Describing the Management of Spondyloarthropathies in French Primary Care: An EHR Study Exploring Regional Differences

OBJECTIVES: Ankylosing spondylitis (AS) and psoriatic arthritis (PA) affect respectively 0.3% and 0.19% of the French population. These spondyloarthropathies’ diagnoses and treatments are often long delayed. This study aimed to unveil correlations between the disease management variables at general practitioners (GPs) and rheumatologists.

METHODS: THIN France (medicalized EHR) data for GPs and rheumatologists and institutional medical demography were used. Delay to diagnosis was calculated from the first recorded symptom to the first AS/PA diagnosis. Delays to treatment and biotherapy were measured between the first AS/PA diagnosis and the first treatment/biotherapy prescription. Referral rates were calculated using rheumatologist consultations recorded in the reimbursement history. A Principal Component Analysis (PCA) was performed in R with the FactoMineR package for each disease-related dataset, using 21 quantitative variables in 12 regions. Correlations were explored through eigenvalues and PCA visualization.

RESULTS: For AS, the first three axes of the PCA held 43.64%, 16.58% and 11.51% of inertia respectively (total=71.73%). For PA, the first three axes held 31.29%, 18.16% and 13.55% of inertia (63.00%). Higher referral rates were associated with higher diagnosis, treatment and biotherapy initiation rates, mostly at rheumatologists. The more numerous the patients, the longer the delays to diagnosis and the shorter the delays to treatment and biotherapy initiation. For AS, medical demography did not show any correlation on any axe. For PA, higher rheumatologist densities were associated with shorter delays. For both diseases, Centre-Val-de-Loire and Provence-Alpes-Côte-d’Azur-Corse had high referral, diagnosis, treatment and biotherapy rates. Brittany and Normandy had low rheumatologist-associated rates and longer delays.

CONCLUSIONS: Longer delays to diagnosis may cause patients to be more severe when diagnosed, which may encourage faster treatment or biotherapy initiations. These results show significant differences in healthcare access between regions which are not satisfactorily explained by medical demography.