Using the threshold technique to elicit patient preferences

Abstract

Patient preference information (PPI) is a topic of interest to regulators and industry. One of many known methods for eliciting PPI is the threshold technique (TT). However, empirical studies of the TT differ from each other in many ways and no effort to date has been made to summarize them or the evidence regarding the performance of the method. We sought to describe the TT and summarize the empirical applications of the method. Forty-three studies were reviewed. Most studies estimated the minimum level of benefit required to make a treatment worthwhile, and over half estimated the maximum level of risk patients would accept to achieve a treatment benefit. The evidence demonstrates that the TT can be used to elicit multiple types of thresholds and can be used to explore preference heterogeneity and preference non-linearity. Some evidence suggests that the method may be sensitive to anchoring and shift-framing effects; however, no evidence suggests that the method is more or less sensitive to these potential biases than other stated-preference methods. The TT may be a viable method for eliciting PPI to support regulatory decision-making; however, additional understanding of the performance of this method may be needed. Future research should focus on TT performance compared with other stated-preference methods, the extent to which results predict patient choice, and the ability of the TT to inform individual treatment decisions at the point of healthcare delivery.