Table 5.

Summary of the methods used to assess patient preferences.

Method	Description of method	Strength	Weakness and limitations	References
Ranking and rating	Direct scaling methods asking patient to rank or score attributes that distinguish treatment	• Feasibility of their implementation • Low cognitive burden	• Lack of direct explicit trade-offs between benefits and harms	(19, 43, 47)
Visual analogue scale	Raking method: Assign preference for a health state on a line anchored by perfect health and death	• Collection and valuation of several outcomes	• Use of ill-defined anchors which limit comparison between individuals • More valuable when used in combination with other methods	(16, 47)
Standard gamble	Choose either a gamble between perfect health and death or a certain but intermediate health state	• Estimation of quality-adjusted life-years (QALYs)	• Cognitively burdensome if several scenarios • Possibility of overestimation of patient's aversion to risk	(1, 19, 34, 41, 47)
Time trade-off	Choose either an intermediate health state for time t or perfect health for time x < t	• Estimation of QALYs • Assessment of risk preferences and minimum benefit	• Emotionally challenging for parents to consider their children having less years of life	(1, 19, 41, 47)
Discrete choice experiment	Choose between scenarios that describe a health state by different levels of attributes of that health state	• Valuation of hypothetical scenarios • Translation of preferences into utilities • Assessment of multiple attributes simultaneously • Ability to inform willingness to pay	• Require large sample sizes to produce statistically significant utilities	(1, 2, 13, 14, 21, 28, 30–32, 34, 40, 41, 43, 47, 48)
Best-worst scaling	Direct valuation of best and worst scenario or profile	• Less cognitively taxing on its participants	• Does not allow for “indifferent” choice	(13, 29, 31, 37, 43)
Multicriteria decision analysis	Direct consideration of an explicit set of criteria and their relative importance	• Decision based on several features simultaneously • Break down complex situations where many variables play a role in the decision-making process	• Potential cognitive burden • Requirement of a preliminary robust model	(2, 8, 27, 33, 42)
Analytic hierarchy process	Type of MCDA: Choose between multiple attributes or criteria in a pairwise compared manner	• Simplify complex decision making with multiple criteria, by reducing the trade-offs made at one time by presenting the choice as a pairwise comparison	• Valuation of limited number of outcomes • Potentially oversimplifying criteria and overlapping endpoints in complex pairwise hierarchies	(7, 8)
Swing weighting	Type of MCDA: First, patients rank the scale swings and afterwards allocate points that indicate the trade-off ratios	• Does not require econometric modeling: preferences are assumed to be directly captured with the elicitation task	• Potential cognitive burden requiring direct numerical assessment	(40)