Executing Digital Quality Measures (dQMs) - $cql, $evaluate, $evaluate-measure, $data-requirements

Note

The features described on this page are available in the following Firely Server editions:

Firely dQM - 🌍 / 🇺🇸

Important

Please see Intro to Digital Quality Measures for an introduction to Digital Quality Reporting in FHIR.

FHIR provides several operations for executing Digital Quality Measures (dQMs), either partially (e.g. evaluating expressions) or fully (evaluating a complete measure). The appropriate operation depends on the specific use case and execution goal:

Library/$evaluate is most commonly used for debugging purposes. dQMs frequently reference multiple Library resources to encapsulate modular logic. When a measure produces unexpected results—such as an incorrect or zero score—it is often useful to investigate why a particular subject meets or fails to meet specific population criteria (e.g., initial population, denominator, numerator). In such cases, it can be helpful to execute a targeted set of CQL expressions or evaluate specific sub-libraries within the measure. This allows implementers to isolate and verify individual components of the logic without executing the entire measure.
Measure/$measure-evaluate is the primary operation for executing a digital quality measure (dQM) as a whole. This operation evaluates a Measure resource against a specified subject (such as a patient, group) using the measurement period and any associated Library parameters. Measure/$evaluate-measure is typically used in production or formal testing scenarios to generate actual measure scores. It is also suitable for automated execution in quality reporting workflows. Unlike Library/$evaluate, which targets specific expressions, Measure/$evaluate-measure executes the full population logic and scoring methodology defined in the measure, making it the most comprehensive method for end-to-end dQM evaluation.
$cql allows direct execution of CQL expressions, either inline or from referenced libraries. It is useful for rapid testing or prototyping when measure logic needs to be validated independently of a Measure resource.

For the preperation of the execution data-requirements can be gathered using the following operations:

Library/$data-requirements retrieves a structured representation of the data required to evaluate a measure, making it a supporting operation for both Library/$evaluate and Measure/$evaluate-measure. This operation identifies the necessary FHIR data types and elements, including value sets, code filters, and date constraints. It returns a Library resource containing DataRequirement elements, which describe the expected inputs for measure evaluation. This operation is particularly useful during implementation, data mapping, and integration planning, as it helps clarify what data must be available for successful execution of a digital quality measure.
Measure/$data-requirements functions identically, but aggregates the data requirements across all libraries referenced by the measure, providing a complete picture of the inputs needed for measure evaluation.

Library/$evaluate

The Library/$evaluate operation executes one or more named CQL expressions within a FHIR Library resource in a given data context (e.g. a patient), and returns the evaluated results. It is primarily used to inspect and debug the logic underlying digital quality measures by allowing targeted execution of individual expressions without running the full measure.

Overview

Operation name

Library/$evaluate

FHIR specification

Using CQL with FHIR Implementation Guide - v2.0.0

OperationDefinition

http://hl7.org/fhir/uv/cql/OperationDefinition/cql-library-evaluate

Scope

Invocation level: type / instance
Supported resource type(s): Library
Idempotent: yes
Affects server state: no

HTTP methods

POST (type and instance level)
GET (type and instance level, when all parameters can be provided as query parameters)

Configuration

The Library/$evaluate operation is provided by the Vonk.Plugin.Cql.Operations.Library.Evaluate namespace.

You can enable or disable this operation by including or excluding this namespace in the Firely Server pipeline options. See Available features and configuration for more information on configuring available plugins.

You can configure the behavior of this operation using the LibraryEvaluateOperation section in the appsettings.

Database requirements

Execution of dQMs relies on retrieving clinical data from the Firely Server data store. Internally, Firely Server uses the $everything operation to collect all relevant data for a subject.

This functionality is only supported when the data store is backed by MongoDB or SQL Server. Therefore, to execute dQMs against data stored in Firely Server, the primary FHIR data database must use either MongoDB or SQL Server.

In addition, the Vonk.Plugin.PatientEverything plugin must be enabled in the pipeline options, as it provides the $everything operation used during data retrieval. See Available features and configuration for more information on configuring available plugins.

The administration database (used for conformance resources such as Library and Measure) can still be hosted on SQLite.

Alternatively, you can configure Firely Server to use only external data sources by enabling the RemoteDataEndpointsOnly setting. In that case, no local data retrieval (and thus no $everything support) is required.

External data endpoints

Firely Server can retrieve clinical and claims data from external FHIR endpoints during execution of Library/$evaluate.

This is used when the useServerData parameter is set to false in a request. In that case, data is not retrieved from the local Firely Server database, but from a configured external endpoint.

"LibraryEvaluateOperation": {
  "RemoteDataEndpointsOnly": false,
  "DataEndpoint": [
    //{
    //    "Endpoint": "<base url>",
    //    "MediaType": "application/fhir+json",
    //    "RemoteDataEndpointAuthentication": "Jwt",
    //    "ClientId": "",
    //    "ClientSecret": "",
    //    "TokenEndpoint": "",
    //    "Audience": "",
    //    "Scopes": "system/*.rs"
    //}
    "ForwardedHeaders": [
      "X-Custom-Auth-Header"
    ]
  ]
}

The DataEndpoint setting defines a list of pre-configured external FHIR endpoints. Each endpoint can be referenced in a request using the dataEndpoint parameter.

Any dataEndpoint parameter provided in a request must match one of the configured endpoints.

Each DataEndpoint entry supports the following fields:

Endpoint: Base URL of the external FHIR server
MediaType: The FHIR media type to use for requests to this endpoint ( application/fhir+json and application/fhir+xml are supported)
ClientId / ClientSecret: Credentials for authentication (if required)
TokenEndpoint: OAuth2 token endpoint (used for JWT authentication)
Audience: Optional audience claim for the access token
Scopes: Space-separated list of SMART on FHIR scopes. Since Firely Server uses a client_credentials flow, only system-level scopes should be used (e.g. system/*.rs).
RemoteDataEndpointAuthentication: Defines how Firely Server authenticates against the endpoint. Supported values include JWT and None

Important

Firely Server expects that the response of the remote $everything operation is returned as a single Bundle page. Pagination is not supported in this context, and Firely Server will not follow additional pages (e.g. via link[relation="next"]) returned by the remote endpoint.

The ForwardedHeaders setting can be used to forward custom HTTP headers from the incoming request to external data endpoints.

Any headers listed in ForwardedHeaders are copied from the original request to Firely Server and included in outgoing requests to configured DataEndpoint entries. This can be used to propagate request-specific context, such as correlation IDs or custom authorization headers, to external systems.

Supported parameters

Firely Server supports the following parameters:

Parameter	Supported	Type	Cardinality	Additional Notes
`url`	✅	`canonical`	0..1	Specifies the `Library` to evaluate via canonical URL. Earlier versions of the IG used `library` for this. Since v2.0.0, `library` is redefined to pass an inline `Library` resource. Firely Server uses `url` only for external logic. Versioned canonical references are allowed, e.g., `http://example.org/fhir/ Library/MyLogic\|1.0.0`.
`library`	✅	`Library` resource	0..1	In-line logic library that contains executable CQL logic. This Library will not be stored in Firely Server. It MAY only contain CQL and will be compiled dynamically. If ELM content is provided, it will be re-used.
`subject`	✅	`string`	0..1	Only Patient references are supported, may be omitted if no “context Patient” is included in the library.
`expression`	✅	`reference`	0..*	The name of the expression to evaluate. If omitted, all expressions in the library are evaluated. CQL Access Modifier extensions are not taken into account.
`parameters`	✅	`Parameters` resource	0..1	Input parameters passed into the evaluation context. These will be mapped from FHIR data types to CQL data types according to the FHIR Type Mapping. Most notably, this includes passing in the measurement period parameter as a FHIR Period.
`raw`	✅	`boolean`	0..1	Return the library results as a string without mapping the CQL result data types back to FHIR. This is a proprietary parameter of Firely Server.
`useServerData`	✅	`boolean`	0..1	When `true`, claims and clinical data are retrieved from the Firely Server database where the operation is executed. When `false`, claims and clinical data are retrieved from the `dataEndpoint` parameter. In both cases, any data passed via the `data` parameter takes precedence.
`data`	✅	`Bundle`	0..1	Inline FHIR data bundle to use as the data context during evaluation. The bundle type SHOULD be either `collection` or `searchset` (as the output of a $everything operation).
`dataEndpoint`	✅	`Endpoint` resource	0..1	Used only when `useServerData` is `false`. Defines the external FHIR endpoint from which claims and clinical data are retrieved. The endpoint must be pre-registered in `LibraryEvaluateOperation` via the `DataEndpoint` option. See Quality Measure Operations. Data supplied via the `data` parameter always takes precedence.
`prefetchData`	❌	Complex	0..*
`includePrivate`	❌	`boolean`	0..1
`contentEndpoint`	❌	`Endpoint` resource	0..1
`terminologyEndpoint`	❌	`Endpoint` resource	0..1	External terminology services should be configured via the Terminology services options.

Important

If the Library references any ValueSet resources, they must be preloaded into the Firely Server’s administration endpoint before executing the Library.

Output parameters

The Library/$evaluate operation returns a Parameters resource containing the results of the evaluated CQL expressions.

Parameter	Type	Cardinality	Description
`return`	Parameters	1..1	A Parameters resource in which each output parameter corresponds to a named CQL expression. Each entry includes the expression name, its evaluated value, and optional type information via an extension ( `cqf-cqlType`).

When to use this operation

Use Library/$evaluate when you want to:

evaluate specific expressions within a Library
debug measure logic
inspect intermediate results of CQL execution

Example: Type-Level Library/$evaluate Invocation

This example evaluates the bp-check-logic library (version 1.0.0) against a specific patient and a defined measurement period using a POST request to the type-level operation.

Request

POST [base]/Library/$evaluate

Request Body

{
  "resourceType": "Parameters",
  "parameter": [
    {
      "name": "url",
      "valueCanonical": "http://example.org/fhir/Library/bp-check-logic|1.0.0"
    },
    {
      "name": "subject",
      "valueString": "Patient/cql-patient-test"
    },
    {
      "name": "parameters",
      "resource": {
        "resourceType": "Parameters",
        "parameter": [
          {
            "name": "Measurement Period",
            "valuePeriod": {
              "start": "2023-01-01",
              "end": "2023-12-01"
            }
          }
        ]
      }
    }
  ]
}

Response Body

Given matching input data (see Example Library for context), specifically, a Patient resource and an Observation with a code of 8480-6 from the LOINC CodeSystem, and an effectiveDateTime that falls within the measurement period — the following output will be returned:

{
  "resourceType": "Parameters",
  "parameter": [
    {
      "extension": [
        {
          "url": "http://hl7.org/fhir/StructureDefinition/cqf-cqlType",
          "valueString": "Fhir"
        }
      ],
      "name": "Patient",
      "resource": {
        "resourceType": "Patient",
        "id": "cql-blood-pressure-check-test-match",
        "meta": {
          "versionId": "d36e61f8-300a-4c2f-8247-9fb4a6837236",
          "lastUpdated": "2025-05-23T18:32:44.106+00:00"
        },
        "birthDate": "1990-06-15"
      }
    },
    {
      "extension": [
        {
          "url": "http://hl7.org/fhir/StructureDefinition/cqf-cqlType",
          "valueString": "Boolean"
        }
      ],
      "name": "HasBPReading",
      "valueBoolean": true
    },
    {
      "extension": [
        {
          "url": "http://hl7.org/fhir/StructureDefinition/cqf-cqlType",
          "valueString": "Boolean"
        }
      ],
      "name": "AdultPatients",
      "valueBoolean": true
    }
  ]
}

Measure/$evaluate-measure

The Measure/$evaluate-measure operation executes a complete digital quality measure (dQM) and returns the calculated results for a given subject or population. It evaluates all referenced Library resources, applies the defined population criteria (e.g. initial population, denominator, numerator), and computes the final measure score.

This operation is the primary mechanism for end-to-end measure evaluation and is typically used in production or formal testing scenarios.

Overview

Operation name

Measure/$evaluate-measure

FHIR specification

FHIR Core Measure Evaluation

OperationDefinition

http://hl7.org/fhir/OperationDefinition/Measure-evaluate-measure

Scope

Invocation level: type / instance
Supported resource type(s): Measure
Idempotent: yes
Affects server state: conditional

HTTP methods

POST (type level)
GET (type level, when all parameters can be provided as query parameters)

Important

Invocation at the instance level ([base]/Measure/[id]/$evaluate-measure) is not currently supported. Use the type-level operation with the measure parameter instead.

Note

The operation can optionally affect server state depending on the persist parameter.

When persist is set to true, the generated MeasureReport is stored on the server. By default (persist = false), the report is returned in the response only and is not persisted.

Configuration

The Measure/$evaluate-measure operation is provided by the Vonk.Plugin.Cql.Operations.Measure.EvaluateMeasure namespace.

You can enable or disable this operation by including or excluding this namespace in the Firely Server pipeline options. See Available features and configuration for more information.

Supported parameters

Firely Server supports the following parameters:

Parameter	Supported	Type	Cardinality	Additional Notes
`url`	✅	`canonical`	0..1	Canonical URL of the Measure to evaluate. Required for type-level invocation. Versioned canonical references are allowed, e.g., `http://example.org/fhir/Measure/ ExampleMeasure\|1.0.0`.
`subject`	✅	`string`	1..1	Reference to the subject for which the measure is evaluated. Supported resource types are `Patient` and `Group`. When a `Patient` is provided, the measure is evaluated for that single subject. When a `Group` is provided, the measure is evaluated for all `Patient` references contained in the Group.
`periodStart`	✅	`date`	1..1	Start of the measurement period.
`periodEnd`	✅	`date`	1..1	End of the measurement period.
`reportType`	✅	`code`	0..1	The type of measure report: `individual`: Evaluates the measure for a single subject (e.g. Patient or Group) and returns population membership and score for that subject. `summary`: Evaluates the measure across a population of subjects and returns aggregated counts (e.g. numerator, denominator). The `subject-list` report type defined in the FHIR specification is not supported. If not specified, the default is `individual`.
`parameters`	✅	`Parameters` resource	0..1	See `Library/$evaluate` configuration for details.
`useServerData`	✅	`boolean`	0..1	See `Library/$evaluate` configuration for details.
`data`	✅	`Bundle`	0..1	See `Library/$evaluate` configuration for details.
`dataEndpoint`	✅	`Endpoint`	0..1	See `Library/$evaluate` configuration for details.
`persist`	✅	`boolean`	0..1	When `true`, the generated `MeasureReport` is stored on the server. When `false` (default), the result is returned in the response only. This is a proprietary parameter of Firely Server.
`raw`	✅	`boolean`	0..1	Return the results as a string without mapping the CQL result data types back to FHIR. This is a proprietary parameter of Firely Server.
`measure`	❌	`Measure`	0..1
`version`	❌	`string`	0..1
`provider`	❌	`string`	0..1
`location`	❌	`string`	0..1
`lastReceivedOn`	❌	`dateTime`	0..1

Output parameters

The operation returns a MeasureReport resource containing the evaluation results.

The report includes:

population counts (e.g. initial population, denominator, numerator)
measure score (if applicable)
subject-level or population-level results depending on reportType

When to use this operation

Use Measure/$evaluate-measure when you want to:

execute a full digital quality measure
calculate population membership and scores
generate results for reporting or submission
validate measure behavior in end-to-end scenarios

Example: Type-Level Measure Evaluation

Request

POST [base]/Measure/$evaluate-measure

Request Body

{
  "resourceType": "Parameters",
  "parameter": [
    {
      "name": "url",
      "valueCanonical": "http://example.org/fhir/Measure/ExampleMeasure|1.0.0"
    },
    {
      "name": "subject",
      "valueString": "Patient/cql-patient-test"
    },
    {
      "name": "periodStart",
      "valueDate": "2023-01-01"
    },
    {
      "name": "periodEnd",
      "valueDate": "2023-12-31"
    }
  ]
}

Response Body

{
  "resourceType": "MeasureReport",
  "status": "complete",
  "type": "individual",
  "measure": "http://example.org/fhir/Measure/ExampleMeasure|1.0.0",
  "subject": {
    "reference": "Patient/cql-patient-test"
  },
  "period": {
    "start": "2023-01-01",
    "end": "2023-12-31"
  },
  "group": [
    {
      "population": [
        {
          "code": {
            "coding": [
              {
                "system": "http://terminology.hl7.org/CodeSystem/measure-population",
                "code": "initial-population"
              }
            ]
          },
          "count": 1
        },
        {
          "code": {
            "coding": [
              {
                "system": "http://terminology.hl7.org/CodeSystem/measure-population",
                "code": "denominator"
              }
            ]
          },
          "count": 1
        },
        {
          "code": {
            "coding": [
              {
                "system": "http://terminology.hl7.org/CodeSystem/measure-population",
                "code": "numerator"
              }
            ]
          },
          "count": 1
        }
      ],
      "measureScore": {
        "value": 1.0
      }
    }
  ]
}

$cql

The $cql operation executes a CQL expression directly and returns the evaluated result. It is useful for rapid testing or prototyping when CQL logic needs to be validated independently of a Library or Measure resource.

Overview

Operation name

$cql

FHIR specification

Using CQL with FHIR Implementation Guide - v2.0.0

OperationDefinition

http://hl7.org/fhir/uv/cql/OperationDefinition/cql-cql

Scope

Invocation level: system
Idempotent: yes
Affects server state: no

HTTP methods

POST

Supported parameters

Firely Server supports the following parameters:

Parameter	Supported	Type	Cardinality	Additional Notes
`expression`	✅	`string`	1..1	Specifies an inline CQL expression to be executed. Only a single statement is supported per request. It cannot operate within a context (e.g. Patient) and will not execute correctly if input parameters are needed.
`subject`	✅	`string`	0..1	Only Patient references are supported.
`parameters`	✅	`Parameters`	0..1	Input parameters passed into the evaluation context. See `Library/$evaluate` configuration for details.
`raw`	✅	`boolean`	0..1	Return the execution results as a string without mapping the CQL result data types back to FHIR. This is a proprietary parameter of Firely Server.
`library`	❌	Complex	0..*
`useServerData`	❌	`boolean`	0..1
`data`	❌	`Bundle`	0..1
`prefetchData`	❌	Complex	0..*
`dataEndpoint`	❌	`Endpoint`	0..1
`contentEndpoint`	❌	`Endpoint`	0..1
`terminologyEndpoint`	❌	`Endpoint`	0..1

Output parameters

The operation returns a Parameters resource containing the result of the evaluated CQL expression.

The result is returned in a parameter named return. The value is mapped back to a FHIR data type, unless the proprietary raw parameter is set to true.

When to use this operation

Use $cql when you want to:

quickly test a simple CQL expression
validate basic CQL syntax or behavior
prototype logic before moving it into a Library
execute logic that does not require a full Measure evaluation

Example: System-Level $cql Invocation

This examples demonstrates a simple calculation executed via the dQM engine.

Request

POST [base]/$cql

Request Body

 {
  "resourceType": "Parameters",
  "parameter": [
      {
          "name": "expression",
          "valueString": "'Hello'&' '&'World'"
      }
  ]
}

Response Body

 {
  "resourceType": "Parameters",
  "parameter": [
      {
          "extension": [
              {
                  "url": "http://hl7.org/fhir/StructureDefinition/cqf-cqlType",
                  "valueString": "String"
              }
          ],
          "name": "return",
          "valueString": "Hello World"
      }
  ]
}