Vascular patients' preferences for local treatment: an application of conjoint analysis

To investigate whether and to what extent vascular patients are willing to trade expected health outcomes for improvements in non-health benefits, such as the availability of local services, in the provision of peripheral vascular surgery.

Conjoint analysis was used to elicit patient preferences via a mailed questionnaire. Respondents were asked to consider pairwise descriptions of vascular services described in terms of a number of key attributes and asked to indicate which scenario description they preferred. Multivariate regression analysis was used to estimate the relative importance of the attributes to respondents.

Of the 339 patients invited to participate, 161 (47%) returned a completed questionnaire, although 36 (11%) were found to be internally inconsistent and could not be used, leaving 125 (36%) for analysis. An unexpected problem arose with respect to the design of the conjoint analysis questionnaire. Specifically, three of the attributes in the regression model were found to be collinear. The source of this problem, how it was overcome and how it can be avoided in future studies is discussed. The results of the regression analysis indicated that the patients who responded had a preference for local treatment to the extent that they were willing to incur increased risks of perioperative mortality and amputation to receive treatment at their local hospital. These results are discussed in connection with those from a different study which elicited patient preferences for local treatment using a standard gamble technique.

Policy-makers, when considering the potential benefits of regionalising health care, should seek to elicit the preferences of those patients who will be affected by the reorganisation.