@GovCanHealth @COVIDSciOntario I'll update these analyses daily. I add them to the thread pinned to my profile so you can fine them more easily.

@GovCanHealth @COVIDSciOntario But again, as I keep saying, watch Quebec for accurate death numbers.

Up until today, Ontario and Quebec should have been reporting the same number of new deaths daily. From here on out, ON should be reporting more new daily deaths than QC.

Let's see if/when that happens.

@GovCanHealth @COVIDSciOntario Let's return to estimated deaths for Canada.

We're about to reach the highest REPORTED daily deaths seen in Waves 1 and 2.

We haven't had deaths reported in most provinces this weekend, and Monday reporting is usually a bit behind. Typically Tues is first good look after w/end.

@GovCanHealth @COVIDSciOntario Right now, the biggest question is how long we'll stay above the W1 and W2 peak. If it's for an extended period, there will be a high death toll.

We may be starting to see a peak for the country around Feb 9. Let's hope....

@GovCanHealth @COVIDSciOntario I know people expect/hope to see a short, sharp peak, or at least that's something that is commonly discussed when people explain why they think this wave won't be as bad as previous waves.

As you can see from the Quebec graph here, I'm not so sure that will actually happen.

@GovCanHealth @COVIDSciOntario Here are new daily deaths in Canada, South Africa, Denmark and the UK.

All that talk in the 2nd and 3rd week of December about the Omicron wave being "mild" wrt deaths was premature, as most of us who watch COVID deaths closely were yelling.

@GovCanHealth @COVIDSciOntario We are not seeing the short sharp death peaks in any country that some expected.

We likely won't unless we greatly reduce cases and keep them tightly under control.

If we can't, there will be a lot of death, and we need to stop hoping/pretending this won't be true.

@GovCanHealth @COVIDSciOntario Here are estimated new daily deaths for Canada to March 1, based on IHME/ICL infection projections to Feb 1st.

I'm not sure yet how accurate the infection projections are to Feb 1, but if they're reasonable, we'll have new daily deaths at or above the Jan 20 number until Mar 1.

@GovCanHealth @COVIDSciOntario I don't think we're looking at a particularly short, sharp wave, at least wrt deaths, and I don't think the death waves in other countries that have had the earliest Omicron waves are looking short and sharp either.

@GovCanHealth @COVIDSciOntario Canada's Omicron wave is at best 1 week behind others.

We should be reporting on what is happening in QC, where death reporting is strong, to help globally.

It's too bad. We should be well-positioned to do some of the research SA and the UK have been doing, but we're not.

@GovCanHealth @COVIDSciOntario And to be honest, inside Canada there's zero interest outside Quebec in understanding why/how so many COVID deaths are being missed, or in learning from QC how to improve this problem.

In fact, other Cdn regions can appease the public by saying "at least we're not as bad as QC".

@GovCanHealth @COVIDSciOntario This is NOT saying that Quebec performance has been ideal--not at all.

But actually, QC is perhaps the best case scenario in the country. QC, NWT and PEI have the highest vax rates for people 70+. It's nearly 99%. And it has higher booster rates in people 70+ than almost all.

@GovCanHealth @COVIDSciOntario Quebec has high rates of mask-wearing, although higher mobility rates compared to pre-pandemic times than Ontario, for example, so cases will likely continue to spread quite quickly.

@GovCanHealth @COVIDSciOntario Some regions, like the Northern territories, Alberta and Saskatchewan have younger populations, which will reduce expected deaths. Provinces west of Manitoba have also seen a lot of deaths since July, which means there are fewer people left to die in the Omicron wave.

@GovCanHealth @COVIDSciOntario I'm hoping that policy-makers outside QC will look carefully at these analyses and take them into account when planning how to manage Omicron (and whatever comes next) over the next 6 months. But I'm not holding my breath either.

@GovCanHealth @COVIDSciOntario Again, if you want to take a look at future estimated deaths, you can always check them here. It's updated daily.

docs.google.com/spreadsheets/d…

@GovCanHealth @COVIDSciOntario We're still expecting 50K new C19 deaths (26K reported) in this bloody stick-our-head-in-the-sand because we're-too-nice-for-it-to-be-true country.

Now that vax protection vs any infection reported by @COVIDSciOntario is stabilizing at ~50% this future estimate is pretty firm.

@GovCanHealth @COVIDSciOntario We can change these future deaths by multiple, overlapping mechanisms:

1. Get everyone 50+ boosted ASAP
2. Get everyone into N95 masks
3. Slow spread so testing is not overwhelmed so we can get people at risk treatment within 5 days of symptom onset

@GovCanHealth @COVIDSciOntario 4. Slow spread so we can get enough medications into the country to reduce future deaths (as much as 30%, or 15,000 prevented deaths if we can do this)

5. Clear communication about staying home when sick (and economic support to do it), and why this is important

@GovCanHealth @COVIDSciOntario 6. Governments provide support needed to do the basics (staying home when sick, high quality masks, prioritizing vax and 3rd dose for everyone at higher risk and 50+).

7. Messaging that this is like first wave--we all need to do our part, and we all need to protect each other.

@GovCanHealth @COVIDSciOntario One of the biggest problems with under-detecting/under-reporting deaths is that people don't understand how serious the situation is.

Most people do act when they understand the real risk, and Canadians have been good about this.

Help people understand what is really happening.

@GovCanHealth @COVIDSciOntario With that, I'll wind up the thread.

Clearly, Twitter is not my medium, or I'd be better at explaining things succinctly :)

Protect others. Everyone's life matters.