The initial validation of a Markov model for the economic evaluation of (new) treatments for rheumatoid arthritis.

OBJECTIVE: Markov models are increasingly used in economic evaluations of (new) treatments for chronic diseases. In this study we propose a Markov model with health states defined by the disease activity score (DAS) to be used to extrapolate efficacy data from short-term clinical trials in rheumatoid arthritis to longer term cost-effectiveness results. Moreover, we perform an initial validation of this model. METHODS: To test the validity of the model, the expected disease course (according to the model) was first compared with the observed disease course in an inception cohort of newly diagnosed rheumatoid arthritis patients. Then the relationship of the health states with utility and costs was investigated. Finally, costs and QALYs were calculated for usual care of patients in the first 5 years of their disease using the model and compared with the literature. RESULTS: The model seemed to be able to extrapolate 1-year efficacy data as seen by a comparable distribution over the Markov states between the model results and the observational data. The health states had a significant relationship with costs and utility, and population characteristics had only a moderate effect on the cost and utility values of the Markov states. The distribution over the Markov states resulted in 3.266 expected QALYs per patient over 5 years. The expected medical and total costs per patient over 5 years were 6754 euro (1997 values) and 12,641 euro, respectively, for standard rheumatoid arthritis care in The Netherlands. CONCLUSION: The developed Markov model seems a valid model for use in economic evaluations in rheumatoid arthritis. The model produced similar utilities, but lower total costs, to those in previously published studies. Although the steps to develop and validate this Markov model were applied in the context of rheumatoid arthritis, they can be generalised to other chronic diseases.

The relationship between disease activity and radiologic progression in patients with rheumatoid arthritis: a longitudinal analysis.

OBJECTIVE: Radiologic progression in rheumatoid arthritis (RA) is considered the consequence of persistent inflammatory activity. To determine whether a change in disease activity is related to a change in radiologic progression in individual patients, we investigated the longitudinal relationship between inflammatory disease activity and subsequent radiologic progression. METHODS: The databases of the University Medical Center Nijmegen (UMCN) cohort and the Maastricht Combination Therapy in RA (COBRA) followup study cohort were analyzed. The UMCN cohort included 185 patients with early RA who were followed up for up to 9 years. Patients were assessed every 3 months for disease activity and every 3 years for radiologic damage. The COBRA cohort included 152 patients with early RA who were followed up for up to 6 years. Patients were assessed at least every year for disease activity and every 12 months for radiologic damage. Disease activity was assessed with the Disease Activity Score (DAS) (original DAS in the UMCN cohort, DAS28 in the COBRA cohort). Radiologic damage was measured by the Sharp/van der Heijde score in both cohorts. Data were analyzed with longitudinal regression analysis (generalized estimating equations [GEE]), using autoregression for longitudinal associations and radiologic damage as the dependent variable. Time, time(2) baseline predictors for radiologic progression and their interactions with time, as well as DAS/DAS28 (actual values or interval means and interval SDs of the means) were subsequently modeled as explanatory variables. RESULTS: Data analyzed by GEE showed a decrease in radiologic progression over time (regression coefficient for time(2) -1.0 [95% confidence interval -1.4, -0.6] in the UMCN cohort and -0.4 [95% confidence interval -0.8, 0.0] in the COBRA cohort). After adjustment for time effects and baseline predictors of radiologic progression and their interactions with time, a positive longitudinal relationship was indicated by autoregressive GEE between the mean interval DAS and radiologic progression in the UMCN cohort (regression coefficient 5.4 [95% confidence interval 2.1, 8.6]), and between the DAS28 and radiologic progression in the COBRA cohort (regression coefficient 1.4 [95% confidence interval 0.8, 2.0]). In the UMCN cohort, the SD of the mean interval DAS was independently longitudinally related to the radiologic progression over the same periods (regression coefficient 20.2 [95% confidence interval 7.2, 33.3]). In both cohorts, the longitudinal relationships between (fluctuations in) disease activity and radiologic progression were found selectively in rheumatoid factor (RF)-positive patients. CONCLUSION: Radiologic progression is not linear in individual patients. Fluctuations in disease activity are directly related to changes in radiologic progression, which supports the hypothesis that disease activity causes radiologic damage. This relationship might only exist in RF-positive patients.