MeasureGroupComponent Codes Don't Reflect Multiple Populations

[jeffeastman]

The odd numerator and denominator counts in my tests are due to a defect in the FhirMeasureEvaluator that does not properly code the MeasureReportGroupComponents to handle multiple numerators, denominators, etc. For example, DMS measure has 3 numerators and 3 denominators that aren't being coded correctly. Here's a snippet from a successful MeasureReport:

  "type": "individual",
  "measure": {
    "reference": "Measure/measure-dms"
  },
  "patient": {
    "reference": "Patient/Patient-2517"
  },
  "period": {
    "start": "2017-01-01T00:00:00-05:00",
    "end": "2017-12-31T00:00:00-05:00"
  },
  "group": [
    {
      "population": [
        {
          "code": {
            "coding": [
              {
                "code": "initial-population"
              }
            ]
          },
          "count": 1
        },
        {
          "code": {
            "coding": [
              {
                "code": "numerator"
              }
            ]
          },
          "count": 1
        },
        {
          "code": {
            "coding": [
              {
                "code": "denominator"
              }
            ]
          },
          "count": 1
        },
        {
          "code": {
            "coding": [
              {
                "code": "numerator"
              }
            ]
          },
          "count": 1
        },
        {
          "code": {
            "coding": [
              {
                "code": "denominator"
              }
            ]
          },
          "count": 1
        },
        {
          "code": {
            "coding": [
              {
                "code": "numerator"
              }
            ]
          },
          "count": 1
        },
        {
          "code": {
            "coding": [
              {
                "code": "denominator"
              }
            ]
          },
          "count": 1
        }
      ]
    }
  ],
  "evaluatedResources": {
    "reference": "#cd3c4385-e2bb-4218-89c0-d044c99c68e9"
  }
}
Aaron Kelly Peters [75] is in initial-population
Aaron Kelly Peters [75] is in numerator
Aaron Kelly Peters [75] is in denominator
Aaron Kelly Peters [75] is in numerator
Aaron Kelly Peters [75] is in denominator
Aaron Kelly Peters [75] is in numerator
Aaron Kelly Peters [75] is in denominator

[jeffeastman]

Adding the population's identifier to the populationReport in FhirMeasureEvaluator.resolveGroupings gives the client more information on the details of the classification.

				MeasureReport.MeasureReportGroupPopulationComponent populationReport = new MeasureReport.MeasureReportGroupPopulationComponent();
				populationReport.setCount(count);
				populationReport.setCode(population.getCode());
				populationReport.setIdentifier(population.getIdentifier());

and yields MeasureReports that can be classified like this:

http://localhost:8080/cqf-ruler/baseDstu3/Patient/Patient-1121/_history/1
	patient is in initial-population
	patient is in numerator 1
	patient is in denominator 1
	patient is in numerator 2
	patient is in denominator 2
http://localhost:8080/cqf-ruler/baseDstu3/Patient/Patient-1128/_history/1
	patient is in initial-population
	patient is in denominator 1
http://localhost:8080/cqf-ruler/baseDstu3/Patient/Patient-1122/_history/1
	patient is in initial-population
	patient is in denominator 1

[jeffeastman]

And lets me produce Cat-3 statistics for the population (albeit slowly);

	denominator 1 = 81
	denominator 2 = 36
	initial-population = 81
	numerator 1 = 36
	numerator 2 = 6

I forked a branch with this fix but evidently don't have karma to push it for a pull request.

[jeffeastman]

I haven't figured out where, but the identifier seems to already be set. So this is just pilot error on my part. Closing...

[jeffeastman]

Reopening because, for some reason I don't yet understand (maybe in the hot-code replacement during debugging?), the identifiers are no longer being set. Checking the codes still yields codes that don't distinguish between numbered numerators, denominators etc.

[jeffeastman]

With more digging, I've located why I thought the identifier was already being set and it did have to do with hot code replacement while I was debugging, so the patch really needs to be made.

[jeffeastman]

I don't see the above patch in the latest checkout. Is there a better way included or is there a reason why it is not?

[jeffeastman]

Closing, thanks