OHDSI Home | Forums | Wiki | Github

Storing Charlson Score and its Parts in the CDM

CDM Builders
,
#1

I have come across a data source that stores the Charlson score as well as its parts. Here is an example of some of the data values from the source:

Item              Valore
Charston_Totale    0
Charston_Totale    1
Charston_Totale    2
Charston_AIDS      0 = No
Charston_AIDS      1 = Si

Storing the total score seems simple to me. I would store the “CHARSTON_TOTALE” in MEASUREMENT, where:

  • MEASUREMENT_CONCEPT_ID = 42538860 - Charlson Comorbidity Index
  • VALUE_AS_NUMBER = “Valore”
  • MEASUREMENT_SOURCE_VALUE = Name of the column “CHARSTON_TOTALE”

However, I had to think about the pieces of the Charlson but what I settled on was this. They would go in the OBSERVATION table like this:

  • OBSERVATION_CONCEPT_ID = 4188893 - History Of
  • VALUE_AS_NUMBER = “Valore” (when numeric)
  • VALUE_AS_STRING = “Valore” (when numeric provide associated numeric description if one exists (1=si), when not numeric store the value)
  • VALUE_AS_CONCEPT_ID = the condition is indicated as present (ex AIDS = Si set to 439727 -
    Human immunodeficiency virus infection) - I would use this to help find concepts for each condition ( FeatureExtraction/CharlsonIndex.sql at master · OHDSI/FeatureExtraction · GitHub)
  • OBSERVATION_SOURCE_VALUE = Name of the column “CHARSTON_AIDS”

Putting into the universe to see if others have better ideas. :unicorn:

Tagging my partner in crime @Sebastiaan_van_Sandi

#2

@ericaVoss I like this approach. I think it makes sense to store the total score as a measurement and the parts and condition history in the observation table.

1 Like
t