To compare three preference-based health-related quality-of-life (HRQL) measures and examine independent correlates of HRQL among overweight and obese women with urinary incontinence (UI) enrolled in a weight loss intervention trial.
Participants completed baseline questionnaires, which included the Health Utilities Index 3 (HUI3) and Medical Outcomes Study Short Form-36 (SF-36). The SF-36 was used to derive SF-6D and estimated Quality of Well-Being (eQWB) scores. Height, weight, medical history, incontinence measures, and level of physical activity also were assessed. The intraclass correlation coefficient (ICC) was computed, and differences in mean scores across HRQL measures were examined. Potential correlates of HUI3, SF-6D, and eQWB scores were evaluated using multivariable generalized linear models.
Mean ± SD scores for the HUI3, SF-6D, and eQWB were 0.81 ± 0.18, 0.75 ± 0.10, and 0.71 ± 0.06, respectively. Significant differences were observed across measures (P < 0.0001), and the overall ICC was 0.36. In multivariable analyses, BMI was negatively associated with HUI3 (P = 0.003) and eQWB (P < 0.001), and UI episode frequency was negatively associated with eQWB (P = 0.015) and SF-6D (P < 0.001).
Significant differences in mean utilities across the HUI3, SF-6D, and eQWB indicate that these measures do not assess identical dimensions of HRQL. Both BMI and UI episode frequency were related to HRQL in this cohort; however, the magnitude of the relationship depended on the preference-based measure used. These findings highlight the need to consider the method used to generate HRQL values for calculating quality-adjusted life-years in cost-utility analyses, since choice of method may have a substantial impact on the outcome of the analysis.