Boeri M, Hauber B. Threshold technique: an important tool to quantify and incorporate preferences in benefit-risk analysis. Presented at the 2019 PSI Conference; June 3, 2019. London, UK.


Throughout the development, regulation, sale, and utilization of any pharmaceutical and medical product and device there is a need to evaluate the balance between benefit and risks. Recent initiatives by regulators, as well as producers, have resulted in an increase in demand for patient-preference studies to complement data regarding the clinical effectiveness and harms associated with the treatment with the willingness to accept treatment-related risks in exchange for treatment benefits. In response to increasing demand for benefit–risk evaluations, researchers have applied a variety of existing preference methods for quantifying the tradeoffs decision makers are willing to make among expected benefits and risks. The methods used to elicit benefit–risk preferences have evolved from different theoretical backgrounds and have been recently catalogued by the IMI-PREFER consortium into two broad categories: preference exploration (qualitative) methods, and reference elicitation (quantitative) methods. The former category comprises individual techniques, group techniques, and individual/group techniques; the latter includes discrete choice techniques, threshold techniques, rating techniques, and tanking techniques. The discrete-choice experiment and swing weighting have been used in pilot projects sponsored by the US Food and Drug Administration and the European Medicines Agency respectively. The threshold technique has also been employed recently in empirical studies as a method to quantify benefit-risk preferences. This methodology was first applied in medical decision making at the end of the last century and has been used in numerous empirical applications since then. This presentation aims to discuss a case study in which a threshold technique was employed to evaluate risk and benefit tradeoffs and compare the methodology to others, such as discrete choice experiment and swing weighting.

Share on: