Assessment of a Modified High-Dimensional Propensity Score Approach for UK Electronic Health Record Data: Evaluating Upper Gastrointestinal Safety of NSAIDs and COX-2 Inhibitors - PubMed
6 hours ago
- #confounder adjustment
- #pharmacoepidemiology
- #electronic health records
- The study extends a modified high-dimensional propensity score (HDPS) approach for UK electronic health records by incorporating hospital data.
- It evaluates the upper gastrointestinal safety of COX-2 inhibitors versus NSAIDs, a well-known association.
- The research used UK primary care data linked with hospitalization and mortality records, focusing on osteoarthritis patients initiating NSAIDs or COX-2is between 2000-2004.
- Results showed that including hospital data in HDPS provided additional confounder proxies, with 26% of HDPS covariates originating from hospitalization data.
- The modified HDPS approach yielded results consistent with previous studies and randomized trials, suggesting improved confounding control in UK EHR data.
- The study highlights the potential benefits of enriching primary care data with hospital data for more accurate pharmacoepidemiological analyses.