Quote from Tufte on the subject:
> ” In general, in a time-series, use a baseline that shows the data not the zero point. If the zero point reasonably occurs in plotting the data, fine. But don't spend a lot of empty vertical space trying to reach down to the zero point at the cost of hiding what is going on in the data line itself.”
https://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=...
A zero baseline for this chat is meaningless and would render the plot useless, unless the purpose of the graph was to show what happens to your body temperature as you turn into a White Walker.
Similarly, the axis chosen for the mortality chart is effective since it shows the typical range of values (and thus does convey the scale effectively), the consistent seasonal trends, and the anomaly (which is very significant in terms of standard deviations above the mean). Setting a zero y-axis would squeeze the graph into 1/3 of the vertical space, leaving 2/3 as useless white space for no benefit at all.
I am comparing the Y values for multiple points with the same X values in different series. I'm only looking for course differences.
For my purpose, it makes sense for the scale to start at zero.
At a glance, the week of April 12, 2020 looks like it has four times as many deaths as any other year since 2015. In reality, it has roughly 1.5x as many. That's still a huge difference, but I had to check the scale and do some quick mental calculations to figure that out - which is more than I would have had to do if the graph's vertical scale started at zero.
I'm not arguing that all graphs must start their scales with zero. I'm not even saying that this is a bad graph. It's just not as effective for the kind of analysis I want to use it for.
[1] https://ourworldindata.org/excess-mortality-covid#excess-mor...
Imagine you had a graph of earth's temperature and decided to set the y axis at absolute zero. What would that look like? Would it be in any way useful?
Or, another way, % change compared to the average:
I would not take that much comfort given the surge in case counts. I don't doubt we are better handling cases now vs April, at the same time hospitals will soon hit the triage stage.
Positive tests tend to trail exposure by 1-2 weeks, while hospitalizations trail another week or two behind positive tests. Deaths can lag even further (not sure on the numbers here). So we're only just getting into the beginning of the post-Thanksgiving spike impacting hospitals.
While most serious cases hit those with compounding factors (age, health, etc.) if you look at the percent of excess deaths, it spans all age groups (again data ends early Nov):
https://ourworldindata.org/grapher/excess-mortality-p-scores...
The other good resource is John Hopkins has a chart overlaying date of policy changes vs cases (or deaths):
https://coronavirus.jhu.edu/data/state-timeline/new-confirme...
I would be surprised if the 2020 line doesn't jump right back up again once the rest of November and post-Thanksgiving numbers are added in given that we're seeing large numbers of reported deaths again in recent weeks.
https://www.cdc.gov/flu/pandemic-resources/1918-commemoratio...
How come number of Covid infection rates is rising, while flu seems to be eradicated? If the measures help against the flu, why not against Covid?
I don't like appeals to authority, but this perspective that people who have spent their entire lives studying epidemiology have just got this all wrong, don't understand what they are doing, and that appeals to "common sense logic" will reveal the truth ... really irritates me.
I mean, just as a basic starting point, most people have immunological exposure to the flu, so a fairly effective method to prevent spread of the flu virus will have outsized effects on the incidence of flu (even more so if that fairly effective method is even more widely utilized by those most at risk from flu). But no humans prior to the end of 2019 had immunological exposure to SARS-COV-2, and so any "leakage" in the efforts to stop its spread will have much greater impact than with influenza.
More generally, this is really basic epidemiology, and I don't understand why you think it's so useful to question this stuff in this way.
There is actually no clear cut scientific story about Sars-Cov-2 yet. That is part of the scare.
I also don't think your logic is sound - influenca used to spread rapidly through populations in all previous years, despite of previous immunological exposure as you describe. There is no reason to assume it would spread less rapidly this year because of "previous immunological exposure".
The effectiveness of masks and social distancing also does not depend on previous immunological exposure.
That's very understandable but I doubt your view would survive contact with the epidemiological literature.
It's really hard to believe I know, but epidemiology papers are all terrible. They seem to always contain basic errors that any lay person can spot, and peer review doesn't catch them, nor does the editing process at supposedly prestigious journals. I've read a lot this year and by now I go in to a new paper being sure I'll encounter something stupid or crazy, because the rate of problems is just so high.
Remember that the only people who study epidemiology their whole lives are in academia, a place where being correct is less important than being published. Although it sounds absurd, 2020 has convinced me that epidemiologists and public health researchers in general know absolutely nothing about disease. They are however very good at closing ranks and claiming nobody outside their little cliques should be allowed to criticise or question them.
More generally, this is really basic epidemiology, and I don't understand why you think it's so useful to question this stuff in this way.
There's nothing basic about the claim you just made, and I'm really curious now if you yourself are an epidemiologist. Because you've gone from asserting that epidemics aren't susceptible to "common sense logic" to saying it's obvious and common sense that lockdowns/masks - which have no observable impact on case curves for COVID at all - will have outside impact on influenza.
But imagine the "extreme" case where excess mortality is zero - that is, even with covid19, about the same number of people die every day.
That would mean, presumably, that since covid19 causes some deaths that would not have happened historically (it doesn't matter what you think that number is) it would have to be balanced by some decrease in "more normal" deaths.
Obviously, one could propose all kinds of mechanisms that might lead to this. One could suggest that because people are either disgusted by or rallying around Trump, death rates are different. One could suggest that the impending conjuction of Saturn & Jupiter on Dec 21st has reduced death rates from non-covid19 causes. And to be sure, at present, there doesn't seem to be a particularly good way to establish this.
But it also seems rather reasonable to say that since the biggest change in human behavior at this point has been driven by attempts to contain covid19, that it is likely these changes that have caused the "balancing" decrease in non-covid19 deaths.
One obvious rebuttal is that the graph does NOT show no excess mortality. This is certainly true. But the magnitude of the it is smaller than would be expected from even the most, ahem, conservative estimates of covid19 deaths. Ergo, there is some reduction in non-covid19 deaths, and occam's razor would put changes due to anti-covid measures near the top of the list of likely causes for that.
That would not be a given assumption (under your scenario of zero excess deaths). Old people who would otherwise have died from another reason could die from Covid19 instead.
Or imagine testing everybody for the common cold, and if somebody dies with the common cold virus, they would be counted as a "common cold death". Then you would see a lot of deaths "from the common cold", without any actual change in death rates.
For one, we don't even understand all the long term effects of having COVID as it is: there's heart damage, lung damage, brain damage, extended fatigue, and that's just what we're seeing evidence of now. [0]
If another disease like COVID strikes in a couple years, it could still tear through the surviving population and cause similar long lasting effects. The thousands that're suffering a terrible death daily, separated from their families and straining our healthcare systems to the breaking point--their deaths are not somehow making our populaitons safer from another disease.
EDIT: your analogy also supposes it's somehow the old/vulnerable populations that're the reservoir for the virus in our populations, when data indicates the opposite: it's far more likely younger populations are the carriers/vectors for COVID, especially since they are more often asymptomatic (w.r.t. acute effects) than more vulnerable chunks of the population. Future bugs will have plenty of population to travel through.
[0]: https://www.mayoclinic.org/diseases-conditions/coronavirus/i...
In the analogy the dry brush are the vulnerable that die. Once the dry brush is gone, it can't burn a second time. Just like small pox wiped out all the vulnerable native americans, but then the offspring of the survivors were increasingly more resilient. Except in this case the fatality rate is 1% (or less) instead of 90%.
> For one, we don't even understand all the long term effects of having COVID as it is: there's heart damage, lung damage, brain damage, extended fatigue, and that's just what we're seeing evidence of now.
This just seems like fear of the unknown. On the contrary, for the vast majority of healthy humans, it seems like the immune-system is well equipped to deal with the virus so there isn't much reason to fear IMO (i.e. most people recover at home with no modern medicine needed). I have several family members that had it and recovered with seemingly no other long term effects, so there is no reason to believe they are going to be screwed later in life.
> If another disease like COVID strikes in a couple years, it could still tear through the surviving population and cause similar long lasting effects.
More fear of the unknown. Novel viruses don't come along very often. Getting sick is not a bad thing, natural selection is not a bad thing, it's the very mechanism that has allowed humans to survive this long on planet earth.
No, the dry brush is what makes the fires get out of control and cause more damage than they "ought to," implying that if the brush is regularly dealt with then the broader population of trees will be better off.
But (a) we're not talking about trees, we're talking about people, and (b) letting swaths of the population die from a disease does not imply the surviving population will be better off in the future.
People can't die twice because we only have a single life, and callously saying "welp, you would have died later anyway, might as well die now so our stats look good" is so unbelievably short-sighted I can't believe I'm seeing this argument made by multiple people on HN.
> This just seems like fear of the unknown..
Uh, yes, it's fear of what the largely unknown disease will do to us. We shouldn't blindly assume that it's "not that bad" if it occasionally rips through the population off of some belief it'll make us safer if we're routinely exposed to pandemics--as follows from your brush-and-forest-fire analogy--when we don't understand the effects of those diseases.
> Natural selection is not a bad thing
I do not understand this point at all. Should we not have allergen warnings on food packaging? Or glasses for folks with bad eyesight? There are plenty of reasons we intervene in the "natural course" of events to make people's lives better. Citing "natural selection" is not in any way a valid defense for just letting diseases meander through populations.
Wait, am I making the analogy, or you? You can't just take my analogy and twist it until it fits your view. Make a new analogy, but don't correct mine.
> I do not understand this point at all. Should we not have allergen warnings on food packaging? Or glasses for folks with bad eyesight? There are plenty of reasons we intervene in the "natural course" of events to make people's lives better
Sure, but where do you draw the line? Should we move heaven and earth to artificially prolong the life of people extremely vulnerable to common diseases? Or just let them die and be replaced with more resilient humans?
My whole point is your analogy is flawed and you shouldn't use it for this purpose because it trivializes significant components of the issue at hand. You said:
> Once all the old/vulnerable wood is burned up, it's much more difficult for another forest fire of the same magnitude to happen ...
Which is demonstrably wrong when you liken _people_ to the dry brush, which is exactly what you did. We aren't optimizing to minimize deaths in individual years (or else we'd just kill everyone right now and the death rate would spike and then drop to zero), we're optimizing for general well-being of the population.
Your analogy is flat wrong, which is why I'm poking holes in it. It also belies a shocking lack of empathy for your fellow humans, which I frankly find alarming.
> Where do you draw the line?
I have no idea, and I don't think this is the venue to argue it, but I definitely don't think we're at it yet, and I would much rather err on the side of trying too hard to protect people's lives than not trying hard enough.
For example, when death numbers from COVID were first being released, people compared them to car fatalities, going "Why are we so worried? Car crashes kill 100x as many people in the same timeframe." The point is not "we should not worry about COVID because other things are more deadly" but "there is an alarming new source of deaths, and we should be working to reduce death as much as we can in the interests of maximizing the population's happiness."
EDIT: I mean, consider, the sitting US president called COVID "a hoax" for a good two months, then kept downplaying it for another two ("it'll be gone after November, you'll see"), and overall completely failed to make a plan at the federal level to, I dunno, provide more sick leave for people so they could actually avoid spreading it to their co-workers, or rent assistance for those out-of-work because we shut down the industries with high levels of communication, or support just wearing a mask when you go outside. We are so far and away from even a reasonable level of response to this it seems wholly disingenuous to pose the "but where do we _stop_ intervening?" argument.
I disagree. If we were optimizing for general well-being of the population we wouldn't be doing things that will push hundreds of millions into poverty[0] and drive up suicide rates[1].
From my perspective all the evidence points to us optimizing either for healthcare capacity or for minimizing covid-19 deaths.
> "there is an alarming new source of deaths, and we should be working to reduce death as much as we can in the interests of maximizing the population's happiness."
Well, clearly a huge % of the population is not happy with the current actions and policy being performed in the name of saving lives as evidenced by this thread.
[0] https://unu.edu/media-relations/releases/covid-19-fallout-co...
[1] https://www.cbsnews.com/news/japan-suicide-coronavirus-more-...
Fair enough, I wasn't clear: we _should_ be optimizing for general well-being, and among other things I think minimum-effort disease-prevention (masks, distancing, reasonable sick leave) are completely reasonable asks.
Similarly, we probably shouldn't be relying on a mega-consumerist culture that, when suppressed for whatever reason, has far-reaching economic effects that doom millions in developing countries to hunger/whathaveyou.
But these issues are not in opposition, we just have a crappy system that doesn't deal with both of them well at the same time. I think it's unreasonable to frame it as "well we can only address these issues in degrees, lest we make the other worse" and not question the system that somehow frames disease prevention and feeding people as opposing goals.
It's certainly possible to address this disease without such drastic impacts, but it relies on the population having an empathic mentality. Look at Japan, as an example[0]: Similar in terms of development, but with a much less invasive response they've only suffered around 2.3k deaths _total_, when the US is already seeing that number daily.
You are completely correct that "a huge % of the population is not happy with the current actions and policy being performed," because we have examples of cultures that are weathering this storm FAR better than those of us in the USA are.... but I think attitudes like yours are contributing to our failure to deal with this pandemic well.
[0]: https://www.economist.com/asia/2020/12/12/the-japanese-autho...
[1] https://ourworldindata.org/spanish-flu-largest-influenza-pan...
> Other large influenza pandemics
> The Spanish flu pandemic was the largest, but not the only large recent influenza pandemic. Two decades before the Spanish flu the Russian flu pandemic (1889-1894) is believed to have killed 1 million people.
> Estimates for the death toll of the “Asian Flu” (1957-1958) vary between 1.5 and 4 million. Gatherer (2009) published the estimate of 1.5 million, while Michaelis et al. (2009) published an estimate of 2–4 million.
> According to a WHO publication the “Hong Kong Flu” (1968-1969) killed between 1 and 4 million people.
> Michaelis et al. (2009) published a lower estimate of 1–2 million.
> The Russian Flu pandemic of 1977-78 was caused by the same H1N1 virus that caused the Spanish flu. According to Michaelis et al. (2009) around 700,000 died worldwide.
> What becomes clear from this overview are two things: influenza pandemics are not rare[...]
Heck, ebola was a decade ago with 300k deaths. Novel viruses appear all the time.
The other side of the problem is that lawmakers have no incentive to set aside money/time/resources for pandemic preparedness.
A quote from HHS Secretary under Bush, Mark Leavitt sums up part of the problem nicely:
> “In advance of a pandemic, anything you say sounds alarmist. After a pandemic starts, everything you’ve done is inadequate.”
Another example of lawmakers having no incentive to think about the future:
>The Public Health Emergency Preparedness program’s funding has gone from about a billion dollars in 2003 to $675 million this year, while the Hospital Preparedness program has gone from more than $500 million at its peak in 2004 to less than $300 million today.
Fairly frequently.
> So far it's once-in-a-lifetime (or less)...
No, it's not, even if you restrict yourself to novel betacoronaviruses that become significant human public health concerns (e.g., SARS, MERS, SARS-CoV-2).
According to the CDC, there seem to be about 12% more deaths than the average of the last three years, or about 300K. Number of deaths has been rising every year, though, sometimes increasing by up to 90k from one year to the next.
So that's 275k-169k=106k extra deaths, which is a napkinmath'd 17% increase over the baseline (assuming that's 106k deaths over a 12wk period, and baseline is ~50k/week).
A 17% rise in extra deaths seems pretty dramatic to me, regardless of the graphs.
Granted, I'd love to see error bars on this stuff, but I don't think the axis starting at 0 is some nefarious plot to dramatize the data.
Absolute number of deaths have been rising from year to year, sometimes with jumps of 90k.
Sure Covid has an impact, but whether it is dramatic is another question.
I thought that _was_ the quesiton: you seemed to imply the graph not starting at 0 was the article editorializing in some drama. My point was the situation is indeed dramatic (perhaps "significant" is a better phrase?) on its own, and they weren't unfairly exaggerating it for clicks/attention/fearmongering/whatever.
It makes sense that 'absolute deaths' would rise as a function of absolute population size, but that's why this data is important to pay attention to: if our death rate is climbing more than expected, there's problems we should probably pay attention to.
Oddly this year the CDC baseline expected death rate diverged from the trend line:
https://twitter.com/Humble_Analysis/status/13356752493633536...
That looks bad. If they were projecting the long term trend forward as they did in prior years, excess deaths from COVID in the USA would suddenly become a lot lower.
One other thing to bear in mind - a lot of these graphs that make excess deaths look dramatic are only looking 5 years in the past, because that's generally the easiest data to get hold of. But death rates are falling with time, as you'd hope to see. If you look further back you don't have to go far to find years with similar death rates. E.g. the UK had one of the highest excess death rates in the world from COVID in the first wave, but when you look at historical data, it was no worse than the the winter of 1999/2000 when no special measures were employed.
Lockdowns are an ahistorically extreme move. For that to make sense the scale of the problem with have to be equally ahistoric but it's not. I've lived through years with excess death rates just as bad as 2020 and never even noticed, because nobody was commenting about it. That's the risk with very short term analysis.
This page has a number of good breakdowns, but here's the US chart comparing 2020 vs other years:
https://ourworldindata.org/excess-mortality-covid#excess-mor...
Can you spot the anomaly?