The impact of structuring information in a patient decision aid

Abstract

Two computer based studies examined the correspondence between decision outcomes from two different types of decision support and the decision derived from a ‘reasoned choice’ model. In Study 1 (N = 72), the impact of presentation (alternative-tree or linear) on participants’ decision, satisfaction with decision and their expectancy-vale (EV) ratings was examined for a hypothetical scenario about a decision to have a heart operation or not. In the alternative-tree format, attributes associated with each alternative were set out either as ‘pros’ or ‘cons’ and participants could click on these headings to receive more information. In the ‘linear’ format participants read through the same information in a similar fashion to a leaflet. There was significantly greater correlation (p <0.05) between the decision and participants’ summed EV ratings in the alternative-tree condition (r = 0.63, p <0.01) than in the linear condition (r = 0.29, ns). Study 2 (N = 216) replicated these findings comparing the alternative-tree (r = 0.43, p <0.05) to an attribute-tree format (r = 0.19, ns) thus controlling for the interactive searching for information on a computer. Furthermore, process tracing measures reflected the use of more complex, systematic processing of information in the alternative-tree condition, consistent with an increased correspondence between participants’ decision and values. Health professionals need to consider more closely the structuring of information about treatment options to allow patients to process information and make decisions in line with their expectations and values.