Eliciting patients' preferences for the intensity of preference-sensitive health care: An integrated research strategy

Statement of Problem. Characterizing the public's preferential attitudes towards relatively more-/lessintense elective options in "preference-sensitive" health care situations is particularly relevant for the prioritization policies of publicly-funded systems. However, there are a number of issues that need to be considered when designing studies aimed at measuring these preferential attitudes in terms of strength-of-preference scores within community settings, including: whom to sample (since sub-groups might vary in their motives for preferring the same option) how many health care "scenarios" to present (since fewer are needed if preferential attitudes are consistent), and what evaluative techniques to use (since engaging techniques are costly). In three sub-studies, I explored whether strength-of-preference scores for elective health care options tend to be: (1) associated with particular "reasons" or "social influences" (2) "state-like" or "trait-like" and (3) influenced by prior use of an "engaging" evaluative technique.Methods. 202 Medicare beneficiaries considered 8 preference-sensitive scenarios involving choices between relatively more-/less-intense elective options. For each, they reported: their favored option a strength-of-preference score for that option and the reasons, sources of social endorsement, and sources of social pressure affecting their choice of that option.Results. Sub-study participation ranged from 196--199 individuals. In the first sub-study, for each of the relevant scenarios, sub-groups who strongly-/weakly favored the most-popular option did not differ in the reasons for choosing that option, and the social influences (endorsement and pressure) on their choice tended to be similar. In the second sub-study, preferential attitudes appeared to be state-like, in that there was little across-scenario consistency (using Kendall's coefficient of concordance) in the strength-of-preference scores. In the third sub-study, the engaging evaluative technique had no significant effect (using Kendall tau-b correlation coefficient) on the strength-of-preference scores elicited with the subsequent non-engaging technique.Conclusions. In community-wide surveys, researchers: (1) need not assume that, because they differ in terms of their strength-of-preference scores, individuals have different motivational patterns for favoring the same option (2) carefully select the sets of scenarios used to elicit strength-of-preference scores, since attitudes may vary depending on what's at stake and (3) might safely use a non-engaging technique when eliciting strength-of-preference scores.