Which States are Under- and Over-Reporting COVID Deaths in 2022?

I ran the numbers to compare excess deaths to reported COVID deaths for each state in 2022 (through June 18), with the idea this would give us a sense of which states are erring on the side of counting deaths as COVID deaths when they aren't, and which are erring on the side of not counting things as COVID deaths that are. I'd done this before, for the whole run of the pandemic, but this is just for 2022, so it should better reflect current practices.

So, for example, Vermont reported a total of 255 COVID deaths in that period. Yet, when you compare the number of people who died in Vermont in that period to the number who would have died based on pre-pandemic mortality rates, they suffered 658 more deaths than expected. So, admitted COVID deaths only account for 38.75% of excess deaths, suggesting they're under-reporting COVID deaths (e.g., a lot of deaths that wouldn't have happened if not for COVID are not being counted as COVID deaths).

At the other end of the spectrum, Rhode Island reported 584 COVID deaths in that time period, while only suffering 48 more deaths than they would have if they'd had the same mortality rates as they were having before the pandemic. Thus, their identified COVID deaths are 1,216.67% as many as would be needed to explain their excess deaths, suggesting they're over-reporting (e.g., counting as COVID deaths some people who happened to die with COVID, but would have died anyway).

When you look at the pandemic as a whole, most states (and the nation as a whole) were significant under-reporters, to the points we probably had a lot more COVID deaths than the official data acknowledges. We were missing huge numbers of COVID deaths early on, when there was little testing. But, in 2022, the counts are almost exactly right for the nation as a whole (179,585 identified COVID deaths, versus 168,625 total excess deaths). As a whole, our methods have improved, and we're counting just slightly more COVID deaths than would be needed to explain our elevated mortality rate.

Interestingly, there no longer appears to be a meaningful tendency for red states to undercount while blue states overcount (which happened earlier in the pandemic). The top under-counters are VT, DE, AK, CT, OR, UT, MT, NH, CO, and AR. The top over-counters are ND, RI, NE, MA, SD, HI, IA, MN, NJ, and MO. So, it's about an equal red/blue split on both ends. Whatever is leading to over- and under-counting these days doesn't seem to have a consistent political bias the way it once did.

AZ's gov. admits he's not reporting shit about Covid anymore. Our hospitals are trying to get out the message that spread is high and people are again hospitalized more.

Here's what the data looks like for each (note North Dakota actually had lower-than-expected deaths in 2022, so it sorts funny, but it's an over-reporter, in that it had no excess deaths yet counted 381 COVID deaths. Posted in three units due to character limit.

Ginger Ale said:

AZ's gov. admits he's not reporting shit about Covid anymore. Our hospitals are trying to get out the message that spread is high and people are again hospitalized more.

Click to expand...

That's consistent with my data, which has Arizona as the 12th biggest "under-reporter." Through June 18, in 2022 Arizona has suffered about 5,242 more deaths than expected (i.e., relative to what you'd have suffered if you'd been at your pre-pandemic averages), but has only admitted to 4,661 COVID deaths, leaving 581 extra deaths, or about 11%, unaccounted for.

Mina said:

I ran the numbers to compare excess deaths to reported COVID deaths for each state in 2022 (through June 18), with the idea this would give us a sense of which states are erring on the side of counting deaths as COVID deaths when they aren't, and which are erring on the side of not counting things as COVID deaths that are. I'd done this before, for the whole run of the pandemic, but this is just for 2022, so it should better reflect current practices.

So, for example, Vermont reported a total of 255 COVID deaths in that period. Yet, when you compare the number of people who died in Vermont in that period to the number who would have died based on pre-pandemic mortality rates, they suffered 658 more deaths than expected. So, admitted COVID deaths only account for 38.75% of excess deaths, suggesting they're under-reporting COVID deaths (e.g., a lot of deaths that wouldn't have happened if not for COVID are not being counted as COVID deaths).

At the other end of the spectrum, Rhode Island reported 584 COVID deaths in that time period, while only suffering 48 more deaths than they would have if they'd had the same mortality rates as they were having before the pandemic. Thus, their identified COVID deaths are 1,216.67% as many as would be needed to explain their excess deaths, suggesting they're over-reporting (e.g., counting as COVID deaths some people who happened to die with COVID, but would have died anyway).

When you look at the pandemic as a whole, most states (and the nation as a whole) were significant under-reporters, to the points we probably had a lot more COVID deaths than the official data acknowledges. We were missing huge numbers of COVID deaths early on, when there was little testing. But, in 2022, the counts are almost exactly right for the nation as a whole (179,585 identified COVID deaths, versus 168,625 total excess deaths). As a whole, our methods have improved, and we're counting just slightly more COVID deaths than would be needed to explain our elevated mortality rate.

Interestingly, there no longer appears to be a meaningful tendency for red states to undercount while blue states overcount (which happened earlier in the pandemic). The top under-counters are VT, DE, AK, CT, OR, UT, MT, NH, CO, and AR. The top over-counters are ND, RI, NE, MA, SD, HI, IA, MN, NJ, and MO. So, it's about an equal red/blue split on both ends. Whatever is leading to over- and under-counting these days doesn't seem to have a consistent political bias the way it once did.

Click to expand...

Horse hockey! I've read the instructions on counting Covid deaths and they are geared toward counting everything as Covid. People who are admitted for serious illness, are tested for covid and if they die of heart failure they put covid as a contributing factor whether they die of covid or not. That's been the case for the whole covid period. Liberals think if you got run over by a dump truck but tested positive for Covid while in emergency adn you die, that's a Covid death.
Just like the claim of a Covid vaccine, no vaccine exist. If you get the shot, you will not get Covid, how many times were we told that? The shot isn't a vaccine, it's a therapeutic. Polio has a vaccine, take the shot no polio.

When you pay a facility extra for a a Covid case or death, what do you think they migth claim

Integrityrespec said:

Horse hockey! I've read the instructions on counting Covid deaths and they are geared toward counting everything as Covid. People who are admitted for serious illness, are tested for covid and if they die of heart failure they put covid as a contributing factor whether they die of covid or not. That's been the case for the whole covid period. Liberals think if you got run over by a dump truck but tested positive for Covid while in emergency adn you die, that's a Covid death.
Just like the claim of a Covid vaccine, no vaccine exist. If you get the shot, you will not get Covid, how many times were we told that? The shot isn't a vaccine, it's a therapeutic. Polio has a vaccine, take the shot no polio.

Click to expand...

I had to check to see if this was a post from 2020.

Integrityrespec said:

When you pay a facility extra for a a Covid case or death, what do you think they migth claim

Click to expand...

It's easy to see how the big lie was so easily pulled off by Trump. We have some vulnerable, ignorant people out there.

Integrityrespec said:

Horse hockey! I've read the instructions on counting Covid deaths and they are geared toward counting everything as Covid.

Click to expand...

Whose instructions, specifically? Do you have a link?

Anyway, regardless of what the nominal instructions are, there are VAST differences from one place to another, in terms of the ratio of admitted COVID deaths to excess deaths. What do you think explains that disparity?

Like take two comparably sized states: Arizona and Massachusetts. So far this year, 38,116 people died in Arizona and only 31,667 in Massachusetts. Much of that difference, though, was driven by background matters predating the pandemic, e.g., more elderly retirees in Arizona, such that MA only expected 30,200, and Arizona expected 32,874. But, even counting that, Arizona had 5,242 more than expected, whereas MA had just 1,467 more than expected.

So, why so many more excess deaths in Arizona? It was 257% more than Massachusetts, for a population that's only about 4.2% higher. What's killing all those extra Arizona residents? Our first guess would be COVID, but they only counted 4,661 COVID deaths, meaning there's a huge gap -- 581 extra deaths this year that AREN'T accounted for by way of their official COVID death count. Meanwhile, Massachusetts goes the other way entirely. They only have 1,467 excess deaths, even though they officially counted 3,106 caused by COVID.

I think it's pretty clear they're not all using the same definitions, or at least not applying them the same way. Some states, like Massachusetts, are erring the side of inclusion (counting something as a COVID death if COVID was even just a complicating factor), while others, like Arizona, are erring on the side of exclusion (only counting it as a COVID death if they are thoroughly convinced the person would have survived if not for COVID).

Integrityrespec said:

That's been the case for the whole covid period. Liberals think if you got run over by a dump truck but tested positive for Covid while in emergency adn you die, that's a Covid death.

Click to expand...

Why is it, then that excess deaths during COVID have been significantly higher than officially recognized COVID deaths? If you were right, we'd expect to be counting a lot more COVID deaths than excess deaths (e.g., that person hit by the dump truck wouldn't increase the excess death count, since people aren't getting hit by dump trucks any more often than previously). Instead, it has generally gone the other way. Why do you think that is?

Integrityrespec said:

Just like the claim of a Covid vaccine, no vaccine exist.

Click to expand...

Several COVID vaccines exist. However, there was a meme circulating among the low-IQ set that suggests that something can't be called a "vaccine" unless it makes everyone immune to the disease (a standard that never applied to any other vaccine in history). If we applied that standard elsewhere then, for example, there'd also be no such thing as a polio vaccine, since polio vaccines also have breakthrough cases.

Integrityrespec said:

Horse hockey! I've read the instructions on counting Covid deaths and they are geared toward counting everything as Covid. People who are admitted for serious illness, are tested for covid and if they die of heart failure they put covid as a contributing factor whether they die of covid or not. That's been the case for the whole covid period. Liberals think if you got run over by a dump truck but tested positive for Covid while in emergency adn you die, that's a Covid death.
Just like the claim of a Covid vaccine, no vaccine exist. If you get the shot, you will not get Covid, how many times were we told that? The shot isn't a vaccine, it's a therapeutic. Polio has a vaccine, take the shot no polio.

Click to expand...

Your failed attempt to shit on Mina's well-researched claims is a joke. Put evidence up that proves her wrong or admit that you have nothing here.

Mina, are you treating deaths as a static equation with Covid as the only variable?
.People are not living as they were in years prior to Covid or from year to year during Covid,
so while appreciating the effort it may be misleading to assume a similar death rate (or percentage of total deaths) in every category year over year for the past 3 years.

JH Oldoc said:

Mina, are you treating deaths as a static equation with Covid as the only variable?
.People are not living as they were in years prior to Covid or from year to year during Covid,
so while appreciating the effort it may be misleading to assume a similar death rate (or percentage of total deaths) in every category year over year for the past 3 years.

Click to expand...

It's true that changes in behavior could result in the pandemic influencing mortality rates indirectly. For example, if people are working from home more often, we'd expect fewer traffic fatalities associated with commutes, which would be expected to drive mortality rates down. Similarly, if masking and distancing meant less transmission of the flu, we'd expect fewer flu deaths, also pushing mortality rates down.

I'm not arguing those things didn't happen. It's just hard to untangle them. Like why were there so very many more excess deaths that aren't accounted for by official COVID-death stats in Arizona than there were in Massachusetts? Was it because Arizona was more reluctant to label something a COVID death, and thus undercounted badly (while Massachusetts overcounted)? That's my guess for the lion's share of the disparity. But it could be other things. Like maybe Arizona suffered a spate of deaths from Ivermectin and hydroxychloroquine reactions, as people experimented with alternative means of fighting the virus.

I guess what it would take to be really sure is to break down death rates across a large group of causes, and see where the bumps were. If, for example, Arizona had most of that extra excess death because they were undercounting COVID deaths, then we'd expect to see a relatively larger bump in various natural deaths from conditions that are known to be exacerbated by COVID (e.g., heart disease), relative to other states that had fewer unexplained excess deaths (i.e., a data fingerprint suggesting more people were dying of heart attacks associated with a prior COVID infection, in Arizona than in those other states). Meanwhile, we might see that other deaths not known to have a meaningful COVID link (drownings, deadly falls, electrocutions) only rose or fell roughly in proportion to what was happening in other states. It would be a tough calculation, but worth doing for someone with the time and inclination.

Ultimately, we may never have a definitive answer. Even something like the huge 2020 murder rate spike could have a fairly direct COVID link (COVID is known to shrink the brain, so could it have impaired impulse control, or inflamed violent impulses, in states where there were more infections?) We'd see that as a behavioral change, but it could be a medically-induced behavioral change.

Mina said:

It's true that changes in behavior could result in the pandemic influencing mortality rates indirectly. For example, if people are working from home more often, we'd expect fewer traffic fatalities

Click to expand...

Though just to point out how difficult it is to make any assumptions, traffic deaths in the US actually went up the last 3 years.
I was surprised to find this, though there might be ready explanations.

...anyway, the only point I wanted to make was that it's a massive equation after 2019. 2019 over 2018 it might be somewhat useful - after....not too much.

Integrityrespec said:

Horse hockey! I've read the instructions on counting Covid deaths and they are geared toward counting everything as Covid. People who are admitted for serious illness, are tested for covid and if they die of heart failure they put covid as a contributing factor whether they die of covid or not. That's been the case for the whole covid period. Liberals think if you got run over by a dump truck but tested positive for Covid while in emergency adn you die, that's a Covid death.
Just like the claim of a Covid vaccine, no vaccine exist. If you get the shot, you will not get Covid, how many times were we told that? The shot isn't a vaccine, it's a therapeutic. Polio has a vaccine, take the shot no polio.

Click to expand...

Perfect example of clueless posting, fueled by extremists who are scared of masks and vaccines but want to dictate what a woman must do with her body.

Mina said:

I ran the numbers to compare excess deaths to reported COVID deaths for each state in 2022 (through June 18), with the idea this would give us a sense of which states are erring on the side of counting deaths as COVID deaths when they aren't, and which are erring on the side of not counting things as COVID deaths that are. I'd done this before, for the whole run of the pandemic, but this is just for 2022, so it should better reflect current practices.

So, for example, Vermont reported a total of 255 COVID deaths in that period. Yet, when you compare the number of people who died in Vermont in that period to the number who would have died based on pre-pandemic mortality rates, they suffered 658 more deaths than expected. So, admitted COVID deaths only account for 38.75% of excess deaths, suggesting they're under-reporting COVID deaths (e.g., a lot of deaths that wouldn't have happened if not for COVID are not being counted as COVID deaths).

At the other end of the spectrum, Rhode Island reported 584 COVID deaths in that time period, while only suffering 48 more deaths than they would have if they'd had the same mortality rates as they were having before the pandemic. Thus, their identified COVID deaths are 1,216.67% as many as would be needed to explain their excess deaths, suggesting they're over-reporting (e.g., counting as COVID deaths some people who happened to die with COVID, but would have died anyway).

When you look at the pandemic as a whole, most states (and the nation as a whole) were significant under-reporters, to the points we probably had a lot more COVID deaths than the official data acknowledges. We were missing huge numbers of COVID deaths early on, when there was little testing. But, in 2022, the counts are almost exactly right for the nation as a whole (179,585 identified COVID deaths, versus 168,625 total excess deaths). As a whole, our methods have improved, and we're counting just slightly more COVID deaths than would be needed to explain our elevated mortality rate.

Interestingly, there no longer appears to be a meaningful tendency for red states to undercount while blue states overcount (which happened earlier in the pandemic). The top under-counters are VT, DE, AK, CT, OR, UT, MT, NH, CO, and AR. The top over-counters are ND, RI, NE, MA, SD, HI, IA, MN, NJ, and MO. So, it's about an equal red/blue split on both ends. Whatever is leading to over- and under-counting these days doesn't seem to have a consistent political bias the way it once did.

Click to expand...

.disease like this always kills far far more than are officially counted. We had nearly a million dead after just the first year. The official number was nowhere near

JH Oldoc said:

Though just to point out how difficult it is to make any assumptions, traffic deaths in the US actually went up the last 3 years.
I was surprised to find this, though there might be ready explanations.

Click to expand...

Yes, I'd seen that. One possibility is based simply on lower traffic meaning higher speeds. Like the lowest traffic death rate per passenger mile is in Massachusetts, which presumably has less to do with them being super skillful drivers, and more to do with the fact that high density means for the most part vehicles aren't moving that fast and accidents aren't deadly.

JH Oldoc said:

...anyway, the only point I wanted to make was that it's a massive equation after 2019. 2019 over 2018 it might be somewhat useful - after....not too much.

Click to expand...

Not sure what you're saying there.

noonereal said:

.disease like this always kills far far more than are officially counted. We had nearly a million dead after just the first year. The official number was nowhere near

Click to expand...

I think that was definitely true early on, but for 2022 it appears to no longer be true -- the official COVID death count is very close to the number of excess deaths.

Mina said:

I think that was definitely true early on, but for 2022 it appears to no longer be true -- the official COVID death count is very close to the number of excess deaths.

Click to expand...

If we use 5he sake criteria as used during previous pandemics (1918) it is not true, that is my point.

Mina said:

Yes, I'd seen that. One possibility is based simply on lower traffic meaning higher speeds. Like the lowest traffic death rate per passenger mile is in Massachusetts, which presumably has less to do with them being super skillful drivers, and more to do with the fact that high density means for the most part vehicles aren't moving that fast and accidents aren't deadly.

Not sure what you're saying there.

Click to expand...

Pardons - I was meaning 2020 over 2019. For a good portion of 2020 the public and government reaction to covid wasn't as it became later on, so there weren't as many behavior modifications in 2020 as in 2021. Remember, we had a president who said it was a cold and others saying it was a hoax so reaction was delayed.

Mina said:

I ran the numbers to compare excess deaths to reported COVID deaths for each state in 2022 (through June 18), with the idea this would give us a sense of which states are erring on the side of counting deaths as COVID deaths when they aren't, and which are erring on the side of not counting things as COVID deaths that are. I'd done this before, for the whole run of the pandemic, but this is just for 2022, so it should better reflect current practices.

So, for example, Vermont reported a total of 255 COVID deaths in that period. Yet, when you compare the number of people who died in Vermont in that period to the number who would have died based on pre-pandemic mortality rates, they suffered 658 more deaths than expected. So, admitted COVID deaths only account for 38.75% of excess deaths, suggesting they're under-reporting COVID deaths (e.g., a lot of deaths that wouldn't have happened if not for COVID are not being counted as COVID deaths).

At the other end of the spectrum, Rhode Island reported 584 COVID deaths in that time period, while only suffering 48 more deaths than they would have if they'd had the same mortality rates as they were having before the pandemic. Thus, their identified COVID deaths are 1,216.67% as many as would be needed to explain their excess deaths, suggesting they're over-reporting (e.g., counting as COVID deaths some people who happened to die with COVID, but would have died anyway).

When you look at the pandemic as a whole, most states (and the nation as a whole) were significant under-reporters, to the points we probably had a lot more COVID deaths than the official data acknowledges. We were missing huge numbers of COVID deaths early on, when there was little testing. But, in 2022, the counts are almost exactly right for the nation as a whole (179,585 identified COVID deaths, versus 168,625 total excess deaths). As a whole, our methods have improved, and we're counting just slightly more COVID deaths than would be needed to explain our elevated mortality rate.

Interestingly, there no longer appears to be a meaningful tendency for red states to undercount while blue states overcount (which happened earlier in the pandemic). The top under-counters are VT, DE, AK, CT, OR, UT, MT, NH, CO, and AR. The top over-counters are ND, RI, NE, MA, SD, HI, IA, MN, NJ, and MO. So, it's about an equal red/blue split on both ends. Whatever is leading to over- and under-counting these days doesn't seem to have a consistent political bias the way it once did.

Click to expand...

I find it hard to believe that in any time period VT had nearly 10 times the death rate from Covid that RI had, so perhaps using "excess deaths" as a benchmark has its problems.

Tlrmln said:

I find it hard to believe that in any time period VT had nearly 10 times the death rate from Covid that RI had, so perhaps using "excess deaths" as a benchmark has its problems.

Click to expand...

It probably has a few problems, mostly associated with the way it would be expected to break down with smaller numbers. That would tend to happen both for very small populations, where the data would be inherently noisy (e.g., Vermont in 2022), where, say, a bad breakout at a single big facility could double the whole state’s COVID death rate for a period. It would also include cases where COVID death tolls get low enough that they start to get overwhelmed by unrelated changes in background mortality (e.g., RI this year). If COVID deaths are fairly high relative to pre-pandemic mortality, they’d be expected to be a driving force in excess mortality. But if they’re small in those relative terms, they’ll get lost in the noise.

For example, if a given state was having plus or minus a hundred deaths in given six month periods even before the pandemic, then if they look like they’ve had a hundred excess deaths so far this year, it could suggest zero COVID deaths or 200 COVID deaths, which are both within that margin of error of a hundred. So, it’s basically worthless for figuring anything out with that fine-grained detail. But if instead it has 1,000 excess deaths in that period, that suggests a lot more is going on than just the usual noise in the data.

If we were to chart the states with “uncertainty bands,” probably a lot of them would have overlapping band, especially for shorter periods of time (like YTD).

Integrityrespec said:

When you pay a facility extra for a a Covid case or death, what do you think they migth claim

Click to expand...

You know, IF the United States of America was the ONLY country where COVID-19 was active, THEN you MIGHT have a wonderfully valid point.

Unfortunately for your position, there are around 229 OTHER countries where COVID-19 is active.

Phys251 said:

Your failed attempt to shit on Mina's well-researched claims is a joke. Put evidence up that proves her wrong or admit that you have nothing here.

Click to expand...

You do realize that your challenge is directed to someone who is operating on the premise that a falsehood that is repeated 10 times is 550% truer than a non-falsehood that is repeated once - don't you?