Background/Purpose: Tailoring therapy to individual patients, for instance by predicting its effectiveness has become increasingly important especially with chronic use of expensive biological DMARDs.To evaluate the cost effectiveness of a personalized approach to biological treatment of rheumatoid arthritis RA) patients indicated for rituximab (RTX) using interferon (IFN) type I response genes as prediction tool.

Methods: 26 consecutive patients diagnosed according to 1987 ACR criteria from the VU University Medical Center and READE starting RTX were followed. Baseline IFN type I response genes expression level was used as a biomarker for response to RTX using an optimal cut-off (1). The Positive Predictive Values (PPV) for good and moderate EULAR response were used to simulate the response using a personalized approach, assuming that the response in patients not selected for the RTX therapy was equal to the average response to treatment (biological) in the population. Using a patient level Markov model, response was extrapolated to disease activity (DAS28), functional disability (HAQ), direct and productivity costs and Quality Adjusted Life Years (QALYs) over 5 years using external data. The Incremental Cost Effectiveness Ratio (ICER) comparing the personalized approach to (observed) usual care was calculated from a societal and health care perspective. Although RTX treatment is less costly than other biologicals no differences in drug cost were assumed and testing costs were assumed € 50 per patient. All the analyses were performed probabilistically and a sensitivity analysis was also performed.

Results: 50% were moderate responders and 50% were non-responders (no good responders). Twenty-nine percent decreased more than 1.2 points on the DAS28. Using the selected IFN cut-off, 35% of patients were selected for treatment and the PPV (for moderate response) was 100%. On average € 236 (2.5 – 97.5 percentile range -6,083 to 5,557) were saved and 0.03 QALYs (-0.08 to 0.15) were gained leading to an ICER of €-8099 per QALY gained. In 37% of simulations the personalized approach was found to be cheaper and more effective, in 32% more expensive and more effective, in 16% less expensive and less effective and in 15% more expensive and less effective. Due to the limited data on prediction using IFN so far and the resulting uncertainty the probability of costs-effectiveness did not increase to more than 70%. Using a healthcare perspective the results were less positive with a mean ICER of € 1920 per QALY gained (represents extra costs and QALY, which is within the often mentioned limit of € 20,000/QALY). Decreasing the effectiveness of alternative biological therapy with 10% resulted in a lower probability of cost-effectiveness but average results still indicated cost saving and QALY gain.

Conclusion: Although results are still uncertain due to the limited clinical data, the use of the IFN type I response genes for personalized biologic treatment seems very promising and should continue to be studied in patients indicated for RTX treatment to obtain more precise estimates of cost-effectiveness results.

References: 1.Raterman et al., Arth. Res. Ther. 2011

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/an-early-economic-evaluation-of-personalized-treatment-with-rituximab-by-prediction-of-effectiveness-using-the-interferon-type-i-signature-in-rheumatoid-arthritis/