Stream: implementers
Topic: RiskAssessment stratification
Michelle (Moseman) Miller (Nov 08 2016 at 20:59):
@Lloyd McKenzie I ran across GF#8404 that you had logged asking to move the CodeableConcept from RiskAssessment.prediction.probability to be a separate element. (If you hadn’t logged it already, I would have requested the same change!)
Question, though, do you think the probabilityCodeableConcept (i.e. proposed qualitativeRisk) can be used for any type of stratification? For example, using an observation analogy, some risk scores are deemed to be high, medium, low based on simple reference range like thresholds that are independent of the population. By contrast, in other scenarios, we might want to stratify the population by quantiles (e.g. quartiles, deciles, etc). Do you think both of these ‘fit’ in the same code (whatever we call it going forward) or are quantiles a different element since they are relative to the population?
Lloyd McKenzie (Nov 08 2016 at 22:23):
I'm comfortable with probabilityCodeableConcept for that. Basing the notion of "high" on "significantly higher than the base population" seems reasonable to me. The real challenge is whether you could treat "high" and "95th percentile" as translations - that might lean towards separate attributes.
Last updated: Apr 12 2022 at 19:14 UTC
