FHIR Chat · Mapping multiple Lab Results · implementers

Stream: implementers

Topic: Mapping multiple Lab Results


view this post on Zulip Maria Hu (Feb 22 2019 at 23:31):

For a lab test with multiple test results (e.g. a Chem 4 test which has multiple results of Sodium, Potassium, Chloride, Bicarbonate, Anion Gap), how can these multiple lab results be mapped to resource Observation when the cardinality of observation.value is only (0..1)? Do I've to send Observation multiple times? Please advise, thanks.

view this post on Zulip Michele Mottini (Feb 22 2019 at 23:36):

A DiagnosticReport with one Observation per value - https://www.hl7.org/fhir/diagnosticreport-example.html

view this post on Zulip Michele Mottini (Feb 22 2019 at 23:37):

Or: one Observation with multiple components (https://www.hl7.org/fhir/observation-definitions.html#Observation.component)

view this post on Zulip Michele Mottini (Feb 22 2019 at 23:38):

Or: a main Observation referencing child ones - https://www.hl7.org/fhir/observation-definitions.html#Observation.hasMember

view this post on Zulip Lloyd McKenzie (Feb 22 2019 at 23:42):

Components wouldn't be appropriate because each of these are separately interpretable. DiagnosticReport or Observation with members would be the way of grouping them.

view this post on Zulip Maria Hu (Feb 22 2019 at 23:43):

Components wouldn't be appropriate because each of these are separately interpretable. DiagnosticReport or Observation with members would be the way of grouping them.

What do you mean by Observation with members? Please elaborate, thanks.

view this post on Zulip Michele Mottini (Feb 22 2019 at 23:44):

The last option I listed

view this post on Zulip Michele Mottini (Feb 22 2019 at 23:45):

And in the US at least the agreed way is the first: use DiagnosticReport

view this post on Zulip Maria Hu (Feb 22 2019 at 23:47):

And in the US at least the agreed way is the first: use DiagnosticReport

DiagnosticReport and reference Observation multiple times with a cardinality of (0..*)? It still involves referencing resource Observation multiple times?

view this post on Zulip Michele Mottini (Feb 22 2019 at 23:48):

Yes - there will be one Observation for each of the lab values, and one DiagnosticReport for the overall lab test referencing all those observations

view this post on Zulip Maria Hu (Feb 22 2019 at 23:51):

Yes - there will be one Observation for each of the lab values, and one DiagnosticReport for the overall lab test referencing all those observations

mmmmm thanks Michele

view this post on Zulip Lloyd McKenzie (Feb 22 2019 at 23:51):

Every test needs to be captured as an independent Observation because they might be passed around or queried independently. The only time you group things is if they're traditionally always interpreted together (e.g. systolic & diastolic for blood pressure) or the drug and the susceptability for a a susceptability test

view this post on Zulip Maria Hu (Feb 22 2019 at 23:53):

Every test needs to be captured as an independent Observation because they might be passed around or queried independently. The only time you group things is if they're traditionally always interpreted together (e.g. systolic & diastolic for blood pressure) or the drug and the susceptability for a a susceptability test

Well, you can't really look at the Sodium result only without looking at Potassium/Chloride/Bicarbonate/Anion Gap together, that's the purpose of ordering a Chem 4 test instead of 5 different standalone tests.

view this post on Zulip Lloyd McKenzie (Feb 22 2019 at 23:55):

It'll be reported together. But once reported, you could easily query to just see all of the sodiums done for the patient so you can analyze a trend and totally ignore the others.

view this post on Zulip Maria Hu (Feb 23 2019 at 00:01):

It'll be reported together. But once reported, you could easily query to just see all of the sodiums done for the patient so you can analyze a trend and totally ignore the others.

ok thanks, I still think the cardinality of observation.value should relax to be (0..*)

view this post on Zulip Lloyd McKenzie (Feb 23 2019 at 00:02):

Each of the observations you described would also have distinct codes, distinct normal ranges, distinct interpretations, etc.

view this post on Zulip Lloyd McKenzie (Feb 23 2019 at 00:02):

Making Observation.value repeat wouldn't accomplish your objective.

view this post on Zulip Maria Hu (Feb 23 2019 at 00:11):

Each of the observations you described would also have distinct codes, distinct normal ranges, distinct interpretations, etc.

ah reference range and all that jazz, ok, I see your point now. Thanks !

view this post on Zulip John Silva (Feb 23 2019 at 16:46):

It is also possible that certain values are part of different 'panel tests'. For example, the Hct (Hematocrit) can be ordered as a separate test or part of a CBC (Complete Blood Count) test, Similarly, a Sodium value can be part of a Complete Metabolic Panel or a Basic Metabolit Panel or an Electrolyte Panel ... see: http://laboratory.uchealth.com/tests/lab-panel-components/


Last updated: Apr 12 2022 at 19:14 UTC