A person trade-off study to estimate age-related weights for health gains in economic evaluation.

Background: An increasing body of literature is exploring whether the age of the recipient of health care should be a criterion in how health care resources are allocated. The existing literature is constrained both by the relatively small number of age comparison groups within preference-elicitation studies, and by a paucity of methodological robustness tests for order and framing effects and the reliability and transitivity of preferences that would strengthen confidence in the results. This paper reports the results of a study aimed at estimating granulated age-related weights for health gains across the age spectrum that can potentially inform health care decision-making.
Methods: A sample of 2,500 participants recruited from the health care consumer panels of a social research company completed a person trade-off (or 'matching') study designed to estimate age-related weights for 5- and 10-year life extensions. The results are presented in terms of matrices for alternative age comparisons across the age spectrum.
Results: The results revealed a general, although not invariable, tendency to give more weight to health gains, expressed in terms of life extensions, in younger age groups. In over 85% of age comparisons, the person trade-off exercises revealed a preference for life extensions by the younger of the two age groups that were compared. This pattern held regardless of the method of aggregating responses across study participants. Moreover, the relative weight placed on life extensions by the younger of the two age groups was generally, although not invariably, found to increase as the age difference between the comparator age groups increased. Further analyses revealed that the highest mean relative weight placed on life extensions was estimated for 30-year-olds when the ratio of means method was used to aggregate person trade-off responses across study participants. The highest mean relative weight placed on life extensions was estimated for 10-year-olds for 5-year life extensions and for 30-year-olds for 10-year life extensions, when the median of individual ratios method was used to aggregate person trade-off responses across study participants. Methodological tests framed around alternative referents in the person trade-off questions and the stability of preferences had no discernible effects on the study results.
Conclusion: This study has produced new evidence on age-related weights for health gains that can potentially inform health care decision-making.