OBJECTIVES: Cost-effectiveness models of biologic treatments for RA have been shown to be strongly affected by improved physical function, as measured by HAQ-DI. This analysis was conducted to focus on patients receiving continuous treatment with tocilizumab (TCZ) 8 mg/kg monthly and conventional anti-rheumatic medications for more than 24 weeks, while assessing HAQ-DI scores. METHODS: HAQ-DI scores in patients with moderate to severe RA were obtained from the 8-mg/kg treatment groups of three extension studies of four randomised, placebo-controlled studies. Analysis focused on observations at or after week 24 to assess the long-term effects of treatment. Week 192 was chosen as a cut-off to preserve sufficient sample size. Analysis of mean HAQ-DI scores (average at each time window) was conducted for all observations. Parallel analysis was conducted separately for patients who received previous therapy with anti-TNF agents that did or did not fail. A trendline was fitted to the data to describe the overall trend and assess its linearity. In addition, a mixed linear model of HAQ-DI as a function of time was estimated at the patient level. RESULTS: A total of 1422 patients were available at week 24. Nearly 40% (546) have been observed through week 156, and 56 have been observed through week 192. All-observed analysis showed a strong linear relationship with a slow, steady decrease in HAQ-DI scores through week 192 (week 24=1.05; week 192=0.88; slope=–0.015; R2=0.87). HAQ-DI improvement was monotonic through week 156. Stratification by previous failure of anti-TNF therapy produced similar findings. Patient-level mixed-model results were consistent, showing a significant downward slope. CONCLUSIONS: Patients with RA treated with 8 mg TCZ in combination with standard anti-rheumatic medications have steadily improving HAQ-DI scores while on treatment through week 192. Cost-effectiveness models of TCZ should include these findings to accurately gauge the utility of treatment.