Measuring Patient Preferences to Inform Clinical Trial Design: An Example in Rheumatoid Arthritis.

Background: Allogeneic bone marrow transplantation (BMT) may be a curative treatment for patients with rheumatoid arthritis (RA), but it has serious risks, including death. It is uncertain whether patients would accept the risks and benefits of BMT and participate in clinical trials. We conducted a discrete choice experiment (DCE) to quantify risk tolerance and benefit-risk trade-offs to inform the design of a clinical trial for BMT.

Methods: We conducted a DCE with three attributes (three levels each): chance of stopping disease progression (50-90%), increased chance of death in year after transplant (3-15%), and chance of chronic graft-versus-host disease (cGVHD) (3-15%). An orthogonal main effects design of nine binary choice tasks were presented for two scenarios: one considering their current situation and a second scenario where the patient has failed seven anti-rheumatic drugs. Participants were recruited from the Rheum4U inflammatory arthritis registry. Choice data were analyzed using a logit model accounting for multiple responses per participant.

Results: Sixty patients participated. Most (82%) had severe disease, and the median number of anti-rheumatic drugs previously taken was 6 (range 0-18). As expected, an increased chance of stopping disease progression increases the probability of choosing BMT, while increased chance of both risks decreases the probability. Patients were willing to accept a 3% increase in risk of death or 6% increase in chance of chronic GVHD for a 10% increase in the chance of stopping disease progression. For the most clinically likely BMT risk-benefit profiles, and the likely initial target population of patients who have failed multiple biologics, between 72% and 91% of patients would choose BMT.

Conclusions: Patients with RA are willing to accept substantial risks for a chance to stop disease progression with BMT, suggesting that a pilot trial of BMT for RA could successfully recruit patients. Preference studies have an important role in informing patient-centered clinical trial planning and design.