Appendix A: Conjoint Survey Profiles
Appendix B: Reference Guide
Appendix C: Determinination of Attributes and Levels
The experimental design of this study employs a series of tradeoff questions to generate preference information. The attributes and levels included in the tradeoff questions are determined through a three-step process, including first a review of the literature on risks and benefits of the urethral treatments; second, discussions with and evaluations of the initial attributes with clinicians who provide urethral treatments and a few patients; and third, a qualitative interview process with patients with urethral stricture disease.
During the qualitative interviews, we asked patients what the most important factors to them are when they are deciding on a treatment. Once the results from the qualitative interviews begin to plateau and there was a clear pattern of important factors, we evaluated the top five factors highlighted by patients as the attributes for this study. We included the copayment amount in order to analyze patients’ willingness to pay. The specific attributes and the number of attributes are selected to generate as much preference information regarding open reconstruction and endoscopic incision procedures, while limiting the cognitive burden placed on patients.
The levels within each attributes were selected to reflect the realistic outcome of the two treatment options this study examines. Open reconstruction procedure has an 85% success rate, is more likely to require zero future procedures, 3 weeks of catheter, and 6 weeks recovery time. On the other hand, endoscopic incision procedure has a 25% success rate and is more likely to require more future procedures, catheter for 1 week or less, and 2 weeks recovery time. Clinicians also discussed with patients the most desirable and least desirable outcomes within each factor they are concerned with. The desirable outcomes highlighted by patients were then used to supplement the determination of levels within each attributes. Taking the realistic outcomes and patients’ concerns, we set the levels at values which allow us to statistically analyze their importance while providing substantively and clinically meaningful results.
The final attributes and levels were then used to formulate survey profiles. We piloted the survey with eight patients and had no issues. The survey was then finalized and administered to the patient sample in our study.