Clinical Quality Framework Common FHIR Assets (US-Based)
0.1.0 - CI Build United States of America (USA)

Clinical Quality Framework Common FHIR Assets (US-Based), published by Clinical Quality Framework. This guide is not an authorized publication; it is the continuous build for version 0.1.0 built by the FHIR (HL7® FHIR® Standard) CI Build. This version is based on the current content of https://github.com/cqframework/cqf-us/ and changes regularly. See the Directory of published versions

Tooling Support

This topic provides a set of requirements based on community engagement and stakeholder feedback to understand the level of effort required to achieve the end goal of:

  • Supporting a user with deep clinical knowledge but limited technical knowledge being able to build and maintain CQL-based questionnaires
  • Potentially prototype the inclusion of CQL authoring capability in a questionnaire authoring environment such as the LHC forms builder

Requirements

Questionnaire Authoring

This topic describes questionnaire-related authoring capabilities that would be needed to support creation of prior-authorization questionnaires:

  • Create and maintain questionnaires
    • Edit basic (shareable) metadata including title, url, status and version
    • Edit advanced (publishable) metadata including description, purpose, approvalDate, lastReviewDate, useContext
  • Create single-item questions
    • Edit basic question information including
      • Question text
      • data type
      • code
      • initial value (as a literal)
      • cardinality
      • required
      • read only
    • Edit advanced question information including
      • Display conditions
      • Terminology server reference
      • Population and extraction capability
  • Create groups of questions
    • Arbitrary nesting of groups

The LHC Forms Builder is an open-source questionnaire authoring environment that supports most of this capability (items in bold are not currently supported)

CQL Support

In addition to questionnaire authoring capability, the following are areas within that environment that would need to be able to support CQL usage:

  • For single-item questions:
    • Initial value expressions
    • Calculated value expressions
    • Enable-when expressions
  • For group questions:
    • Population expressions
    • Extraction expressions
    • Enable-when expressions
  • Consider the ability to map the code of an question to an existing expression
    • Similar to the way that LHC supports inclusion of a question “based on” a particular LOINC code
    • Consider providing the expression associated with that code automatically
  • Consider the ability to map questions or groups of questions to an existing expression
    • Importing a question based on the result of an expression, e.g. Patient history

This support would generally fall into two categories:

  • Discovery and Reference: The ability for the environment to search for and provide a reference to an existing expression within a CQL library that could be used as is
  • Expression and Element Authoring: The ability for the environment to support authoring new elements and expressions of CQL directly.
Discovery/Reference
  • Support searching for relevant libraries:
    • Libraries that reference a particular implementation guide
    • Libraries with a particular topic
  • Support including a found library in the references for a questionnaire
  • Support searching for relevant expressions:
    • Expressions that reference a particular data type/profile
    • Expressions that reference a particular code system or value set
    • Expressions that reference particular codes (either directly or indirectly through value set references)
  • Support including a reference to a found expression in the expression for a given questionnaire item

Existing mechanisms that could be used (and potentially enhanced) to support this aspect include:

  • NPM Packages - The existing NPM packaging mechanisms for FHIR IGs support distribution of Library resources. The existing NPMPackageManager interface in the Java-based tooling could be used to provide these search capabilities
  • CRMI Artifact Repository - The Artifact Repository capability statement defined by the Canonical Resource Management Infrastructure implementation guide provides a basis for the searching capability required here, enhanced search parameters would need to be defined for:
    • Libraries that reference an IG
    • Libraries with expressions that reference a particular data type/profile
    • Libraries with expressions that reference particular codes (either directly or indirectly through value set references)
Expression and Element Authoring
  • Support creation of a library
    • Inline expressions (the entire expression created is inline in the questionnaire element)
    • Constructed expressions (the expression in the questionnaire element is a reference to an expression in a Library)
  • Support construction of element expressions
    • Support relevant element types
      • e.g. Allergy/Intolerance, Condition, Procedure, MedicationRequest, etc.
    • Support terminology selection
      • ValueSet
      • Code
    • Support filter expressions
      • Logical groupings (and/or)
      • Presence/absence
      • Terminology (in/~)
      • Comparison
      • Temporality
      • Relationships to other elements (including contextualization)
    • Support reuse of existing elements
  • Support construction of inclusion expressions
    • Boolean-valued expressions that can reference elements (existing or in-line)
    • Filtering expressions

The CDS Connect Authoring is an open source CQL authoring environment that supports much of this functionality