The additive utility assumption of the QALY model revisited

Quality-adjusted life years are valid representations of the preferences of individuals for health outcomes only under a set of restrictive assumptions. One of the key assumptions is additive utility independence (AUI). Recently, Bleichrodt and Filko [Bleichrodt and Filko, 2008. Journal of Health Economics 27 (5), 1237-1249] presented a new test for AUI, the test for generalized marginality (GM). Based on a student survey showing that violations observed at an individual level cancel out at the group level they concluded that use of the QALY model for economic evaluations in health care is supported. In this comment we argue that this conclusion is not warranted for 2 independent reasons: (i) the GM test is not sufficient to claim AUI both at an individual and group (i.e., aggregated) level and (ii) the student survey is not appropriate to make generalized statements about preferences at the population level.