Deaths

Day 26 – April 11, 2020

For the sake this example, I’m about to virtually kill a lot of people. Please don’t feel bad — they never existed.

Let’s imagine you want to drive from Vancouver to Seattle… and let’s further imagine that there’s a winding road that follows the coastline all the way down. Sidenote — for part of it, there is… there’s a 20-mile winding road from Bellingham to Burlington called Chuckanut Drive that’s well-worth the detour. Spectacular views and much more.

So… there’s the imaginary coastline road, and there’s Highway 99/I-5 which actually exists.

And you’re a new driver, kind of nervous… the thought of the fast-moving traffic on the highway scares you a bit. But you also know the coastline road is very winding, and you’ve heard of cars losing control and going over the cliff. You do your research and quickly find a report that tells you that over the last two years, accidents killed 45 people on the highway and 24 on the coastline road. Again, I’m making this all up. Nobody was hurt in the creation of this posting.

No brainer, you think, the coastline road is twice as safe as the highway… because that’s the ultimate measure of safety, and there can’t be too much more to it…

Well, there can be… and if you keep reading down my imaginary report, you’ll find that the coastline road seems to have about one accident a month. Like clockwork, once a month, a single-occupant vehicle loses control, rolls down the 100-foot cliff and kills the driver. That accounts for the 24 deaths.

On the highway, as it turns out, some idiot last year was celebrating something… and rented out one of those monstrosity stretch Hummer limos, filled it with 43 of his closest buddies… and apparently everybody, including the driver, got drunk… and the limo, with its full tank of gas, crashed into a telephone pole, exploded into flames and killed all 45 occupants.

That changes things a bit, doesn’t it…

Those 24 single-car accidents each have a little circle around them. The HummerLimo has a single, big circle around it. Around here, the Lynn Valley Care Centre has a big circle around it too, as does the administration office of Lions Gate Hospital. The Mission Institution. The Okanagan Correctional Centre in Oliver. The Blueberry River First Nations community near Fort St. John. All of them have their own little circles.

So… let’s talk about clusters.

From an epidemiological point of view, a cluster is defined by infections that are grouped by where and when they happened. If those two things are close to each other, they’re part of the same cluster. Infections can be clustered, deaths can be clustered. Really, any statistic can (and usually should) be associated where it’s relevant.

In B.C., the Lynn Valley Care Center (where the average age of residents is 87) recorded Canada’s first COVID-19 death on March 8th. Since then, it’s accounted for an additional 17. Here in B.C., our mortality rate per resolved case is 6.0% when you count Lynn Valley, and 4.3% when you don’t. This is in no way minimizing the importance of each and every one of those people; rather, it’s just to point out where they should fit in statistically.

Treating those unfortunate 18 deaths the same as any random 18 in the city would skew things significantly; there’s clearly a lot more to learn about those 24 individuals that drove off the cliff than by analyzing the demographics of the 45 people in the limo.

This is all something to keep in mind when reading reports that tend to favour skewing data towards their intended conclusion… something I’m saying for my own benefit as much as yours, because these days I’m being bombarded with articles and reports and opinions, many of which are diametrically opposed to each other. All of them claim credible evidence. I’m trying to keep a level head, and you should too.

As for today’s numbers… well, it’s green all the way across the bottom. From a purely aesthetic point of view, that’s as good as it’ll ever get… and it’s what I consider great evidence that what we’re doing is working. Let’s keep at it, and let’s hope those people that aren’t such good listeners don’t become clusters of their own once this long-weekend is over.

View Original Post and All Comments on Facebook

Day 17 – April 2, 2020

I added a little table (just above the logarithmic chart) to the spreadsheet yesterday, and today I will explain it. It’s a simple “look-up” table for “Time To Double”, useful if you want to know how a certain percentage maps to a TTD. For example, let’s say you have $1,000 to invest, and you want to double it to $2,000 in 7 years. What interest rate would you need? The answer is 10.5%. If you can wait 10 years, you’d only need a rate of 7.2%. How long to double your investment if you’re being offered 20%? The answer is 3.8 years.

These percentages and their related time periods can measure years… or days, which is the relevant discussion.

Let’s begin with a simple example, where we start with the number 100. And we are adding 20 to it every day. After 5 days, it’s doubled to 200. A TTD of 5. Now we keep adding 20 per day… so it’s going to take another 10 days to go from 200 to 400. And to double from 400 to 800, it’ll require a further 20 days. The only thing doubling here is the TTD itself… and this represents linear, not exponential growth. Certainly, it’s growing… and in this example, that 100 will grow indefinitely… but, as it does, its TTD gets bigger and more distant.

Now let’s imagine an example where on day 1, we’re at 100. But by day 4, we’re are 200. And at day 7, we’re at 400…. and we’re at 800 after only 10 days. So this is clearly a TTD of 3, and if you look at the continuing growth… 800, 1600, etc… it’s not hard to imagine what this would look like on a graph… an ever-increasingly steep curve. With a consistent TTD, there is exponential growth. The steepness of that curve has everything to do with the actual TTD, and that’s important because no matter what the finish line, it’s important how quickly we get there. In this case, we want to get there as slowly as possible.

The big graph on the bottom left shows those curves, overlapped on each other, showing how numbers have evolved for different jurisdictions from similar starting points. The logarithmic graph to its right shows the same data, and when you graph exponential data on a logarithmic scale, consistent exponential growth shows up as a straight line. Those 4 TTD lines of 2, 3, 5 & 10 days are the best example. A logarithmic presentation also helps to show the deviation, positive or negative. As that exponential growth increases or decreases… ie, as the TTD increases or decreases, the lines for each country (or province) will move… and obviously, to the left (into the steepness) is bad, and to the right (flattening out) is good.

Logarithmic graphs can be a little misleading in the way they squish data, and can misrepresent reality. But from the point of view of displaying trends, they’re pretty good. We can look at the encouraging B.C. line. We can look at the Canada line, and at least relate to the fact that we’re on a very different trajectory than what the U.S. is following. As much as the numbers back east have jumped, and as exponential as the growth continues to be, it’s less exponential, ie slower, ie the TTD has gone up, ie… from a trending point of view, not worse than what led up to it.

Even without the graphs, the numbers speak for themselves, and the growth percentages are there, day to day, both for Canada and for B.C. You can plug those numbers in to the little table… from today, from a week ago… and see what TTD would correspond.

That being said, what exactly are we measuring? These TTDs are important to chart the rates of growth, but rates of growths of what? Known cases? Presumed cases? Hospitalizations? Patients in critical condition? Deaths?

The only thing I’ve been dealing with are confirmed cases and their growth. My data deals with the confirmed known spread of the virus…. but all of those other numbers are also important, and will be tackled in due course. Topics for another day.

View Original Post and All Comments on Facebook

Day 12 – March 28, 2020

There’s that feeling you get when you sit down in a rollercoaster… first of all, what the hell am I doing, do I really need to be doing this… but once the thing starts moving, there’s no way out, so the impending dread as you start going up that first big hill… click-click-click as the chain underneath pulls the train slowly… wow, this thing is going a lot higher than I thought… click-click-click… this was such a stupid idea… click-click-click… ugh, this is a lot steeper than it looked… click- ohh.. no more clicks. We all know that means…

… and as the train gains momentum and sends you flying down that first huge drop, two things will come to mind… one, this part of it will thankfully be over soon and two, now you have a clear idea just how steep it was. Which serves to illustrate where we are today with respect to the numbers coming our way in the next little bit… there is a finish line to them, a week to ten days… and we once we hit that bottom, we will know exactly how steep things were.

Given where we are today and as per what I wrote yesterday, I don’t think we need to close our eyes and scream and hope for the best. It’s looking better than that. At least, on paper and at least, for now.

In B.C., although we had the largest one-day increase in cases yet, it’s perfectly in-line with our linear growth. Dr. Henry, for now, would like to see that number consistent at 12% which is roughly where it’s been. The average of the last 10 days is actually 11.1%. Today’s number was 11.6%. Yesterday was 9.2%. Nice solid straight yellow line, right in the sweet spot. And might I add, I am tracking total cases as they accumulate, not factoring in recoveries and deaths. The outcomes of these cases is a whole separate topic. But on that note, while we saw an increase of 92 new cases, we also saw 121 cases moved to the “recovered” column. As far as these numbers are concerned, today in B.C., there are less active cases than yesterday.

Canada’s number is bigger, but also consistent and also, slowly, hopefully, for now… going down day by day. Yes, of course the number of cases is increasing, but that rate of increase is itself decreasing. See that column… 4 out 5 days of green numbers. The rate of growth is slowing. For now. Are we still following the U.S. trajectory? Visually, and numerically, we’re not. Not so long ago, and you can still see it on the chart, Canada’s data was almost exactly perfectly 8 days behind the U.S: Feb 29, Mar 1, Mar 2… the U.S. had 68, 75 and 100 cases. 8 days later, Mar 8,9,10 — Canada had 66, 77, 94 cases. Perfect lockstep. And if you eyeball those numbers as you slowly go down the two columns, you see them in lockstep… and then they slowly start drifting apart.

The hope is that we wouldn’t follow them down the hellhole-course they’re presently on, and, for now, we’re not. We’re at 5,655 nationwide cases. 8 days ago, the U.S. was at 24,218. Had we “kept up”, today’s number would be 4x what it actually is. We’re now more than 11 days behind them.

So what does it all mean…

I’d like to address some of the comments that question the usefulness of these numbers in general, how the testing is inadequate, this isn’t reality, this is a useless exercise because the numbers are all bullshit. That the real case numbers are anywhere from 10x to 50x and it’s anyone’s guess. And therefore, blahblahbblah.

So, first of all, the way to solve big problems is to break them in half. Solve each half independently, and once you do, the big problem is solved. And if one or both of those halves is too complicated to solve, break it in half again and solve that. Keep breaking it in half until you have manageable pieces to solve.

The enormity of our present situation requires breaking it into hundreds of pieces, but here are some of the big ones, each of which needs to be broken down into many smaller pieces:
– the actual number of cases out there, factoring in recoveries
– the actual number of cases that require hospitalization
– why are some demographics hit so differently than others
– the testing infrastructure, and the strategy and adequacy of it
– the ability of our medical infrastructure to handle the cases
– the actual number of people dying from this
– the economic implications of allowing this to go on too long
– the herd immunity thing
– the treatment options, effectiveness of therapies, and timelines
– when and where is the vaccine

Without tackling all of that, notwithstanding each of those topics is its own book, and that’s only a small snippet of topics that need addressing, where we’re at right now is trying to solve chunks of a problem with incomplete information. One thing we have to our great advantage is learning from what others have or haven’t done ahead of us. Like one big change that was implemented today here is that the number of patients on ventilators doubled. Because suddenly a lot more people got a lot more sick? No. Because we learned from data elsewhere that putting people on ventilators sooner has a huge impact on positive outcomes. We didn’t know that two days ago, and now we do, and now we use it to our advantage.

Just because we don’t know something is no reason to throw our hands in the air — “these numbers are all crap anyway” — but to tackle this particular aspect…

Knowing the actual number of cases out there would have a profound effect on many aspects… first of all, how many actual new cases are there each day… how many of them will the person never even know, how many will they get sick but not too sick, how many will need a hospital, how many will die. If we could snap our fingers and know all that, it’d be great. One school of thinking that might kick in is that if actually the number is not 10x or 50x but actually 500x, and many of us have had it and never even knew it, and now we’re immune and will be for several months and even if we’re not, who cares, clearly I can fight this thing off so let me get a little sick and impose my herd immunity and get back to work since the actual mortality rate is only 0.2% etc.

Don’t think everyone has their heads stuck in the sand thinking the published numbers are the extent of this. One day, in hindsight, we’ll know those numbers. It’s possible that one day, we will have instant, cheap and available tiny-traces antibody testing. You’ll be able to wander into Starbucks, and along with your chai latte, spit into some throw-away little thing that’ll turn red if you’ve had it, stay blue if you haven’t. But until we get to that point, to a great extent, all we’re doing is buying time. Flattening the curve to suppress the load on our medical infrastructure. Isolating ourselves so we don’t infect others, especially those who are much likelier to get lethally ill. Keeping this thing controlled and contained until we’re certain we can manage it. It’ll likely never go away, and the waves of it appearing in the future will hopefully wind up in the “no big deal” pile.

But for now, the published numbers, the important numbers… the ones that are putting load on our medical system… the 884 confirmed, the thousands of others likely presumed but not confirmed… don’t think they don’t know about it. Don’t think when they tell you to stay home for 14 days, pay attention to your symptoms but don’t come in — that they’re not tracking you. You, who may well have it who think you don’t count — trust me, you do. Not in my numbers, not in their published active-cases numbers, but you’re out there somewhere, included in all of the projections of what might happen and how they’re going to take care of you if you get really sick. Some of you think you have it, but don’t. Some of you have it and don’t know it. Neither of you got tested, so hey what the hell they don’t know what they’re doing this is bullshit… yeah, no. Not at all. They’re not going to waste a test to confirm a mild test. There’s an N% chance you have it, depending on your age and other risk factors. Take care of yourself with the provided guidelines, and you’ll most likely be ok. And if you’re not, critical care awaits you with open arms. As opposed to everyone who thinks they might have it coming in and overwhelming a system that, certainly at the moment, is not prepared to test 2 million people overnight. If you’ve had it, one day you’ll know.

The fact that our hospitals are not overrun… the fact that we’re prepared at present to handle anything but the absolutely worst-case scenario… the fact that were are notably flattening our curve, both provincially and nationally… and the fact that we’re doing that with incomplete information, tackling big, multi-faceted problems… don’t worry too much about absolute numbers and how you feel they don’t reflect reality. They’re serving us well.

Speaking of serving us well, please take a moment to step outside at 7pm tonight (and every night) to cheer the heroes of this nightmare — hope you never need their help, but the army of medical workers of this province and this country, and indeed, around the world… deserve to (loudly) hear our gratitude and appreciation.

None of us like this. They don’t. You don’t. I don’t. But let’s remember… as hellish as it may be, the rollercoaster ride eventually ends.

View Original Post and All Comments on Facebook

Go to Top