Per QICore 4.1.1 StructureDefinition for MedicationRequest, it looks like doNotPerform is fixed to false. However, the QICore 4.1.1 patterns wiki page states "...the MedicationRequest profile does not fix the value of the doNotPerfom element...". Will you consider updating the wiki page to provide more background as to why specifying "doNotPerform is not true" in the CQL is required to express MedicationRequest? Thank you in advance!

https://hl7.org/fhir/us/qicore/STU4.1.1/StructureDefinition-qicore-medicationrequest.html
https://github.com/cqframework/CQL-Formatting-and-Usage-Wiki/wiki/Authoring-Patterns---QICore-v4.1.1#medication-ordered

The current Hospitalization functions provide a way to define criteria with respect to an "episode-of-care" whether that is a single in-patient encounter, or an in-patient encounter with an immediately preceding emergency department visit. However, there are scenarios that need to be able to account for episodes of care that may include observation time, as well as potential emergency department visits. To account for this, new hospitalization functions that account for observation status should be added, as described here:

https://github.com/esacinc/CQL-Formatting-and-Usage-Wiki/blob/master/Source/Cooking%20With%20CQL/27/67_ObservationStatus.cql#L188

There is a Global.ToDate defined in the Common library. CQL 1.3 defines a System ToDate function that should be used instead, so the ToDate in the Global Common Library should be removed.

For example, updates aliases like O.status to conform to the CQL style guide

https://github.com/cqframework/CQL-Formatting-and-Usage-Wiki/blob/master/Source/Libraries/MATGlobalCommonFunctions-7.0.000.cql

This is more of a stylistic question, but should ToDateInterval function be renamed to reflect that this function provides a closed interval?

https://github.com/esacinc/CQL-Formatting-and-Usage-Wiki/wiki/eCQM-Known-Issues#cql-clarifications-and-enhancements says that May 2019 will be a Normative ballot for CQL. Latest thinking is that May 2019 will be a STU ballot, and a Normative ballot would follow in a future ballot cycle.

Once this known issue with the production release of the translator (1.3.10) is addressed, the built-in CQL age calculation functions can be used. Consider updating the GCL to remove these functions as part of the next AU (2020).

https://github.com/esacinc/CQL-Formatting-and-Usage-Wiki/blob/master/Source/Cooking%20With%20CQL/43/38_SingletonExists.cql#L31

Remove "exists" from this line.

https://github.com/cqframework/CQL-Formatting-and-Usage-Wiki/blob/master/Source/Libraries/CumulativeMedicationDuration-1.0.000.cql#L9-L38

These 2 code systems, and the associated codes, are not in VSAC.
I could not find the code systems using https://vsac.nlm.nih.gov/context/cs

Looking for them using FHIR also comes up empty
curl -s -k -u apiKey:your-key-here https://cts.nlm.nih.gov/fhir/r4/CodeSystem?_format=json^&name=timing
curl -s -k -u apiKey:your-key-here https://cts.nlm.nih.gov/fhir/r4/CodeSystem?_format=json^&name:contains=timing

Therefore, these codes cannot be added when creating this library in MAT (for QDM); apparently MAT does not verify with VSAC that those code systems exist since a FHIR version is in MAT.

http://tx.fhir.org/r4/CodeSystem?_format=json&name:contains=timing finds 4 code systems including both urn:oid:2.16.840.1.113883.4.642.4.76 and urn:oid:2.16.840.1.113883.5.139 (with 2 versions of ..5.139)

https://r4.ontoserver.csiro.au/fhir/CodeSystem?_format=json&name:contains=timing finds them

MedicationRequest.doNotPerform is fixed to false in QICore 4.1.1. Therefore, is there a need to check for "doNotPerform is not true" in the CQL?

https://github.com/cqframework/CQL-Formatting-and-Usage-Wiki/wiki/Authoring-Patterns---QICore-v4.1.1#medication-ordered
http://hl7.org/fhir/us/qicore/STU4.1.1/StructureDefinition-qicore-medicationrequest-definitions.html#MedicationRequest.doNotPerform

The ToDaily(Code) function in this library incorrectly uses the lower number of doses for frequencies that are specified as a range of hours:

https://github.com/cqframework/CQL-Formatting-and-Usage-Wiki/blob/master/Source/Libraries/CumulativeMedicationDurationFHIR4-1.0.000.cql#L174

In addition, add this comment to the description of the function:

Note that for ranges, the boundary resulting in the highest frequency result is used. E.g. `Every two to four hours` results in `12.0` and `One to two` times a day results in `2.0`.

Based on implementer feedback in CQLIT-117, the assumption that there would only ever be one immediately preceding emergency room visit within an hour of an inpatient encounter for a given patient does not hold. Guidance from the resolution is that the most recent preceding emergency room visit should be used. The hospitalization functions should be updated to reflect this change.

Since different terminology systems are managed by different organizations and have different update cycles, measure content typically avoids referencing specific versions of terminologies (value sets and code systems) until there is a publication date, at which time terminology version information is typically provided through a different mechanism (the Binding Parameters Specification for example).

However, the 2.0.0 version of the common library references code system versions specifically. These versions should be removed.

Because CQL 1.2 defines Age calculations to account for birth time, existing measures use the CalendarAgeInXXX functions to determine patient age. However, in CQL 1.3, a new Date type was introduced to allow Age functions to use birth date or birth time, as appropriate. The common library should be updated to take advantage of this new functionality.