To the Editor:
I have read the article entitled “Identifying Rheumatoid Arthritis Cases within the Quebec Health Administrative Database” by Zeinab F. Slim, Cristiano Soares de Moura, Sasha Bernatsky, and Elham Rahme (The Journal of Rheumatology, December 2019)1 with interest and I would like to add some comments.
In this article, rheumatoid arthritis (RA) prevalence based on a self-reported survey (CARTaGENE) was 2.9% (564 out of 19,704 patients) with a 95% CI of 2.6–3.1 before misclassification adjustment. This estimate would constitute the highest RA prevalence reported in the world since 1980, followed by Cuba (2.7%) in 20002, Finland in 19893, and Australia in the National Health Survey (1.9%) in 2014–20154. Despite a large sample size that closely resembled the general population in Canada5, it is much higher than the database-estimated prevalence by Lacaille, et al (0.8%)6, Widdifield, et al (0.9%)7, or Jean, et al (0.8%)8.
This estimate is highly biased for the reason that self-reported diagnosis of RA is not sufficient for the ascertainment of RA cases without clinical verification. It is an unreliable method for research or clinical purposes in various populations, because of the low percentage (21–34%) of people who will have confirmed RA9,10. Our concerns are about the validity of self-reported data in the CARTaGENE study to estimate the RA prevalence in Quebec (Canada), without clinical verification by a doctor or by using any RA classification criteria.
