Objective: To estimate the variation among rheumatologists in clinical outcomes and frequency of office visits for patients with rheumatoid arthritis (RA), after accounting for patient demographic and clinical characteristics and treatments prescribed.
Methods: Multiple regression analysis using random effects for rheumatologists and adjustments for patient characteristics and treatments received, based on data derived from a panel study of persons with RA.
Results: During the years 1984-1993, rheumatologists accounted for a moderate amount of the total variation in clinical outcomes and nearly one-third of the total variation in frequency of office visits. For example, in 1993 rheumatologist associated variation in 4 clinical outcomes ranged from 16 to 25%, while the variation in office visit frequency attributable to rheumatologists stood at 46% of the total variation. However, rheumatologist associated variation in clinical outcomes was not statistically significant in any year, while variation in office visits was highly significant in all years (p < or = 0.0001). Although there was an increase in the percentage of variation attributable to rheumatologists for all outcomes examined across the years of this study, the time trend reached statistical significance only for frequency of office visits (2.4% per year; p = 0.0135) and functional status (1.6% per year; p = 0.0034).
Conclusion: The magnitude and strength of rheumatologist associated variation in frequency of office visits, without comparable strength in the variation in clinical outcomes, may suggest inefficiencies in the use of resources for the care of persons with RA. Further work is needed to directly examine the relationship between health outcomes and resource utilization.