Re:Elderly patients 23% more likely to die if surgery is on ...
by esquire_900 on Wednesday December 16, 2020 @04:51AM
(#25440726)
You make a good point, though I'd like to add that one would expect 1/365 = .274% of procedures to be on a random day (like a birthday). .21% is not so far off that it can't be random chance.
[ Parent
| on HN ]
Re:Elderly patients 23% more likely to die if surgery is on ...
by eCa on Wednesday December 16, 2020 @05:26AM
(#25440962)
That, and since most planned surgeries ought to take place Mon-Fri, if anything it’s a higher percentage than I would have expected.
Re:Elderly patients 23% more likely to die if surgery is on ...
by Dylan16807 on Wednesday December 16, 2020 @05:28AM
(#25440981)
Day of the week cancels out.
Re:Elderly patients 23% more likely to die if surgery is on ...
by ChrisLomont on Wednesday December 16, 2020 @07:33AM
(#25441796)
> .21% is not so far off that it can't be random chance
How far off exactly would it need to be before it can be random chance?
Re:Elderly patients 23% more likely to die if surgery is on ...
by esquire_900 on Thursday December 17, 2020 @03:15AM
(#25453670)
Interesting question, wondered that myself as well. My statistics are a little rusty, so doing a quick simulation shows that the range is about .25-.29 (code sample 1). However, things like operations aren't completely random, and you might be more correct by looking at something like a Poisson process, resulting in a range of about .14-.44 for this sample size (code sample 2).
All in the end to be taken with a grain of salt of course, as the planning component of operations eliminates most statistical properties :)
``` operations = np.random.randint(0, 365, 980000) pd.Series(operations).value_counts().sort_values() / 980000 ```
``` x = np.random.poisson(.0089, 1000000) x = (pd.Series(x).cumsum() / 24).round() x.groupby(x).count().iloc[:365].sort_values() / 980000 ```