OHDSI Home | Forums | Wiki | Github

Qualifier for Measurements

(Philip Solovyev) #1

Dear community,

Working on the conversion of the clinical trial data to the CDM, it turned out that in some cases, the CDM measurement table lacks a qualifier field. Here is a couple of problems we have encountered with.

During a Right Heart Catheterization procedure, while a patient was on supplemental oxygen, arterial O2 saturation had been measured. To keep the fact that supplemental oxygen took place, we used an observation 40483579 (Arterial oxygen saturation) with a qualifier 45884194 (Receiving supplemental oxygen.) At the same time in the Measurement domain, there is a concept 4013965 (Oxygen saturation measurement, arterial) that is similar by sense and seems to be more appropriate.

Another example is mean pulmonary artery pressure that had been measured during exercise. It had been incorporated into CDM also as an observation 4353857 (Pulmonary artery mean pressure) with a qualifier 21499046 (Exercise). Meanwhile, it would have been stored as a measurement 3028074 (Pulmonary artery Mean blood pressure) if there had been a qualifier field in the Measurement table.

So I guess many ETLers would like to have a qualifier for measurements as it is in observations. As far as I know, in the Oncology extension the measurement table has been extended by two extra fields – modifier_of_event_id and modifier_of_field_concept_id. I guess they together can be used as a qualifier, so the other possible way would be adopting these fields by the regular CDM with changing naming to somewhat more general.

(Vojtech Huser) #2

I remember when I was advocating for the modifier in observation. This is another case where post-coordination grammar for concept field, postcoordination via modifier and precoordination di-/ trilemma appears.

(Christian Reich) #3

Why not pre-coordination? Sounds like those oxygen and pulmary blood pressure combos with the attributes are commonly used.