Patient preference plays a role in clinical practice and is at the heart of patient-centered research; thus, ignoring its impact on outcomes could result in missing important determinants of study outcomes. Furthermore, patients may have a different psychological response to a treatment they deem more favorable. Although the traditional clinical trial setting in which individuals or clusters are randomly allocated to 1 of multiple treatment arms is the gold standard for assessing a treatment effect (the average effect a particular treatment will have in a specified population), this design ignores the role that patient preferences for a particular treatment may have on study outcomes; they are not estimable in this design. The 2-stage trial design enables researchers to disentangle the treatment effect from those effects resulting from choosing a treatment. Although the use of the 2-stage trial design is becoming more prevalent, especially as the emphasis on using decision aids continues to grow, there is still a large gap in the methods available to design and ultimately analyze this trial design. Often, the primary outcome of interest is not measured on a continuous scale; typically we see binary outcomes (eg, whether patients are satisfied with their treatment) or count data (eg, number of days alcohol free/drug free). In addition, we are often interested in accounting for important covariates in the design of the study (ie, stratification variables, such as age, sex, and/or type of insurance coverage) that may have an impact on the outcome of interest or may influence the preference rate (eg, men may have a stronger preference for a surgical intervention, whereas women may have a stronger preference for a medical intervention). No methods existed to accommodate these scenarios. A well-designed and well-powered study is essential, and the number of participants enrolled is a major determinant of the statistical power of the trial. Therefore, advancing the methodology of the 2-stage trial design and increasing its accessibility for designing such trials is extremely important.
