Rethinking survey samples in patient preference studies: a first-degree relative sample versus a general public proxy sample in rheumatoid arthritis prevention

Background: When selecting samples for patient preference studies, studies may include the general public. However, whilst cost-effective and efficient, it is unclear whether a general public sample is an appropriate proxy in some settings.

Objectives: To compare preferences for rheumatoid arthritis (RA) preventive treatments between the United Kingdom general public (n=1,003) and first-degree relatives (FDRs, n=289) of confirmed RA patients who are at an increased risk of RA.

Methods: Participants were asked to imagine they had joint pain and a 60% chance of developing RA within the next two years. Using a discrete choice experiment (DCE), participants were offered a choice between no treatment and two hypothetical treatments alternatives to reduce their risk of developing RA. Treatments were defined by six attributes (reduction in chance of developing RA, mild to serious side effects, and frequency/route of administration) with varying levels. A random parameter mixed logit model was used to calculate relative importance of attributes and maximum acceptable risk (MAR) for a given reduction in chance of RA. Time to complete the survey, comprehension questions, and self-reported ease of completion were assessed to measure attendance to the survey.

Results: The FDR sample spent more time completing the survey (29.8 vs.16.6 median minutes) and performed significantly better on the 3 comprehension questions compared to the general public (70;3%, vs 30.3%). Most respondents thought the DCE was easy/very easy to understand (75.1% vs. 71.7%); however, significantly fewer FDR respondents (43.3% vs.59.8%) found it easy to answer. The relative importance ranking of attributes was similar between samples, but weights and resulting MAR differed appreciably between groups. In both samples, effectiveness (reduced chance of RA) was the most important attribute and chance of mild side effects was least important. For the general public sample, mode of administration was more important than for the FDR sample.

Conclusions: In the context of RA prevention, the general public may be a reasonable proxy for a more at-risk sample (FDRs) for attribute ranks but not for MAR, which informs benefit-risk tradeoffs. Survey engagement of the general public differed from the FDRs, suggesting that a proxy sample for benefit-risk tradeoffs needs to be clearly justified given the limitations.

Funding: This study was part of PREFER. PREFER received funding from the IMI 2 Joint Undertaking (grant No. 115966), which receives support from the EU's Horizon 2020 research and innovation program and European Federation of Pharmaceutical Industries and Associations (EFPIA).