• Sessions
• Presentations
• Posters
• Rooms

Presentation

Session: Comparing Utility Measures
Room: Congress Hall B
Time: Thu 10:15-11:30

Presenter: Nick Kontodimopoulos (Hellenic Open University. Faculty of Social Sciences)

Abstract

Background: In diabetes mellitus (DM), health-related quality of life (HRQOL) is not typically attributed to a single factor, such as the disease itself, but rather to the combined effect of multiple confounding variables including comorbid conditions, macro- and microvascular complications and obviously demographic, anthropometric, clinical and treatment satisfaction variables. Despite many studies having extensively addressed diabetes HRQOL with a number of generic and diabetes-specific instruments, none have actually attempted to compare the sensitivity of such instruments to specific diabetic complications, and furthermore to explain why potential disparities in sensitivity may exist. This could be particularly interesting in preference-based utility instruments in which overall HRQOL is encompassed in a single index.

Objectives: To examine the hypothesis that the 15D, with more dimensions and a richer descriptive system, compared to the EQ-5D and SF-6D, is more sensitive to a common and highly prevalent complication of diabetes, namely coronary heart disease (CHD) and, if confirmed, to investigate if this might be explained at the dimensional level.

Methods: 319 Type II DM patients (62.4% female, mean age 65.3) were surveyed with the 15D, EQ-5D, SF-36 and the Diabetes Treatment Satisfaction Questionnaire (DTSQ). Independent variables including gender, age, years with diabetes, BMI, blood glucose, treatment satisfaction, and existence of hypertension, hyperlipidaemia, COPD, arthropathy, vascular disease, diabetic foot, retinopathy, neuropathy and nephropathy were introduced in univariate regression models to identify significant predictors of 15D, EQ-5D and SF-6D utilities.

Results: DM patients with CHD (N=107, 61.7% female, mean age 71.0) reported lower utilities than DM patients without: EQ-5D, 0.623 vs. 0.756; SF-6D, 0.708 vs. 0.775; 15D, 0.740 vs. 0.841, (P<0.001 throughout). The regression models explained 52.2%, 39.6% and 28.1% of 15D, EQ-5D and SF-6D variance respectively. Five common significant predictors (gender, age, treatment satisfaction, arthropathy and diabetic foot), and five others significant in at least one model were identified and controlled for with ANCOVA. The respective results (CHD, yes vs. no) were: EQ-5D, 0.702 vs. 0.718, P=0.589; SF-6D, 0.737 vs. 0.761, P=0.084 and 15D, 0.787 vs. 0.817, P<0.05. With the 15D, 297 distinct health states were reported compared to 48 and 125 with EQ-5D and SF-6D respectively. CHD diabetics had significantly lower scores on all 15D dimensions except for mental function, and after controlling for the same covariates, these differences remained statistically significant in six dimensions: moving, hearing, breathing, sleeping, distress and sexual activity, and borderline significant in elimination.

Conclusions: Initially, EQ-5D, SF-6D and 15D utilities discriminated well between DM patients with and without CHD, but after correcting for the confounding effect of common diabetes variables, only the 15D produced statistically significant utility differences between the two groups. Many of the 15D dimensions were sensitive enough to capture HRQOL differences attributed only to CHD. The rich 15D descriptive system resulted in a distinct health state for almost each patient, implying powerful discriminative ability and sensitivity. This is evidence in favour of the 15D in diabetics, but also emphasizes the need for further testing in other DM complications and in more diverse patient samples.

Key Terms

Utilities, 15D, EQ-5D, SF-6D, sensitivity, diabetes, coronary heart disease

Authors:

Nick Kontodimopoulos (Hellenic Open University, Patras, Greece. Faculty of Social Sciences) , Zina Chadjiapostolou (“Mamatseio” General Hospital, Kozani, Greece) , Eleni Arvanitaki (“Ag. Varvara” Health Centre, Heraklion, Greece) and Dimitris Niakas (Hellenic Open University, Patras, Greece. Faculty of Social Sciences)

Search with Google