A mixed methods investigation of methods of valuing health : are preferences over health states matters of taste, complete, and informed?

• Main Author: Karimi, Milad
• Other Authors: Brazier, John ; Paisley, Suzy
• Published: University of Sheffield 2015
• Subjects: 610
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.682292

Health state values are elicited using choice-based methods that are based on several assumptions. Three assumptions were investigated in this thesis, namely that preferences are: matters of taste, complete and articulate, and informed. Violations of these assumptions threaten the validity of choice-based methods. The aim of this thesis is to investigate the appropriateness of these three assumptions. A sequential mixed methods design with three studies was used. Qualitative interviews with a think-aloud protocol were used to investigate whether preferences are matters of taste. A mixed methods study was conducted to test completeness by investigating the effect of reflection and deliberation. A quantitative study tested whether preferences are informed. Preferences over health are not matters of taste but depend on beliefs about how ill health affects an individual on domains such as happiness, independence, and personal relationships. Preferences were not shown to be incomplete because reflection and deliberation did not change mean health state values. Although individuals are uncertain about their values, reflection and deliberation does not seem to systematically alter their preferences. Preferences may not be informed because participants’ beliefs about the consequences of ill health do not conform to the experiences of patient. The thesis contributes to knowledge of the role of beliefs in health state valuation, the effect of deliberation and reflection, and whether preferences are informed. Methodological contributions include developing a method of determining whether preferences are informed and the application of mixed methods. A key finding is that preferences over health states are not entirely informed and therefore choice-based methods of valuing health may not be entirely valid. Recommendations for further research include implementing the methods in this thesis in a larger study and testing the effect of providing information about consequences of ill health to individuals valuing health.