BACKGROUND/PURPOSE: Many drug treatment options are available in the United States for people diagnosed with RA, including conventional DMARDs (cDMARDs), biologic DMARDs (bDMARDs), NSAIDs, and glucocorticoids. Treatments used for each patient over their remaining lifetime depend on effectiveness, side effects, tolerability, and financial factors. Because of the wide range of possible treatment pathways, observational data on the impact of different treatment pathways on disability from RA are limited. In this study, we developed a simple modeling approach using published data to simulate the impact on RA disability of different treatment pathways.
METHODS: We performed a targeted literature review to identify estimates of the rate of change in HAQ disability (HAQ) scores over time for patients with moderate to severe RA on different treatment regimens. We developed a spreadsheet model to estimate individual and population outcomes for two categories of drug regimens: a) bDMARDs and b) cDMARDs, NSAIDs or glucocorticoids (non-biologics). We estimated outcomes for the following treatment pathways: 1) non-biologics for remaining lifetime, 2) bDMARDs for remaining lifetime, 3) bDMARDs until age 65 years then either 100% or 80% switching to non-biologics and the rest staying on bDMARDs. We assumed that, if one drug regimen was not successful, the patient would switch to another regimen in the same category. Inputs in the model included age, HAQ score at initiation of treatment, sex, and relative risk of mortality by HAQ score. The model estimated outcomes for patients using each treatment pathway, allowing a new cohort of 1,000 patients to enter treatment each year. The primary model outcomes were: expected population distribution of mild, moderate, or severe disability; population mean HAQ score; and patient life expectancy.
RESULTS: We modeled a cohort of 1,000 people aged 56 with moderate to severe arthritis (mean HAQ 1.2) initiating treatment each year. Based on clinical trial and observational data, we assumed for the first year a decrease in the mean HAQ of ~0.2 for non-biologics and ~0.6 for biologics. For subsequent years, we assumed an annual change in HAQ of +0.002 at ages < 65 years and +0.027 at ages 65+ years for non-biologics and of -0.009 at ages < 65 years and +0.016 at ages 65+ years for bDMARDs. For each simulated treatment pathway, the figure presents an expected RA population distribution of disability severity. Mean population HAQ scores ranged from 1.10 (non-biologics) to 0.70 (bDMARDs), and individual life expectancy ranged from 22.11 years (non-biologics) to 24.11 years (bDMARDs). Results were sensitive to starting age and HAQ range.
CONCLUSIONS: Based on the assumed HAQ progression rates, more RA patients would have mild disability and fewer severe disability, the mean population HAQ scores would be lower, and life expectancy greater when bDMARD treatment regimens are included in the treatment pathway. The simulation model can be used to estimate additional population outcomes including costs and health care resource use.