JMIR Publications

ABSTRACT

Background:

Clinical Quality Language (CQL) paired with Fast Healthcare Interoperability Resources (FHIR) supports standards-based electronic clinical quality measures (eCQMs), but translating national claims-based indicator definitions into executable logic remains challenging because terminology, exclusions, and data semantics often require local adaptation. To our knowledge, the Taiwan National Health Insurance (NHI) medical-center quality indicators have not previously been represented as an executable CQL library.

Objective:

This study aimed to develop a FHIR R4-based execution platform and executable CQL libraries derived from 39 Taiwan NHI medical-center quality indicator codes, and to evaluate technical executability, internal computational consistency, and initial real-world differential concordance, using cross-institution medication overlap as the principal use case.

Methods:

We developed 39 executable CQL indicator libraries across 6 principal clinical FHIR R4 resources, using 21 data elements. Libraries were compiled to Expression Logical Model (ELM) and executed with cql-execution and cql-exec-fhir against deterministic, author-defined synthetic fixtures; rules depending on claims-specific fields, code systems, or linkage semantics that were unavailable in the study data model were implemented as documented adaptations. We then conducted a retrospective 2-campus study using deidentified NHI outpatient claims and cross-institution medication records. For the 20 indicators supported by that export, a separately coded Python reference pipeline and the platform CQL pipeline computed aggregate numerators and denominators from the same records, represented as a long table and as FHIR R4 resources, respectively.

Results:

All 39 CQL libraries compiled and executed against the predefined synthetic fixtures, and their outputs matched the author-defined expected results (39/39 libraries; 136/136 early-refill grace-period control groups). The real-world dataset included 88 patients (Taipei, n=65; Linkou, n=23), 1597 medication records, and 630 visits from January 1 to June 2, 2026. Across the 20 medication-based indicators, the 2 pipelines produced identical aggregate numerators and denominators (20/20), and 7 of the 8 cross-institution drug-class overlap indicators had nonzero numerators. All 2317 FHIR resources passed base FHIR R4 structural validation, and ATC mapping coverage was 98.3% (1570/1597 records).

Conclusions:

Taiwan NHI quality indicator logic can be represented as executable CQL operating on FHIR R4 data, with technical execution and internally consistent results across synthetic verification and initial aggregate differential testing. Indicators depending on claims-specific fields or code lists were implemented as documented adaptations. The library establishes a defined foundation for the next validation phases: expert rule adjudication, patient- and event-level concordance testing, TW Core conformance, and cross-engine evaluation.