ENVISIONING A REAL OR IMAGINARY DEPENDENT: IMPACT ON UTILITY ASSESSMENT

Purpose:

Individuals often have to make proxy decisions on behalf of an elderly or child dependent. We examined whether utilities assessed on behalf of dependents differed by whether the dependent was real or imagined.

Method(s):

Utility data were obtained from subjects in the Chicago and Indianapolis metropolitan areas for a study intended to assess public perceptions of the utility of health states for themselves and others at different ages (child and elderly). Utility estimation was performed by two methods: standard gamble and time trade-off for four health states (diabetes, severe bilateral vision loss, severe seizure disorder, and severe mental impairment). Estimation was performed for both the subject and either a child dependent, an elderly dependent, or both. For each type of dependent, if the subject had one or more real dependents, a dependent was selected at random and the subject was instructed to estimate utility as a proxy for that dependent. If the subject had no dependent, they were asked to consider an imaginary dependent. For each dependent and each health state, utility distributions were compared for a null hypothesis of either equivalent means or medians, using Welch two sample t test and Mood’s median test. The main outcome of interest was whether utility values differed when considering an imaginary or real dependent.

Result(s):

Utilities were assessed for 165 imaginary children, 427 real children, 242 imaginary elderly, and 123 real elderly dependents. Comparing the utility distributions for each health state between an imaginary and real dependent yielded no statistically significant differences. This was consistent for both null hypotheses.

Conclusion(s):

Preferences for proxy decision making quantified by utilities show no difference between subjects having an imaginary or real dependent for each of the four health states of diabetes, severe bilateral vision loss, severe seizure disorder, and severe mental impairment. Because utility assessment for policy decisions are often done on the general public, it is important that utilities (on average) do not appear to differ between those with real versus imagined dependents.