Person Trade-Off

Elicitation Method

A typical PTO elicitation asks respondents to imagine that they are a decision maker faced with having to choose between two equally expensive healthcare treatment programs that improve quality of life or save lives for varying groups of patients. There is only enough money to fund one of the two mutually exclusive programs. Respondents must decide which program they would fund. Fixing the number of patients in one of the programs, respondents are asked how many patients would need to be treated to make them indifferent between the two programs. For example, Program A might extend the life of 100 healthy individuals for 1 year. Program B might cure 100 individuals of a chronic health condition. Many people may choose to fund Program A because of the imperative to save lives. If this is the case, respondents are then asked how many patients must be cured in Program B for it to be equally good as Program A. For example, respondents may give a median value of 1,000. Thus, 1,000 individuals would need to be cured of the chronic health condition to make the program equally good as another program that saves the lives of 100 healthy people.

Computing Preference Weights

Similar to individual utilities, preference weights can be computed from PTO responses on a 0-to-1 scale, where 0 is equal to death and 1 is perfect health. If Ai is the baseline number of individuals treated in Program A and Bi is the number of individuals who must be treated in Program B for it to be equally good as Program A, then a preference weight for Program B (WB) can be computed as

Thus, in the example above, the preference weight for curing the chronic health condition is 0.9 (1 − 100/1000). The particular equation used to derive a preference weight depends on the elicitation and the baseline used for comparison. For example, some elicitations compare curing one chronic (or acute) health condition versus another. To compute weights directly on a 0-to-1 scale, a comparison must be made with saving a life or preventing the onset of a condition. Weights can also be computed indirectly by “chaining” a series of elicitations. However, this approach has not been tested empirically and may introduce new sources of biases into the estimates.

...