Modeling for QI-Core Encounter.Procedure raises concerns about consistency across FHIR and across QI-Core.

For reference - FHIR

Encounter.diagnosis.condition QI-Core link: [Encounter.diagnosis.condition

Definition: Reason the encounter takes place, as specified using information from another resource. For admissions, this is the admission diagnosis. The indication will typically be a Condition (with other resources referenced in the evidence.detail), or a Procedure. (For systems that need to know the which was the primary diagnosis, these will be marked with the standard expression primaryDiagnosis (which is a sequence value rather than a flag, 1 = primary diagnosis).

Encounter.diagnosis.use QI-Core link: [Encounter.diagnosis.use

Defintion: Role that the diagnosis has within the encounter (e.g., admission, billing, discharge…)

Encounter.diagnosis.rank QI-Core link: [Encounter.diagnosis.rank

Definition: Ranking of the diagnosis for each role type (positiveInt)

Given this modeling approach and the fact that Encounter.diagnosis.condition also referenes procedures, should QI-Core just use the Encounter.diagnosis.condition, Encounter.diagnosis.use and Encounter.diagnosis.rank to reference BOTH principal procedure and principal diagnosis (i.e., rather than create extensions)?

Current R4 QI-Core Encounter.procedure extension includes:

QICore.encounter-procedure-definition#extension.extension:Type.value[x] - one of the codes in the procedure.type value set - whether the procedure was primary or secondary

Note inconsistent modeling in FHIR: the FHIR R4 Claim resource defines Claim.Procedure.type using a value set (example) of primary and secondary with the following definitions:

Primary - the first procedure in a series required to produce an overall patient outcome

Secondary - The second procedure in a series required to produce an overall patient outcome

Hence, is primary referencing the same concept as principal given the way the claim resource defines it. Principal diagnosis is not necessarily part of a series and it seems to have a different definition.

Also note: The Claim.sequence is defined as a number to uniquely identify diagnosis entries and Claim.procedure.sequence as a number to uniquely identify procedure entries.

Conclusion - The Encounter.diagnosis attributes are modeled differently than the same concepts for claim.diagnosis (or claim.procedure). Therefore, QI-Core needs to address one of these options until such time as FHIR provides a more consistent approach.

1. Perhaps QI-Core can merely use the Encounter.diagnosis.condition attribute to handle an encounter-related procedure _ or _ and encounter-related diagnosis, and subsequently use the same role and rank attributes without having a separate extension for procedure.

2. Alternatively, the QI-Core can use the extension to reference the encounter-related procedure. However, the example procedure-type value set seems to be inappropriate (i.e., primary and secondary). It seems more appropriate to model the procedure similar to diagnosis, with a role, rather than using type.

For reference: Although Claim.procedure and claim.diagnosis reference a sequence, it seems to reference the same thing as intended by rank, and, therefore, should probably not be used to reference a principal procedure. Therefore, QDM to QI-Core mapping should change to reference the rank rather than sequence.

Claim.procedure.type is when the condition was observed or the relative ranking

Claim.diagnosis.sequence is the number to uniquely identify diagnosis entries

Claim.diagnosis.type is when the condition was observed or the relative ranking