search

YerevaNN / mimic3-benchmarks

Python suite to construct benchmark machine learning datasets from the MIMIC-III 💊 clinical database.
https://arxiv.org/abs/1703.07771
MIT License
806 stars 329 forks source link

Is there a mapping between hadm_id and episode? #97

Open johntiger1 opened 4 years ago

johntiger1 commented 4 years ago

From my understanding, there is a 1:1 mapping between hadm_id and an episode for a subject. In other words, a patient episode simply summarizes a hospital admit which has a key hadm_id. Is there a mapping file? This would let us join back to get other data from MIMIC. Thanks

hrayrhar commented 4 years ago

Hi @johntiger1 ,

I don't think we have the mapping ready somewhere, but it will be possible to construct it by going over all episodes after processing the root folder.

Hrayr

shan23chen commented 3 years ago

From my understanding, there is a 1:1 mapping between hadm_id and an episode for a subject. In other words, a patient episode simply summarizes a hospital admit which has a key hadm_id. Is there a mapping file? This would let us join back to get other data from MIMIC. Thanks

Hi, I am facing the same question, did you figure out the mapping by chance? Thanks!

shan23chen commented 3 years ago

Hi @johntiger1 ,

I don't think we have the mapping ready somewhere, but it will be possible to construct it by going over all episodes after processing the root folder.

Hrayr

I am also wondering which file you guys use to partition the csv files into all the episodes. Thanks!