7 ms·
Other people have been commenting to this effect, but I'd take a step back and make a broader point: when something makes a bold claim (i.e. "No excess deaths i
by hn_throwaway_99 6y ago
Other people have been commenting to this effect, but I'd take a step back and make a broader point: when something makes a bold claim (i.e. "No excess deaths in Sweden compared to the last ten years") using deliberately obfuscated data, we should call it what it is: lying with statistics.
That is, the graph you would really want to look at is one that shows weekly deaths as a line graph over many previous years, that way it's easy to see, at any week in the year, whether the current year is higher or lower than previous years or averages. That is what the Financial Times graphs show that another commenter posted, and they very clearly show the excess deaths in 2020 vs previous years for Sweden (and many other countries).
What's the point of the graphs in this post? I mean, 2020 isn't even done yet, so the graphs literally tell you nothing about Covid because there are no previous graphs to compare to. So what's the point of posting this, with this title, besides being deliberately deceptive?
- stmfreak 6y agoWhether weekly or annual, you wouldn’t be wrong to expect 2020 to have a noticeable uptick from Covid deaths. The fact that it does not points to one of two things: reporting deaths has a lag time or, covid didn’t kill an unusual number of people compare to flu and pneumonia. I will point out that the US CDC tracks excess deaths on a weekly basis and you can clearly see we are above average from March until the present, as well as for the year overall.
- meroes 6y agoToo black and white. You need some measure of general population health, ICU capacity, access to affordable healthcare, etc to compare two populations' strategies and effects of covid. Any of these and more could easily be a confounding variable. Covid may not have produced excess deaths so far in Sweden compared to US, but maybe they have better work from home and thus fewer car accidents compared to a larger portion of people driving to work in the US.
- hn_throwaway_99 6y agoI think you may be missing my point, though. If you look at the set of graphs that another commenter posted that compare weekly rates, https://www.ft.com/content/a2901ce8-5eb7-4633-b89c-cbdf5b386938 https://www.ft.com/content/a2901ce8-5eb7-4633-b89c-cbdf5b386... (go halfway down the page), you'll see that Sweden HAS produced excess deaths so far in 2020: 20% by this estimate, vs. 24% for the US.
- rebuilder 6y agoIncidentally, many countries in the excess deaths graphic on FT seem to have had pretty serious spikes in mortality early in the year in previous years. Look at Germany, for example - something much deadlier than covid-19 seems to have struck them in a recent year. What is that?
- hnarn 6y ago> something much deadlier than covid-19 Don't use this type of phrasing. The only way to know how deadly Covid-19 is compared to other unidentified changes in excess morality is to have access to an alternate reality where we did nothing about the spread. In our reality, we did a few pretty drastic things about the spread, which means that the comparison is essentially useless from the perspective of excess deaths. The only Covid-19 lethality comparison that makes sense is to compare the survival rate of those infected with those infected by another disease (or perhaps some other type of affliction).
- rebuilder 6y agoFair point. My intended meaning was to say the mortality rate in several countries had much higher spikes than the one for 2020.
- tialaramex 6y agoGiven it's in winter, probably an Influenza variant. Because Influenza (unlike some other causes of "flu" symptoms) is most widespread in winter, each hemisphere uses data from the other's winter to plan their seasonal vaccine. But this is like predicting the weather, even when you've had practice and are good at it on average, sometimes you miss badly. If you either guess wrongly what to vaccinate against based on the data from the other hemisphere, or Influenza mutates significantly after your cut off date to evade your profile, the vaccine is much less effective than normal, and higher than normal rate of hospitalizations and deaths results. Influenza is endemic, we never even wiped out the variant from the "Spanish Flu" pandemic last century, it's still with us is an attenuated form.
- JeremyNT 6y ago> The fact that it does not points to one of two things: reporting deaths has a lag time or, covid didn’t kill an unusual number of people compare to flu and pneumonia. Or people have modified their behaviors to avoid Covid, and thus are less likely to suffer deaths due to other causes of mortality (i.e., they drive less because of social distancing, they die of flu and other contagious diseases less because of social distancing, etc). If nobody modified their behaviors at all, you might have a point. But because people have modified their behaviors to avoid Covid, they're now trading one kind of risk for another.
- cpncrunch 6y agohttps://ourworldindata.org/excess-mortality-covid https://ourworldindata.org/excess-mortality-covid
- rainyMammoth 6y agoIf I read those graphs correctly, except during the peak of each waves, the death rates seems very similar to other years? ie, people dying of covid are people that would have died anyways from other causes.
- lesdeuxmagots 6y agoNot quite. The relatively similar all cause does not imply that it would be the same people dying. There could instead people different people dying. For example, there are much less people dying by automobile accidents (a leading cause of death for young people). In 2020, they could be replaced by very different types of death (COVID, Suicide, etc.) That said, it certainly COULD be relatively similar people. You’d likely see relatively similar individuals dying of pneumonia as dying from COVID.
- Barrin92 6y ago>ie, people dying of covid are people that would have died anyways from other causes. yes, with some exceptions (Spain, the UK, the US seem to register an actual significant spike in deaths). But overall it's true that the age of the average covid death is as high, or in some countries even higher than the average lifespan, meaning it's likely the person would have died of some other cause. I've always thought that not comparing deaths, but comparing 'number of high quality life years' lost or something similar would make more sense, even if it is somewhat more abstract.
- beagle3 6y agoThe term you are looking for a is QALYs - quality adjusted life years. It is a metric already used to e.g. rank public spending on expensive drugs and treatments.
- 6y ago
- jiofih 6y agoEspecially considering the death toll from Corona is about 6k people so far (which is a lot for a small country). It is about the same as all deaths caused by respiratory diseases in 2019: https://www.statista.com/statistics/528942/sweden-number-of-deaths-by-cause-of-death/ https://www.statista.com/statistics/528942/sweden-number-of-...
- thu2111 6y agoThe correct response to that statistic is to wonder why COVID appears to have so neatly replaced deaths from other respiratory viruses, and to wonder - given the relatively vague symptoms - if there's some reclassification going on. The wrong response is to ignore the point of the Swedish numbers specifically, even though the conclusions are painful. A reasonable conclusion isn't "oh that's terrible they're lying with statistics" and then proposing a meaningless alternative (nobody cares if COVID shifts expected mortality around by a few months, they only care about deaths that were avoidable and wouldn't have happened anyway). The correct response is to remember that Sweden was supposed to be a bloodbath in which every single hospital would be overflowing onto the streets, in which people would be collapsing unable to breathe in their hundreds of thousands whilst the rest of the world looked on in pity, safely locked down at home. Clearly that didn't happen, which invalidates the claims of the 'experts' that said it would. When you read their papers the problems that led to these erroneous claims are very clear, but given the top rated comment on this thread it'll probably be a long time before the wider HN community is able to critically examine the deep structural flaws in the field of epidemiology. If ever.
- ad404b8a372f2b9 6y agoCan you elaborate on the clear problems you mentions? I agree that the statistics were too easily dismissed but I don't understand how to reconcile them with the rest of the world where it was/is a bloodbath with overflowing hospitals.
- thu2111 6y agoThe rest of the world was NOT a bloodbath with overflowing hospitals. Look at the UK. Excess death once of the worst in Europe but not actually bad at all, lower than 1999/2000 in fact. Hospitals were ghost towns during these periods, cleared out with operations cancelled and people told to avoid attending to make way for a wave of COVID patients that never came. Same story is repeated all over the world, hence viral videos of dancing medics on TikTok. The hospitals were quiet, there was no bloodbath. See past the groupthink and mass hysteria: the stats don't show anything except a normal year made worse by lockdowns.
- johbjo 6y agoThe graph in the link is badly chosen to demonstrate this, but there is no excess mortality during january-september relative to previous years. It is clear from monthly data that from march to may there was excess mortality. From june to september the mortality was lower than previous years. The main effect of covid in mortality statistics is that it shortens lives by about 6 to 12 months, which causes a bump in mortality followed by a slump.
- hn_throwaway_99 6y ago> It is clear from monthly data that from march to may there was excess mortality. From june to september the mortality was lower than previous years. If you have a data set that clearly shows this, I would welcome a conversation about it. The data I have seen from the Financial Times site does not support this, but again, I'm more than happy to look at other data that backs this up.
- oceanplexian 6y agoIf a weekly chart demonstrated a different result, but excess deaths for the year didn't change, then a reasonable conclusion would be that COVID-19 brought some deaths forward in time, but not any further than a year. That makes plenty of sense. The vast majority of those who died of COVID-19 were in advanced age or had preexisting conditions. Some fraction of those people would have died even if there was not a pandemic. This is neither controversial nor disputed.
- tinus_hn 6y agoHow is this lying? Numbers don’t lie. If you extend this (taking the rest of the year to be 1 1/2 month) you end up with 89176. Too bad if that doesn’t fit your narrative.
- brmgb 6y agoNo, if you extend, you get 90365. You have to multiply by 12/10.5 but that's low balling because winter months usually have more deaths than summer months. Also deaths are usually reported monthly. I have no idea of how statista got the november figure and apparently I would have to subscribe to see the source.
- tinus_hn 6y agoThat is true, but the result still is not a remarkable number. My Swedish is not good enough to look up the official numbers.
- brmgb 6y agoI did in another comment in this thread. Sweden has a website in English. Excess deaths are impossible to miss.
- tinus_hn 6y agoSo I calculated the total from January 1 to November 9, the numbers are 2015 77033 2016 75524 2017 76522 2018 77089 2019 73374 2020 79088 So, the difference is about 2000. Noticeable, not as dramatic as 6000 but noticeable.
- brmgb 6y agoIdeally, I would like a rolling average of monthly death per capita on a year to smooth out seasonal variation on a suffisently large time span to substract any long-term trend. Sweden publishes enough data to do it actually [1] so I did [2]. There is a very strong downward trend in 2018, stability in 2019 then a literal a cliff in 03/2020 suddenly bringing you back to the 2017 figure. The average of deaths per capita (given per 10000) which was generally stable in 2019 jumps from 7.18 in february to 7.24 in march, 7.49 in april, 7.63 in may, 7.69 in june and then is more or less stable until september at which point I have no more data. So yes, the HN submission title is worthless and most likely intentionally misleading. The statistica data from 2020 are on 10 months. 2020 is going to be an outlier as soon as november and december are added. It's obvious even if you just multiply by 1.2 (december and november typically have more deaths than summer months). [1] https://www.scb.se/en/finding-statistics/statistics-by-subject-area/population/population-composition/population-statistics/ https://www.scb.se/en/finding-statistics/statistics-by-subje... [2] https://docs.google.com/spreadsheets/d/1tovNw9RReZ-8eF8y-JbnkEEQotJtSe1tmxo6dd9bLIM/edit?usp=sharing https://docs.google.com/spreadsheets/d/1tovNw9RReZ-8eF8y-Jbn...