In Multi-Criteria Decision Analysis-based decision support for person-centred care, the person's quantitative preferences (as criterion weightings) are combined with quantified evidence and expert assessments (as option performance ratings on all criteria) to produce a personalised quantified opinion (as a set of expected value option scores). In our current decision support tools, we use the best available (central point) estimates for option performance ratings. The uncertainty surrounding the performance rating estimates, routinely reported by researchers as intervals around the means, are ignored. While defensible, this paper responds to questioning of this disregard. Apart from the inappropriate 'inverse variance' method, we find no attempt to integrate parameter uncertainty into decision analyses, simply an emphasis on reporting it fully, leaving decision makers unsupported in the burden of dealing with the separated outputs - e.g. Means and Credible Intervals. The paper suggests that uncertainty can be brought within Multi-Criteria Decision Analysis-based decision support by treating the means and uncertainties of all outcomes and process considerations as separate criteria, having them traded-off in an individually preference-sensitive manner at the point of decision. An empirical proof of method via an online example on bone health medications is provided, involving six options, two considerations and four criteria.
Keywords: Multi-Criteria Decision Analysis; decision support; preferences; uncertainty.