(C-176) Development and Verification of Robust Comprehensive Code Lists to Ensure Consistency Across Global Real-World Evidence Studies

Background: The FDA guidance issued December 2023 indicates that the “FDA considered the operational definition to include three components:” time frame, specific codes/component(s) assessed, and algorithm for combining the components. The authors have previously presented on the importance and value of computable operational definitions (CODefs) used in cohort selection, case ascertainment, cohort characterization, and outcomes assessment for real-world evidence (RWE) research. Often these CODefs are based on code lists (value sets). However, significant inconsistency arises, largely due to different coding systems across geographic regions and often no single reference source containing a comprehensive list of codes/medications for a given clinical concept.

Objectives: We developed and verified a robust framework using multiple sources and clinical input for CODef development to enable effective global RWE programs, with comprehensive and consistent code lists.

Methods: A robust process was used to develop comprehensive value sets across multiple code systems:
1. Identify the clinical concepts needed for the RWE study;
2. Review multiple global/regional reference sources (eg, general, clinical, medical coding);
3. Aggregate codes/medications across the multiple sources to generate an initial comprehensive value set;
4. Conduct medical review to verify the value set;
5. Map to relevant code systems (e.g., ICD-10-CM, ICD-10-WHO, SNOMED CT, RxNorm, ATC-5, HCPCS) and link together;
6. Organize linked value sets by relevant indication and therapeutic area.

Results: A few examples:
- Antidepressant Therapies (step #1): 43 references (step #2) identify the 7 drug classes that collectively include 104 total meds after medical review (steps #3 & 4) - 69 ATC-5, 68 RxNorm, and 13 HCPCS codes (step #5), which is part of the Depression indication library (step #6).
- Antidiabetic Therapies: 38 references for 10 drug classes of 146 meds - 119 ATC-5, 112 RxNorm, and 20 HCPCS codes.
- Atherosclerotic Cardiovascular Diseases: 6 references for 9 distinct medical conditions - 793 ICD-10-CM, 60 ICD-10-WHO, and 422 SNOMED-CT US codes.
Further examples and details will be presented.

Conclusions: This methodology demonstrates the creation of comprehensive value sets that ensure consistency across code systems for global RWE studies, which allow for appropriate comparison of scientific results/evidence across studies.