PDB28 QUALITY ADJUSTED LIFE YEARS (QALYS) IN THE DIABETES PREVENTION PROGRAM (DPP) USING QUALITY OF WELL BEING (QWB) AND SF-6D INSTRUMENTS IN SAME SAMPLE: COMPLETE CASE ANALYSIS

Abstract

A within-trial 3 year DPP cost-effectiveness analysis (CEA) was published in 2003. The authors reported QALYs using utilities from the QWB for lifestyle (LS), placebo (PBO), and metformin (MET). QALY calculations were not done correctly in the DPP (year-end utilities were used as QALYs instead of average values between yearly measures). Available to them, but not used, were the more complete SF-6D utilities (calculable from their SF-36 data). The main objective of this research was to estimate incremental QALYs using both instruments on the same sample and with the corrected method of calculation, comparing them with previous results and suggesting implications for CEA. We had access to a 95% subset of the DPP data. We calculated QWB-based QALYs correctly. We then mapped the SF-36 data to the SF-6D and calculated QALYs. The corrected calculation used area under the curve, based on beginning and end of the year utilities. It required (unreported) baseline utilities values. If baseline was equal across the treatments (plausible given the large sample and randomized allocation), the baseline drops out of the QALY differences, but the corrected method still implies changes in these differences. The incremental QALYs using QWB as originally calculated were 0.072, 0.022 and 0.050 for LS versus PBO, MET versus PBO, and LS versus MET. The corrected incremental QWB-based QALYs were 0.051, 0.00004, 0.051. SF-6D–based incremental QALYs were 0.030, -0.012, 0.042. QWB and SF-6D instruments in the DPP yielded different incremental QALYs in same sample using correct QALY calculations; both were different from the published incorrect values. Small differences in QALYs have been shown previously to be important in the DPP CEA conclusions. These differences are quite significant and likely impact CEA results.