OHDSI Home | Forums | Wiki | Github

Concept Type consolidation - please take a look

CDM Builders
#58

Thanks, @Christian_Reich!

#59

Is there a lookup from the ‘old’ type concepts to the new consolidated type concepts?

#60

I was looking for the same thing. It seems like the old type concepts. For example, ‘Radiology Report’ (44814641) does not map to the new standard code. @Christian_Reich, will this mapping be updated in the vocabulary or will these older codes become invalid?

#61

We can add. We have them internally. Didn’t think it would help, but hey, seems it does. Next release.

#62

Hi,

Is a measurement recorded by the monitor or the ventilator defined by type_concept EHR 32817 ? Or is there another which references measurement collected by a machine ?

#63

It might be good to add non-standard to standard mapping relationships from the spreadsheet to OHDSI vocabulary.

#64

@alexander do you mean these?

Looks like a really good idea.

Have you guys used this table and validate the mapping proposed by @DTorok ?

#65

We had those started, but then didn’t bring them in. Reasons: (i) lack of use case (nobody has Type Concepts as source data), (ii) ambiguity in the maps and (iii) lack of a Wide Mapping table. Some of them have to be split into Status concepts.

#66

We used them in our ETL and after updating version of dictionaries we had to re-map few concepts manually as they are not standard any more.

#67

I can’t say it was a big deal for us, but it would be better to have non-standard to standard mapping after the consolidation of some concepts in the vocabs to make the update easier. It looks a little weird when type concepts are not standard anymore and are not mapped to standard ones.

#68

I agree. This has been an issue for some our sites. If the “maps to” could be added to these deprecated concepts, it would make life easier for many ETLs.

#69

We did it intentionally in order to avoid ambiguity in mapping of those that can’t get 1-to-1 links.

#70

@Christian_Reich - do you have type concept id that I can use for app generated data or manually entered data from a paper form?

#71

What’s “app generated”? Some patient typed it into an app? Use Patient self-report. Or some application that calculated the data item? Use Standard algorithm.

What’s a paper form? A Case Report Form?

#72

@Christian_Reich - could you create a type concept for app generated data and manually entered data from a paper form?

Measurement Type Concept id for
#73

@stephanieshong:

You don’t like the suggestions I made?

#74

let me provide you with bit more context -
use case 1) Montreal Cognitive Assessment tests are completed on the app but the assessment score measurement by the app is then entered into the survey form by the Procter.
use case 2) When it is not possible to enter the assessment score via online survey form, the paper form is submitted and it is then manually entered at a later time.

#75

let me know if the additional explaination is helpful to you to determine the correct type concept id

#76

@stephanieshong:

I am still confused what you need. So, essentially, you are saying there is an app who determines a score, and that score is then entered into the survey by some “proctor”. Correct? Sometimes the survey form is online, sometimes is is paper. Did I get that right?

If so, why is that relevant? Who cares whether it is an electronic or paper form? Is there a scientific question that believes it makes a difference what capture method is used for data entry?

#77

This seems like it may be a larger gap in the concept types for assessment scores.

These are not: Surveys, Patient Self-Report, Health Risk Assessment, or Standard Algorithm

Observation from Measurement being mapped to Lab reduced this coverage and the former would have been the appropriate concept type for all assessments like:

  • Short Physical Performance Battery (SPPB)
  • Berg Balance Scale (BBS)
  • Montreal Cognitive Assessment (MoCA)
  • Mini-Mental State Examination (MMSE)

An initial proposal: Create the concept Assay which subsumes types distinguished by both setting and system they are collected. For example: clinical assay, laboratory assay, imaging assay. All assays would have have the Vocabulary IDs: Procedure, Measurement, and Observation.

Examples:

  • This record was obtained from an laboratory assay (Setting: lab, system: LIS->EHR)
  • This record was obtained from an clinical assay (Setting: clinic, system: Assessment Tool->EHR)
  • This record was obtained from an imaging assay (Setting: imaging, system: PACS->EHR)

However, I don’t think we need to separate this into how an assessment is administered. The score ultimately ends up in the EHR, and the use-case of validity studies seems too granular for what concept types are for. Plus, the validity would hopefully already be established by studies prior to different formats being used in clinical settings.

t