Tests of preferential independence in the QALY model

The author presents evidence for the descriptive adequacy of the quality-adjusted life years (QALY) model as applied to health profiles. One important assumption of the model is preferential independence: if two profiles have the same health state during year X, then preference between them does not switch if the level of health changes during year X. In experiment 1, 30-year health profiles were used to perform 27 empirical tests of independence with 98 subjects. Independence was reliably satisfied in all 27 tests. In experiment 2, 15 additional tests were conducted. These tests had been specifically designed to be more sensitive to independence violations, but independence was still mostly satisfied. In both experiments, the conclusions about independence hold regardless of what discount rate is used. These results act as a "lower bound" on the validity of the QALY model for health profiles.