Practice variation in the treatment of rheumatoid arthritis among German rheumatologists

Objective. To describe practice variation in the treatment of rheumatoid ar thritis (RA) among German rheumatologists with regard to drug and non-drug therapy. Methods. We used data of 7,326 patients with RA registered in a national Ge rman rheumatological database in 1998. In the database, every patient with an inflammatory rheumatic disease seen at one of the German Collaborative A rthritis Centres is registered once a year with a standard clinical data fo rm and a patient questionnaire. We compared health care provided by 29 rheu matological outpatient units. For drug and non-drug treatment unit prescrip tion rates, ranges and outliers were calculated. Logistic regression analys is was used for case mix adjustment and for the identification of practice patterns. Results. We observed variation concerning the frequency of use of single di sease modifying antirheumatic drugs (DMARD). The median of the prescription rates in the 29 units for methotrexate (MTX) was 55% in 1998 (1(st) quarti le: 51%, 3(rd) quartile: 63%); sulfasalazine had a median of 15% (quartiles : 10%/19%), antimalarials a median of 8% (quartiles: 5%/21%). Combination D MARD therapy was used in 11% (quartiles: 6%/18%). Prescriptions of low dose steroids (less than or equal to 7.5 mg) had a median of 45% (quartiles: 35 %/55%), and nonsteroidal anti inflammatory drugs (NSAID) had a median presc ription rate of 58% (quartiles: 50%/70%). High variation was also found con cerning active physiotherapy (median: 41%; quartiles 34%/55%) and passive p hysical measures (median 14%, quartiles 9%/37%). Differences in case mix (a ge, sex, rheumatoid factor, disease duration, severity, disability) only ex plained a small proportion of the total variation. When the units were grou ped according to the frequency of prescription of DMARD combination therapy , treatment patterns could be identified. Units with higher rates of DMARD combination therapy used more drugs for the prevention and treatment of ost eoporosis, more active physiotherapy but fewer NSAID and fewer passive phys ical therapies. Conclusion. Variation in drug and non-drug treatment indicates significant differences in health care provision. Trends in the drug management of RA a re adopted differentially by the members of the rheumatology community. The large variability in non-drug therapies may, apart from differences in ava ilability, suggest a lack of agreement on therapeutic effectiveness.