Finding lab data for MIMIC-IV-ED patients? #1558
Unanswered
tchang1997
asked this question in
MIMIC-IV-ED
Replies: 1 comment
-
I agree <1% feels too low. Probably just a slight mismatch in timing. I would add a fuzzy window of +- 8 hours to the intime/outtime for ED stays and see if you still have a very low number of labs. I would be interested to see the percentages if you post them here. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
I am trying to work with lab data for patients in MIMIC-IV-ED. So far, I've been using the
labevents
table in MIMIC-IV. However, I've run into difficulties finding data for patients with a nullhadm_id
. I've tried to link such patients with thelabevents
table without much success — I describe my strategy in more depth here.(1) For patients with a non-null hadm_id, obtaining their
labevents
info is straightforward sincelabevents
stores thehadm_id
.(2) I mostly have problems with patients with a null hadm_id. I tried to link
labevents
to ED stays as follows:labevent
, retrieve all ED stays with a matchingsubject_id
charttime
falls between theintime
andouttime
attributesIn other words, I'm trying to find all lab tests done for a particular patient at any time, and filling in the
stay_id
for each lab event based on whether thecharttime
falls within any of that patient's ED stays.Based on this strategy, however, <1% of all patients with a null hadm_id are tested, which seems unreasonably low, even though we're looking at non-admitted patients. I get higher testing rates when linking to a test in
microbiologyevents
instead oflabevents,
but they still seem low. Is lab data for non-admitted patients in MIMIC-IV-ED available anywhere else in MIMIC? Or is there a more reliable strategy for linking labevents to ED stays?Beta Was this translation helpful? Give feedback.
All reactions