Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Almost third of UK Covid hospital patients readmitted within four months

theguardian.com

Post reply on HN
Print view View original

11–20 of 217 posts

Re: Almost third of UK Covid hospital patients readmitted within four months

#11

Post by azakai »

peytn wrote: post #6
morsch wrote: post #4

Earlier quoted context omitted.

Four times more often than the control group. Over a mean follow-up of 140 days, nearly a third of individuals who were discharged from hospital after acute covid-19 were readmitted (14 060 of 47 780) and more than 1 in 10 (5875) died after discharge, with these events occurring at rates four and eight times greater, respectively, than in the matched control group. Rates of respiratory disease (P

Worth mentioning that the control group wasn’t matched for hospital admission status, while the study group was solely people who ended up in the hospital. I’m guessing that the authors likely could have acquired the data to match a control group of people who had been admitted to the hospital, so I’d be curious to see what the results would’ve been there. According to the paper, this control group was picked because…

Yes, but that seems like a very relevant type of control group. If a person got COVID and ended up in the hospital, and that person wants to know what possible long-term harm that could cause them (that is, compared to had they not gotten COVID and not been hospitalized), then the article focused on exactly that.

Re: Almost third of UK Covid hospital patients readmitted within four months

#12

Post by peytn »

azakai wrote: post #9
bedhead wrote: post #8

There is a ton of selection bias here, these are already generally speaking very unhealthy people.

That's why the study looked at a control group, to try to account for that bias (as the abstract says, the control group was chosen to match "personal and clinical characteristics"). It's always possible they missed a factor, that's always a risk in such research, but they worked hard to address that possibility.

Not matching for hospital admission status is a pretty big factor given the bold, highly publicized top-line numbers:

> We selected controls from the general population rather than matching to non-covid hospital admissions to determine the increased risk after hospital admission for covid-19 versus no hospital admission for covid-19 (that is, compared with the expected risk for people with similar personal and clinical characteristics in the general population).

Re: Almost third of UK Covid hospital patients readmitted within four months

#13

Post by azakai »

DanBC wrote: post #7
roenxi wrote: post #2

The article doesn't give us quite enough information - what percentage of hospitalised Covid patients are over 70? 70 year olds are getting to the point where they spend a lot of time in hospital. The "hospitalised for Covid" part of the sample group is introducing a lot of bias. Fit, young and healthy people are much less likely to end up in hospital. So it makes sense from the first that the group is suffering more…

The ICNARC reports give more information. It's only for ICU. https://www.icnarc.org/Our-Audit/Audits/Cmp/Reports Page 29 has a chart for spread of age. Most people (88%) live without any assistance in day to day activity.

The full article says it is not only for the ICU, and most were not in the ICU:

> 47,780 patients with covid-19 (4,745 admitted to the intensive care unit and 43,035 not requiring admission to the intensive care unit) were included in the analysis

https://www.bmj.com/content/372/bmj.n693

Re: Almost third of UK Covid hospital patients readmitted within four months

#14

Post by azakai »

peytn wrote: post #12
azakai wrote: post #9

Earlier quoted context omitted.

That's why the study looked at a control group, to try to account for that bias (as the abstract says, the control group was chosen to match "personal and clinical characteristics"). It's always possible they missed a factor, that's always a risk in such research, but they worked hard to address that possibility.

Not matching for hospital admission status is a pretty big factor given the bold, highly publicized top-line numbers: > We selected controls from the general population rather than matching to non-covid hospital admissions to determine the increased risk after hospital admission for covid-19 versus no hospital admission for covid-19 (that is, compared with the expected risk for people with similar personal and clinic…

(I disagree, see discussion above: https://news.ycombinator.com/item?id=26672461 )

Re: Almost third of UK Covid hospital patients readmitted within four months

#15

Post by makomk »

azakai wrote: post #11
peytn wrote: post #6

Earlier quoted context omitted.

Worth mentioning that the control group wasn’t matched for hospital admission status, while the study group was solely people who ended up in the hospital. I’m guessing that the authors likely could have acquired the data to match a control group of people who had been admitted to the hospital, so I’d be curious to see what the results would’ve been there. According to the paper, this control group was picked because…

Yes, but that seems like a very relevant type of control group. If a person got COVID and ended up in the hospital, and that person wants to know what possible long-term harm that could cause them (that is, compared to had they not gotten COVID and not been hospitalized), then the article focused on exactly that.

Not really. The UK has been testing pretty much everyone who's admitted to hospital for COVID since the start of the outbreak regardless of why they were admitted, which means there's going to be quite a few people in the COVID and in hospital group who were actually admitted due to other, unrelated conditions. Particularly if they were being admitted to hospital regularly...

Re: Almost third of UK Covid hospital patients readmitted within four months

#16

Post by sokoloff »

azakai wrote: post #11
peytn wrote: post #6

Earlier quoted context omitted.

Worth mentioning that the control group wasn’t matched for hospital admission status, while the study group was solely people who ended up in the hospital. I’m guessing that the authors likely could have acquired the data to match a control group of people who had been admitted to the hospital, so I’d be curious to see what the results would’ve been there. According to the paper, this control group was picked because…

Yes, but that seems like a very relevant type of control group. If a person got COVID and ended up in the hospital, and that person wants to know what possible long-term harm that could cause them (that is, compared to had they not gotten COVID and not been hospitalized), then the article focused on exactly that.

That’s what they attempt to focus on, but inadvertently introduce bias in terms of people’s proclivity to show up at the hospital for any given severity of issue.

People who present once for one thing are probably more likely to present a second time: either because they have a lower threshold for seeking help, they have overall poorer health and so present more often, or because they have a referring primary doctor who is biased towards admitting over treating outpatient.

Re: Almost third of UK Covid hospital patients readmitted within four months

#17

Post by surfpel »

peytn wrote: post #6
morsch wrote: post #4

Earlier quoted context omitted.

Four times more often than the control group. Over a mean follow-up of 140 days, nearly a third of individuals who were discharged from hospital after acute covid-19 were readmitted (14 060 of 47 780) and more than 1 in 10 (5875) died after discharge, with these events occurring at rates four and eight times greater, respectively, than in the matched control group. Rates of respiratory disease (P

Worth mentioning that the control group wasn’t matched for hospital admission status, while the study group was solely people who ended up in the hospital. I’m guessing that the authors likely could have acquired the data to match a control group of people who had been admitted to the hospital, so I’d be curious to see what the results would’ve been there. According to the paper, this control group was picked because…

The most useful result here is to isolate the impact of covid hospital admission status on the probability of mortality, not the difference between covid and other hospitalizations.

Re: Almost third of UK Covid hospital patients readmitted within four months

#18

Post by azakai »

makomk wrote: post #15
azakai wrote: post #11

Earlier quoted context omitted.

Yes, but that seems like a very relevant type of control group. If a person got COVID and ended up in the hospital, and that person wants to know what possible long-term harm that could cause them (that is, compared to had they not gotten COVID and not been hospitalized), then the article focused on exactly that.

Not really. The UK has been testing pretty much everyone who's admitted to hospital for COVID since the start of the outbreak regardless of why they were admitted, which means there's going to be quite a few people in the COVID and in hospital group who were actually admitted due to other, unrelated conditions. Particularly if they were being admitted to hospital regularly...

Good question to ask, but no, they didn't look at people who tested positive but were admitted for something else:

> Individuals were included if they had a hospital episode from 1 January to 31 August 2020 with a primary diagnosis of covid-19

(from the full paper)

Re: Almost third of UK Covid hospital patients readmitted within four months

#19

Post by azakai »

sokoloff wrote: post #16
azakai wrote: post #11

Earlier quoted context omitted.

Yes, but that seems like a very relevant type of control group. If a person got COVID and ended up in the hospital, and that person wants to know what possible long-term harm that could cause them (that is, compared to had they not gotten COVID and not been hospitalized), then the article focused on exactly that.

That’s what they attempt to focus on, but inadvertently introduce bias in terms of people’s proclivity to show up at the hospital for any given severity of issue. People who present once for one thing are probably more likely to present a second time: either because they have a lower threshold for seeking help, they have overall poorer health and so present more often, or because they have a referring primary doctor…

That's a fair point. That could inflate the later admission numbers.

However, I don't think it can explain the far higher death rate.

Re: Almost third of UK Covid hospital patients readmitted within four months

#20

Post by peytn »

surfpel wrote: post #17
peytn wrote: post #6

Earlier quoted context omitted.

Worth mentioning that the control group wasn’t matched for hospital admission status, while the study group was solely people who ended up in the hospital. I’m guessing that the authors likely could have acquired the data to match a control group of people who had been admitted to the hospital, so I’d be curious to see what the results would’ve been there. According to the paper, this control group was picked because…

The most useful result here is to isolate the impact of covid hospital admission status on the probability of mortality, not the difference between covid and other hospitalizations.

Anyone who’s even dipped a toe in this policy area can tell you that when old people and hospitals meet, gaudy statistics follow. It’s common knowledge in the field. To not include that context or any points of comparison seems irresponsible to me given the publicity.
Post reply on HN