fhir4ds.dqm

Digital Quality Measure orchestrator and audit engine for FHIR-based CQL measures.

1. At a Glance

The high-level orchestration interface:

from fhir4ds.dqm import (
    MeasureEvaluator,  # Main orchestrator
    AuditEngine,       # Evidence trail generator
    AuditMode,         # Enum: NONE, POPULATION, FULL
    AuditOrStrategy,   # Enum: ALL, TRUE_BRANCH
)

---

2. Class Reference

MeasureEvaluator

MeasureEvaluator(conn, audit_or_strategy=AuditOrStrategy.TRUE_BRANCH)

The primary class for executing end-to-end quality measures.

Parameter	Type	Description
conn	duckdb.DuckDBPyConnection	DuckDB connection with FHIR data loaded.
audit_or_strategy	AuditOrStrategy	(Optional) Strategy for OR evidence collection.

Methods

• evaluate(measure_bundle, cql_library_path, parameters=None, audit=False) -> MeasureResult: Evaluate a measure bundle against loaded FHIR data.
• summary_report() -> str: Generate a summary report of the last evaluation.

---

3. Data Structures

MeasureResult

The object returned by evaluation methods.

Property	Type	Description
populations	dict	Aggregated counts for each population (IP, NUMER, etc).
patient_results	list	Detailed results for every patient in the set.
evidence	dict	Logical evidence trails (if audit mode enabled).
sql	str	The optimized DuckDB SQL used for execution.

---

4. Configuration Enums

AuditMode

Controls the depth of evidence captured.

• NONE: No evidence capture (Fastest).
• POPULATION: Captures resources that satisfied population criteria.
• FULL: Captures every logical sub-expression evaluation.

AuditOrStrategy

Defines how evidence is collected for OR operations.

• ALL: Collect evidence from all branches of the OR.
• TRUE_BRANCH: Only collect evidence from the specific branch that evaluated to true.

---

5. Compliance

Metric	Status
CMS eCQM (QI-Core 2025)	42/46 Measures (100% accuracy)
Audit Coverage	100% expression-level traceability