(B-061) Utilization of Nationwide Claims Data Mapped to OMOP Common Data Model to Enhance Post-Marketing Drug Safety Studies

Background: Timely and comprehensive post-marketing surveillance is crucial for generating real-world evidence to support regulatory decision-making. In South Korea, nationwide claims data provided by the Health Insurance Review and Assessment (HIRA) Service have been standardized using the Observational Medical Outcomes Partnership Common Data Model (OMOP CDM) under the Observational Health Data Sciences and Informatics (OHDSI) initiative.

Objectives: To conduct a case study to explore utilization of HIRA OMOP CDM for drug safety studies.

Methods: We utilized national claims data (~56 million) individuals between January 2012 and December 2021, standardized to OMOP CDM Version 5.3.1. A retrospective cohort study was conducted to assess the association between sacubitril/valsartan and the risk of dementia, a post-marketing study required by the U.S. FDA. The study employed an active-comparator new-user design using the CohortMethod package within the HADES framework developed by the OHDSI community. Three key cohorts were defined: (1) Target cohort: patients initiating sacubitril/valsartan for heart failure, (2) Comparator cohort: patients initiating renin-angiotensin-aldosterone system (RAAS) inhibitors for heart failure, (3) Outcome cohort: patients diagnosed with dementia and prescribed dementia medications within seven days before or after diagnosis. A large-scale propensity score (PS) was estimated using large-scale regularized regression, and PS stratification and 1:4 PS matching were conducted to balance the two groups. Patients were followed up from the initiation date of sacubitril/valsartan or RAAS inhibitors, depending on the cohort, and both on-treatment and intention-to-treat approaches were used. Hazard ratios (HR) with 95% confidence intervals (CIs) were estimated using large-scale Cox regression. To assess potential confounding, various negative control outcomes were examined.

Results: We identified 9,159 new users of sacubitril/valsartan and 36,359 new users of RAAS inhibitors. For on-treatment analysis, no significant association between sacubitril/valsartan and dementia risk was observed after stratification (HR 0.72, 95% CI 0.33-1.44) or 1:4 matching (1.47, 0.32-7.56). For intention-to-treat analysis, no significant association was observed after stratification (0.86, 0.56-1.28), 1:4 matching (0.80, 0.44-1.39). Null findings were found in negative control outcome analyses.

Conclusions: The results were consistent across all analyses and with findings from randomized controlled trials and observational studies. Our findings support the reliability of HIRA OMOP CDM for post-marketing surveillance, reinforcing its potential to generate real-world evidence for regulatory decision-making.