Releases: smart-on-fhir/cumulus-library-data-metrics
Releases · smart-on-fhir/cumulus-library-data-metrics
8.0.0
Whenever data-metrics stratifies Encounters by category, it now stratifies by the class field (e.g. AMB or EMER) instead of the type field (e.g. a SNOMED code). This provides a more reasonable number of categories for charting purposes.
This does mean that a column name in c_resource_count changed from type to class, so the major version and data package version are getting bumped.
7.0.2
7.0.1
7.0.0
This release adds some new metrics and features:
- New attachment-oriented metrics, to help evaluate NLP-readiness of your document references and diagnostic reports.
- Relatedly, now that Cumulus ETL provides a marker when it strips attachment info, the US Core metrics added checks for
DiagnosticReport.presentedFormandDocumentReference.context.attachment, where appropriate.- For performance reasons, the addition of this extra validation field for DocumentReferences meant that we now split the
c_us_core_v4_countDocumentReference tables into two for performance reasons.
- For performance reasons, the addition of this extra validation field for DocumentReferences meant that we now split the
- Added a
medianstatistic column toc_resources_per_pt - All cube counts generated now have low-patient-count combinations (rows with less than ten patients) cut out, to protect privacy and make it harder to de-identify the patients.
- Fixed a bug when running data-metrics against a freshly created and empty Athena database.
6.0.0
5.1.0
5.0.1
When stratifying by category, data-metrics now follows US Core more closely by adding additional Condition category codes (health-concern and 16100001 - death diagnosis from SNOMED) and restricting DiagnosticReport category codes to just the ones mentioned by US Core (LAB and three specific LOINC note codes).