47

From CNN

Dr. Tom Frieden, the former director of the US Centers for Disease Control and Prevention, laid out "10 plain truths" about Covid-19 on Wednesday as he spoke at a House Appropriations Committee hearing on the pandemic response. [...]

"Even now with deaths decreasing substantially, there are twice as many deaths from Covid-19 in New York City as there are on a usual day from all other causes combined," Frieden said.

Without getting into in-depth debates about how correctly those deaths are attributed to Covid-19, is what Frieden says correct, as far as the death statistics that have been recorded? There's no doubt excess mortality in NYC, as there is in similarly badly hit areas e.g. in Italy, but is the twice as many deaths attributed to Covid-19 as all other causes combined (at other times) true for NYC?

Community
  • 1
Fizz
  • 57,051
  • 18
  • 175
  • 291
  • 4
    Worth adding the specific date of the interview (or his datum point if different) to the question, considering the fluidity of the situation. I've seen some very misleading figures quoted that were only a few days out of date. – user_1818839 May 07 '20 at 17:58

2 Answers2

76

According to the New York Times, weekly deaths peaked around 7,000 in late March, whereas the average death rate was around 1,000. Reported COVID deaths for the period March 11-May 2 were around 18,000, whereas total deaths were around 23,000, 297% above normal. And, as can be seen from the chart, deaths still remain around 3,000.

enter image description here

Daniel R Hicks
  • 5,550
  • 3
  • 36
  • 38
  • Another NYT article: https://www.nytimes.com/interactive/2020/04/27/upshot/coronavirus-deaths-new-york-city.html – Daniel R Hicks May 07 '20 at 11:51
  • I'm always baffled by the "yeah, but more people die from X" every year..... it's not like not being the leading cause of death means those deaths don't happen. Lots of extra bodies would be a bad thing, regardless of rank. – PoloHoleSet May 07 '20 at 14:53
  • 41
    I wonder if, say around September or October, death rates will actually be lower than normal, because so many COVID deaths are related to co-morbidities that would have proved fatal in the near future. – Ask About Monica May 07 '20 at 15:43
  • 57
    @AskAboutMonica -- Most people over a certain age have what are classified as co-morbidities. Hypertension for example is very prevalent in the elderly and is counted as co-morbidity. Most people who lose their lives to COVID-19 are losing years from their expected life spans, rather than months. – antlersoft May 07 '20 at 16:27
  • 4
    @AskAboutMonica: "lower than normal"... depends what you want to consider "normal", but the influenza season does result in a cyclical yearly pattern of deaths, with a peak in winter. (The average over many years basically looks like a sine wave.) When one speaks of excess mortality, that's relative to the monthly or weekly average to the same month/week of the prior years (to account for this cyclicity). See some graphs in the other questions e.g. https://i.stack.imgur.com/WYtbC.png – Fizz May 07 '20 at 18:10
  • 8
    And antlersoft is correct, that for most age groups the loss of life-time (from Covid-19) is rather substantial. https://www.economist.com/graphic-detail/2020/05/02/would-most-covid-19-victims-have-died-soon-without-the-virus – Fizz May 07 '20 at 18:14
  • 15
    @AskAboutMonica some useful context, being over 65 and male are two risk factors. All told males who make it to 65 (including the >65% who have hypertension, 27% who have diabetes) are expected to live another 18 years. Males who make it to 90 are expected to live another 4.1 years. https://www.cdc.gov/nchs/data/nvsr/nvsr68/nvsr68_07-508.pdf – De Novo May 07 '20 at 18:21
  • 6
    From what I've heard, it _is_ possible that death rates will be below average in fall, but for a different reason: the measures we are taking against Corona (COVID-19) will also negatively impact the possibility of other viruses, such as influenza (flu) and rhinovirus (common cold), to spread. That is, provided we keep social distancing ourselves and don't let COVID re-ignite. – CompuChip May 08 '20 at 07:32
  • 1
    @AskAboutMonica It's just a mathematical fact that the death rates will at some point be lower. Imagine the extreme case that a disease would kill randomly half of the population. In this case the death rate would be halved obviously. Or image a disease killing all people over 60. The death rate afterward would be extremely low in this scenario. Of course these are extreme cases but they show that there must be an influence. – Jannick May 08 '20 at 08:35
  • 5
    Or even better - disease which kills entire population. It would then reduce amount of future deaths to zero forever. – Artur Biesiadowski May 08 '20 at 10:11
  • 3
    @ArturBiesiadowski Reductio ad absurdum! Vale! – Oscar Bravo May 08 '20 at 11:52
  • 2
    @Fizz Do note that the graph you linked to has Y-axis beginning at 50000. The graph in this answer also shows the normal yearly variation, but shows it in scale relative to total deaths - the normal yearly variation is only 20% or so. – jpa May 08 '20 at 12:54
19

There is a study from the Yale School of public health that attempted to study this carefully, but their analysis was restricted to March through April 4. Here is a newspaper article describing the study.

The basic answer to your question is "probably yes," but the situation is somewhat cloudy. The study found that excess deaths were about double the number recorded as covid. Many/most of these may actually be deaths due to covid that weren't recorded as such, but many may also be deaths of people who were afraid to go to the hospital, or who couldn't get appropriate care because hospitals were trying to clear the decks for covid patients. They may therefore have died from things like appendicitis or cancer that wasn't properly treated. For example, there are projections that the cancer death rate in England could go up 20% because people aren't getting proper care. There has been an unusually low number of people showing up in ERs with heart attacks, strokes, and other problems. This is not believed to be because people aren't having strokes. It's probably that they're having strokes but staying home, and either living or dying.

An additional complicating factor is that because people are staying home, they may not be dying as often from causes such as gunshots or traffic accidents.

For this reason, it is not a valid methodology just to subtract deaths from time-averaged deaths and conclude that you have measured the number of deaths from covid, as in the graph shown in the answer by Daniel R Hicks. That method of analyzing the data may overestimate deaths from covid, possibly by some fairly big percentage. However, the large spike shown in that graph (from a time period later than the one covered by the Yale study) is, I would think, almost certainly an indication that at least for some short time in NYC, the true covid death rate was a multiple of the normal death rate.

  • 14
    Arguably these are really deaths by covid. Not by the virus itself, but by the way it affected the medical system and society. – user000001 May 08 '20 at 06:22
  • 2
    Patients with heart attacks aren't showing up in ERs because it was removed from EMT protocol to bring them: https://www.nbcnewyork.com/news/coronavirus/grim-new-rules-for-nyc-paramedics-dont-bring-cardiac-arrests-to-er-for-revival/2356265/ Transport criteria at the height of the epidemic included pronouncing on scene for basically any non-traumatic cardiac arrest where return of pulse couldn't be achieved in field. – Tiercelet May 08 '20 at 13:34
  • 5
    @user000001 exactly, it's little different than a war where excess deaths is also the commonly used statistic. the only reasonable explanation for such a huge spike in deaths is the coronavirus, through both direct and indirect effects. Contrary to this answer, this is a perfectly reasonable methodology, though obviously people trying to minimize the virus' impact will disagree. – eps May 08 '20 at 16:05
  • 3
    Though arguably having to stay home with family could actually lead to an increase in gunshot (and knife, blunt instrument, poison &c) deaths, since most murder victims have social or romantic relationships with the people that killed them: https://www.bjs.gov/content/pub/press/MILUC88.PR – jamesqf May 08 '20 at 16:14
  • 1
    @Tiercelet: *Patients with heart attacks aren't showing up in ERs because it was removed from EMT protocol to bring them* No, you're misunderstanding. EMTs are being told not to try to revive people who have undergone cardiac arrest. A heart attack is not cardiac arrest. –  May 08 '20 at 20:13
  • 4
    @BenCrowell you're correct about cardiac arrest and heart attacks not being the same thing, but your summary of what the EMTs were/are being told is wrong. They *were* trying to revive people, but if they couldn't do it in the field, they weren't bringing them to the hospital after. – mbrig May 08 '20 at 20:22
  • 1
    I trained as an EMT in Illinois, under an instructor who had asked, nervous as a boy asking for a first date, whether the "CPR saves lives" slogan wasn't at least misleading. CPR is only appropriate after somene's heart has stopped, meaning that they're clinically, and legally, dead. Combine this with a healthy dose of "Perfect CPR doesn't happen [even with professionals]," meaning you can't make a mistake that will leave a patient more dead because that's already happened, and an assertion that massaging the heart is only 35% as effective as real heart action, and... – Christos Hayward May 09 '20 at 20:34
  • 1
    ...it means that CPR is basically holding on to a slender chance that an ER can get a patient back. As of my training paramedics had something like 10 drugs they could provide [with more options as authorized by the dispatcher], and EMT's had only one to offer, namely purer oxygen. Which leads me to my point: It is a known possibility that EMT's can administer CPR and have a patient revive, and I don't know if paramedics have access to a drug to jolt someone back, but mere EMT's (as of my training less qualified than paramedics) have an almost negligible chance of reviving a patient on-scene. – Christos Hayward May 09 '20 at 20:38
  • @ChristosHayward: Why weren't the EMTs also authorised to use those 10 or so drugs, then? – Vikki May 09 '20 at 23:32
  • @eps It's perfectly good methodology for discussing overall impact of the virus, but much less effective (almost useless) when discussing how lethal the virus is itself. This answer is pretty obviously framed as the latter, though, obviously people conflating things and being disingenuous will disagree. – Clay07g May 10 '20 at 14:39
  • 1
    @Sean: I suppose you can stay an EMT and nothing more, but an EMT is basically an apprentice paramedic. You can't become a paramedic in one fell swoop; you can train and become an EMT, and then after practicing with a certain number of hours of ambulance ride time, you can train further and become a paramedic. There's a literal and idiomatic difference in pay grade, and some fire departments prefer to have workers be firefighter II/paramedics (or on their way to reaching that status) across the board. – Christos Hayward May 11 '20 at 12:34
  • 2
    I think the simplest way to more truthfully portray these excess deaths is as "COVID19-*influenced* deaths". They may not have been from the disease itself, but the disease and social modifications around it were significant contributing factors to the deaths (people choosing to stay home with life-threatening conditions, beds/equipment/personnel not being available, ad nauseum) – Doktor J May 11 '20 at 19:48
  • @user000001 there is some amount of causal proximity required. Most people would not attribute millions of deaths in the 1910s to Bosnian assassins, even if mechanically there is some truth to that. Because there is COVID-19, hospitals shut down non-emergency care, so screenings dropped, so more people died of strokes... – Paul Draper May 07 '21 at 17:54