OHDSI Home | Forums | Wiki | Github

Hospital readmissions

Researchers
#1

Hi all,

What are your favorite papers on all-cause 30-day hospital readmissions in US population from administrative data?

(Yes, George, I’m aware that you think the focus on this problem is misguided :)).

Cheers,
David

#2

Ah, clearly the answer must be my paper on it [Colin Walsh, Hripcsak. “The Effects of Data Sources, Cohort Selection, and Outcome Definition on a Predictive Model of Risk of Thirty-Day Hospital Readmissions.” JBI 2014.], and my guilt for misguiding Colin. George

1 Like
#3

@Dsontag, I’m not sure I’d call it a ‘favorite’, but I came across this
recent publication today and thought I’d share it with you:
https://www.ncbi.nlm.nih.gov/pubmed/27784047. The one thing I did like in
this paper was an attempt at comparing different machine learning methods
for the same task, and they did look at both calibration and discrimination
for evaluation metrics. I think the work could have been helped a bit by
following the full framework that @jennareps and @Rijnbeek have been
leading us to in the OHDSI predictive modeling workgroup:)

#4

Thanks, Pat and George!

#5

Hi David,

I actually like this HCUP Methods report: Overview of Key Readmission Measures and Methods which compares several different methods of computing 30 day readmission rates. E.g. how are qualifying stays defined, numerator / denominator etc.

#6

Thanks Akshay, this is very helpful!

t