Cumulus Library 1.4.0 Release #12
dogversioning
announced in
Announcements
Replies: 1 comment
-
We've released a v1.4.1 hotfix for this release cleaning up some issues with quoting in the discovery study - same install command as above |
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
-
You can install the latest library release with
pip install -U cumulus-library
. The release notes are over on the project github.There are two bits of new functionality that we'd like to go into a little bit more detail on:
FHIR Resource filtering
This update will change the way your data looks on export, so it warrants a little bit more explination.
We had a number of records that we were including that probably shouldn't have been - taking Encounters as an example, we were including things like future scheduled followup appointments, which generated a dangling tail of records in charts going (potentially) into dates that had not yet occured.
We're now using the
status
field on records like these (Encounters, Observations, and DocRefs) to limit the exported counts to looking at records that are final, and not superceded by an updated record. We still keep these records around in a database, so if you were working on a study that might have use of these temporally related records, you could use them - you'd just have to generate counts by hand, rather than using one of the provided count builder functions.New study: Discovery
We've ran into several instances of differences between EHR vendor implementation causing unexpected issues, especially when using codings to filter FHIR resources as part of cohort selection. This has resulted in us doing some live debugging sessions over slack - and while that's helpful, this study is an attempt to generate information that would allow us to detect issues like this before they happen.
Currently, this will extract codings/CodeableConcepts that are root level elements of a FHIR resource, which can be uploaded to the aggregator via the same mechanism as a count-driven study. We can then use this to inform future study design, or to help onboard a new partner into the Cumulus Network.
Beta Was this translation helpful? Give feedback.
All reactions