Journal: Clinical Epidemiology
Authors: Catherine Vasilakis-Scaramozza, Katrina Wilcox Hagberg, Rebecca Persson, George Kafatos, Joe Maskell, David Neasham, Susan Jick
NLM Citation: Vasilakis-Scaramozza C, Hagberg KW, Persson R, Kafatos G, Maskell J, Neasham D, Jick S. Comparison of Rheumatoid Arthritis Information Recorded in UK CPRD Aurum and CPRD GOLD Databases (Companion Paper 3). Clin Epidemiol. 2023 Dec 16;15:1207-1218. doi: 10.2147/CLEP.S434831. PMID: 38126003; PMCID: PMC10731986.
Abstract
Purpose: To report distribution of codes associated with a rheumatoid arthritis (RA) diagnosis recorded in Clinical Practice Research Datalink (CPRD) Aurum compared to the previously validated CPRD GOLD database as a critical step toward making decisions about CPRD Aurum’s suitability for medical research.
Patients and methods: We analyzed the distribution of codes for RA diagnoses, labs, and treatments in the new CPRD Aurum database, compared to the CPRD GOLD database by selecting relevant indicators of RA diagnosis, treatment, and clinical care. We included all patients in England in CPRD Aurum and CPRD GOLD with an incident diagnosis code for RA on or after 1 January 2005 and at least two years recorded data before first RA diagnosis.
Results: We found 53,083 and 18,167 patients with a new diagnosis code for RA in CPRD Aurum and CPRD GOLD, respectively. In both databases approximately 67% were female with similar mean ages at first diagnosis. There were few differences in RA-related recording patterns between the two data sources. Before first RA diagnosis, CPRD Aurum patients had more RA-specific labs and other supporting clinical codes. After diagnosis, CPRD Aurum patients had more RA diagnoses coded and more often had 10+ general RA labs than patients in CPRD GOLD. More CPRD GOLD patients had 10+ prescriptions for conventional disease-modifying antirheumatic drugs (cDMARD) compared to CPRD Aurum. Otherwise, the distribution of drugs used to treat RA was similar between databases. The standardized incidence of RA was similar between databases.
Conclusion: Overall, among patients with a diagnosis code for RA, recording of diagnoses, prescription drugs, and labs were similar between CPRD Aurum and CPRD GOLD. Slight differences were found for a few variables, but overall, we found consistency between the databases. In addition, standardized incidence of RA was similar between databases.
Keywords: CPRD Aurum; CPRD GOLD; Clinical Practice Research Datalink; data quality; validation.
