McClure, N.

Applying the minimally important difference for health-related quality of life utility scores in the calculation of quality-adjusted life years
McClure, N.S., Paulden, M., Ohinmaa, A., Johnson, J.A.

In cost-utility analyses, the quality-adjusted life year (QALY) is calculated as the area under the curve (AUC) for the combined assessments of time and health-related quality of life (HRQL). Researchers have recommended using a regression model to calculate QALYs, adjusting for differences in baseline utility score, as well as the effects of 'regression toward the mean'. Interestingly, support for a regression model implies that small QALY changes are not meaningful, and thus are valued at zero. I propose another method in the calculation of QALYs, incorporating the minimally important difference (MID) for the HRQL utility score, which is defined as the smallest change in utility score that is considered meaningful to patients.

A simulation model was used to show how applying an MID for the HRQL utility score can be incorporated into the QALY calculation. Next, a case-study comparing enhanced care to collaborative care in the treatment of depression for patients with type 2 diabetes was used to demonstrate how the MID-QALY calculation method can be used on patient-level data. In all analyses, results of the MID-QALY calculation method were compared to a change from baseline approach, AUC, and regression method.

Preliminary results from the simulation-based study suggest that the MID-QALY calculation method performs at least as well as a change from baseline approach, whereas on average, the AUC method performed worst and the regression method performed best (i.e., lowest mean absolute error). Preliminary results from the case study suggest that adjusting for differences in baseline EQ-5D-5L utility score impacts the direction of incremental benefit (AUC = 0.03). While there was consistency in direction, the magnitude of incremental benefit varied from -0.01 to -0.03, with a MID-QALY calculation result of -0.02.

An MID for HRQL utility scores can be used to calculate QALYs from patient-level data.