Presentation (CTI)

OBJECTIVES: To validate estimates of primary care resource use and costs from OMOP-mapped Clinical Practice Research Datalink (CPRD) Aurum data against those from the source CPRD Aurum, using a UK cohort of postmenopausal women with fragility fractures.
METHODS: We conducted a retrospective study (2010-2018) of women aged >50 with a first recorded fragility fracture (index). Clinically-relevant primary care encounters and associated costs were estimated over two years post-index date (or until censoring) using both OMOP-mapped and source CPRD Aurum datasets. Multiple encounters with the same healthcare specialty, on the same day, and for the same patient were considered duplicates and only one such record kept. Outcomes included encounter counts and cost estimates (2024 GBP). Differences were assessed using means and 95% confidence intervals from 1,000 bootstrap samples.
RESULTS: The OMOP-mapped cohort included 22,900 women (mean follow-up: 578 days), and the CPRD cohort 23,106 (577 days), with 22,571 appearing in both. At least one encounter was recorded for 97.3% (OMOP) and 97.0% (CPRD). OMOP data showed 9.2% more encounters (537,756 vs. 488,449), with significant variation across specialties. In OMOP, GPs accounted for 73.9% of encounters, nurses 7.9%, other specialties 12.5%, and 5.8% had “No matching concept” (GP Registrar, Seasonal and Locum GP, and Consultant in CPRD). In CPRD, these proportions were 82.9%, 8.3%, and 8.8%, respectively.Mean encounters per patient were 24.1 (CI: 23.8-24.5) in OMOP vs. 21.8 (21.5-22.1) in CPRD. Mean costs were £939.1 (CI: £927.2-£951.6) in OMOP vs. £858.9 (CI: £847.8-£869.1) in CPRD.
CONCLUSIONS: Primary care resource use and costs were slightly higher in the OMOP-mapped dataset. Differences likely stem from specialty mapping, which can be refined. OMOP-mapped analyses can yield results comparable to source CPRD, supporting the conduct of health economics research in single and federated approaches leveraging OMOP-mapped datasets.