![[ Visit Client Website ]](images/banner.gif)
Methods: A 15-minute survey was administered via the internet in December 2007. Panel participants were recruited via telephone using random digit dial, to avoid biases typically associated with internet panels. Respondents were 18 years and older and represented the distribution of characteristics matched to the general US adult population (n=1012). Response rate was 67%. Respondents were asked to value hypothetical descriptions of influenza illness (uncomplicated influenza illness, hospitalization) and potentially rare vaccine-related adverse events (anaphylaxis, Guillain-Barré syndrome) using time trade-off (TTO) or willingness-to-pay (WTP) questions. Respondents were randomized to 8 different ages for an affected hypothetical individual: 1 year, 3 years, 8 years, 15 years, 35 years, 55 years, 70 years, 85 years (between-subjects design). Respondents were also asked to value the same health state for themselves. All other aspects of the health state description were held constant. Summary statistics for each health state and age were calculated. The Kruskal-Wallis test was used to measure differences in responses across ages of affected individuals in the hypothetical scenarios.
Results: Median values for time and willingness to pay were highest for young children. For example, median TTO values (in days) for an influenza-related hospitalization were: 1 year old – 122.0; 3 year old – 91.5; 8 year old – 30.5; 15 year old – 9.5; 35 year old – 8.0; 55 year old – 29.3; 70 year old – 14.0; 85 year old – 30.5. This pattern was generally consistent across responses and type of valuation. Responses differed significantly (p-value < 0.05) by age for the following health states: uncomplicated influenza (WTP), hospitalization (TTO, WTP), and Guillain-Barré syndrome (TTO, WTP).
Respondents were consistently willing to trade or pay more to avoid a hypothetical illness for themselves than for someone else. When valuing for themselves, elderly respondents were willing to trade or pay more than working-age adults.
Conclusions: Approaches that assume values do not differ with the age of a patient may bias analyses that use these values. If patient age is likely to affect valuations, then age should be included as an attribute in the valuation exercise.