Skip to main content
. Author manuscript; available in PMC: 2023 Nov 20.
Published in final edited form as: Ann Rheum Dis. 2020 Oct 8;80(3):286–292. doi: 10.1136/annrheumdis-2020-218282

Figure 2. Predicted burden of multimorbidity in incident rheumatoid arthritis (RA) compared to non-RA patients after diagnosis.

Figure 2.

Predicted burden of chronic conditions from generalized estimating equation models comparing RA and non-RA patients. Shown are the number of chronic conditions in RA and non-RA as well as the interaction term between RA status and follow-up time (years). Panel A, primary analytic approach requiring 1 year in the dataset without RA diagnostic codes or medications and evaluating 44 chronic conditions. Panel B, similar analytic approach evaluating 36 chronic conditions after removing those known to be associated with RA or could be misclassified as RA. Panel C, restricting the population to individuals with at least 3 years of data prior to index date and beginning follow-up at 2 years before the index date (date fulfilling RA algorithm). Confidence intervals are shown but fall within the width of the predicted lines.