SEverity of condition

[diatomsRcool]

Do we need an attribute on Condition for reporting the severity? Like severe asthma, or mild sleep apnea? Or would there be two different concept ids for severe, mild, moderate conditions?

[diatomsRcool]

mondo doesn't have a class for "severe asthma"

[bfurner]

FHIR's Condition resource includes severity as an attribute. Can certainly include

[diatomsRcool]

Maybe take a look at Sleep Apnea Status. Would this be a Condition with a severity attribute or a Measurement Enum?

[bfurner]

What are the permissible values for Sleep Apnea Status? I see the mapping to OMOP Athena term, but that doesn't include severity

[diatomsRcool]

This is what google says. Based on the available documentation, I think these are the values that will be in the data. Without looking at the data, I can't make a better guess.

[bfurner]

it would be helpful to look at some example data for this. i'm still not sure what to expect in the data for a single patient on a single study (i.e. numbers only, categorical only, numbers + categorical). if someone can either walk through this or pull summary statistics for AHI data in our priority data sets that would be super helpful.

[diatomsRcool]

I think Ozzy may have pulled all the unique values from each variable. Let me look.

[diatomsRcool]

According to this documentation, sleep apnea is qualitative and AHI is quantitative and they are held as two separate variables.

[diatomsRcool]

OK, at this point, for our purposes, I think sleep apnea should be a MeasurementObservation.value_enum. The broader question of different ways to represent the same observation can be a question for another day. Maybe a post-pilot question. How does that sound?

[diatomsRcool]

This one might force our hand tho.

[diatomsRcool]

Now this is just me brain dumping....
One way to indicate that multiple values coming from the same "measurement" is to link them both to the same procedure. For example, a SleepStudyProcedure would produce two MeasurementObservations an AHI and a sleep apnea status. Does that sound right?

[bfurner]

As is often the case, reasonable arguments can be made for mapping these things in different ways.

Assuming that Sleep apnea and Asthma are both status indicators (i.e. 'Present', 'Absent', 'Unknown'), these seem like they should be recorded as Condition records.

For "severity" measurements of conditions, we have a few options that aren't mutually exclusive. 1) add a severity attribute to the Condition class with an enum that has permissible values of 'Mild', 'Moderate', 'Severe' and record severity measurements there. 2) record severity measurements in a MeasurementObservation record

AHI feels like a MeasurementObservation to me. While it represents a measure of the severity of sleep apnea, it is a standardized index and people may want to specifically look for 'AHI' measurements and if the values are recorded in a Condition.severity attribute the fact that these values represent AHI values will be lost. If they are recorded as a key-value pair in an MeasurementObservation, the observation_type can explicitly be 'AHI' and the value_enum can record the severity value.

AsthmaSeverity on the other hand does seem like it might best be recorded in a Condition.severity attribute since it does not seem to be a standardized index.