CDM

From Clinfowiki
Revision as of 12:55, 26 September 2014 by Vojtech huser (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Common Data Model is a model developed by OMOP. OMOP has transitioned to IMEDS. Another related initiative is OHDSI.

In 2014 - CDM v5 draft specification were released.

CDM v4

  • Added death table
  • schema [link]

Procedure occurrence

Every procedure done to the patient is stored in this table. Key column is procedure_concept_id. Dates are only stored on day granularity. (time is missing)

CDM v5

  • changes to CDMV

Procedures

Suggestion

Vojtech Huser I am working with the CCAE data in CDM4 and based on this I would like to submit for consideration a 2 new columns for the CDM5 procedure occurance table.

The new column names are:

relevant_condition_concept_id relevant_condition_source_value

In fact, IMEDS CCAE CDM4 dataset has "relevant_condition_concept_id" despite this being not in official CDM4 specs.


If a provider orders a procedure, claims data often contain "the reason" for the procedure. E.g., CTP code 81241 for testing "coagulation Factor V test" may include as reasons for testing diagnostic codes, such as CDM VocabID:432296 Coagulation factor deficiency .

However, the source value is not saved and (guess what), the most frequent reason for testing (after NULL) is CDMV id:0 "No matching concept")

If we try to be consistent and allow saving source value code where we try to detect CDMV concept, I think reason for procedure is a good source value to preserve. All claims data have this info.

CDM Vocabulary (CDMV)

CDMV supports relationships between concepts. Vocabulary ID is an important variable. SNOMED is 1, CPT is 4. In version 5, this may get a bit easier (ultra-short strings will be used)