Probability of various outcomes once infected

[Sojourner]

Has anyone come across recent, high quality data on the likelihood of outcomes (e.g., hospitalization, ICU, death) for people in specific, narrow demographic groups? For example, I'd be very interested to know what percentage of men in their 50s—with no comorbidities, who don't smoke and aren't obese—end up having to be hospitalized after becoming infected with COVID-19.

I have seen numerous charts and graphs over the weeks and months that come close to what I'm looking for (e.g., deaths by age group, deaths by comorbidities, deaths by race, etc.), but nothing that shows the numbers when all the various risk factors are combined.

 

[rodi]

I haven't seen anything like that. Just the ones you mentioned (deaths by race, age, gender).

I would imagine any data would vary from state to state...

 

[rodi]

I found this - it looks specifically at covid patients who were hospitalized and their outcomes.

https://nyulangone.org/news/health-system-pandemic-epicenter-identifies-outcomes-new-risk-factors-patients-hospitalized-covid-19

The study it refers to is here:
https://www.bmj.com/content/369/bmj.m1966

 

[audreyh1]

That’s a useful link. I found this important:

The investigators found that while age and comorbidities were strong predictors of hospitalization, critical illness, and death, patients with low levels of blood oxygen (88 percent or lower) upon admission and those with markers of inflammation such as elevations in c-reactive protein, which is a substance produced by the liver in response to inflammation, were most likely to have the worst outcomes during hospitalization.

So low oxygen levels or high inflammation markers. IMO you can’t know in advance whether that will happen to you or overall how your immune system will react or overreact. That makes it seem so much more like Russian roulette.

Also, I suspect that the exposure dose matters a great deal to how sick someone gets. Again, we’re talking about chance rather than purely health here. Risky behavior means higher chances of large dose exposure.

 

[CardsFan]

I am sure this varies widely, but in my county we have just over 1000 confirmed cases, and 73 deaths. County population is just over 400,000. I can't find any data on co-morbidity for the county, but found a startling statistic. 60 of the deaths were people in Long Term Care facilities. That's 82%! And, 72% of all cases were in LTC facilities.

 

[harllee]

But remember it is not an all or nothing outcome--it is not that you get 100% well or that you die. Some people who get Covid-19 do not die but are being left with long term serious side effects. As far as I can tell there are not any real statistics on this.

 

[CDRE]

This new calculator created by U.K. researchers takes your health stats into account and paints a clearer picture of your risk of complications.
This online calculator was created by researchers from University College London (UCL), UCL Hospitals NHS Trust, the University of Cambridge and Health Data Research U.K. It draws from a data pool of more than 3.8 million health records from patients in England — making its calculations more accurate.
OurRisk.CoV | COVID-19 Phenomics

 

[Sojourner]

This new calculator created by U.K. researchers takes your health stats into account and paints a clearer picture of your risk of complications.

Hmmm, interesting. Thanks for sharing this link.

Selecting age group 51-55, male, and no underlying conditions (i.e., no comorbidities), I got the following results:

We estimate there are 1,324,122 people in England with similar characteristics.

Of these, we would expect approximately 4,888 to die each year, before COVID-19 ever existed.

With the selected assumptions regarding the COVID-19 emergency we might expect there to be an additional 244 deaths amongst this group of people.

So, if I'm interpreting this correctly, the odds of someone in this narrow demographic dying in a given year, prior to COVID-19, were about 4,888/1,324,000. Equivalent to 0.369%, or one in 271. Adding deaths due to COVID-19, that goes up to about 0.388%, or one in 258. Still very low odds in an absolute sense, but a nontrivial relative increase in the risk of dying.

 

[Sojourner]

I found this - it looks specifically at covid patients who were hospitalized and their outcomes.

https://nyulangone.org/news/health-system-pandemic-epicenter-identifies-outcomes-new-risk-factors-patients-hospitalized-covid-19

Very interesting article. Easy, quick read, with lots of hard data. I found this quote near the end of the article particularly sobering.

"If you are young, that doesn’t mean that you can’t get the virus. Almost a third of our hospitalized patients are younger than 55. Similarly, having chronic diseases increases risk, but one in five of our hospitalized patients had no chronic conditions."

 

[CardsFan]

Hmmm, interesting. Thanks for sharing this link.

Selecting age group 51-55, male, and no underlying conditions (i.e., no comorbidities), I got the following results:



So, if I'm interpreting this correctly, the odds of someone in this narrow demographic dying in a given year, prior to COVID-19, were about 4,888/1,324,000. Equivalent to 0.369%, or one in 271. Adding deaths due to COVID-19, that goes up to about 0.388%, or one in 258. Still very low odds in an absolute sense, but a nontrivial relative increase in the risk of dying.

I have no way of knowing if your numbers are accurate, but if they are I would say it IS a trivial increase. We are talking a .02% change. My guess is the standard deviation for the no-covid number is greater than that.

 

[eytonxav]

high quality data

That sounds like an oxymoron.

 

[JoeWras]

Very interesting article. Easy, quick read, with lots of hard data. I found this quote near the end of the article particularly sobering.

I think this data is very stale. The nearly 50% hospital admission was taken from March 1 to April 8. This was a time of limited tests. Recall the authorities were telling people to self isolate with symptoms and not even look to get a test unless they got severe, because testing availability was so limited.

So the study is nice, but has bad input data simply because the input data set was limited to severe cases only.

---

I'd like the same info the OP is looking for. Unfortunately, we tend to get filtered hype (click bait) from our media instead.

 

[FIREd_2015]

Has anyone come across recent, high quality data on the likelihood of outcomes (e.g., hospitalization, ICU, death) for people in specific, narrow demographic groups? For example, I'd be very interested to know what percentage of men in their 50s—with no comorbidities, who don't smoke and aren't obese—end up having to be hospitalized after becoming infected with COVID-19.

I have seen numerous charts and graphs over the weeks and months that come close to what I'm looking for (e.g., deaths by age group, deaths by comorbidities, deaths by race, etc.), but nothing that shows the numbers when all the various risk factors are combined.

From Geneva, Switzerland infection fatality risk study "...Of the 286 reported deaths caused by SARS-CoV-2, the youngest person to die was 31 years old. Infected individuals younger than 50 years experienced statistically similar IFRs (range 0.00032-0.0016%), which increases to 0.14% (95% CrI 0.096-0.19) for those 50-64years old to 5.6% (95% CrI 4.3-7.4) for those 65 years and older (supplement). After accounting for demography and age-specific seroprevalence, we estimate a population-wide IFR of 0.64% (95% CrI 0.38-0.98)..."

https://www.medrxiv.org/content/10.1101/2020.06.10.20127423v1.full.pdf