Comparability of patient-reported health status: multicountry analysis of EQ-5D responses in patients with type 2 diabetes.
Salomon JA., Patel A., Neal B., Glasziou P., Grobbee DE., Chalmers J., Clarke PM.
BACKGROUND: Valid and comparable measures of health outcomes are needed for clinical trials, studies on quality of healthcare, and population health monitoring. OBJECTIVE: To examine comparability of patient-reported health status across populations. RESEARCH DESIGN: Logistic regression analysis of health status across regions, controlling for demographics, risk factors, and clinical event history. SETTING: Multicenter clinical trial in 20 countries, grouped into 3 regions defined by geography and levels of economic development (Asia, Established Market Economies, Eastern Europe). SUBJECTS: 11,140 people with type 2 diabetes. MEASURES: Patient-reported health status in 5 domains (mobility, self-care, usual activities, pain/discomfort, anxiety/depression) using EQ-5D. RESULTS: Examining unadjusted response probabilities, patients in Eastern Europe were consistently more likely than patients in other regions to report problems in all domains. Compared to Asia, probabilities of reporting problems at baseline in Eastern Europe were more than 3 times higher for mobility or usual activities, and more than 6 times higher for self-care. Patients in Asia were less likely than patients in Established Market Economies to report problems in all domains except anxiety/depression. Substantial regional reporting differences persisted after controlling for demographics, common risk factors, and history of major disease complications. Compared to Established Market Economies, adjusted odds ratios for reporting problems in at least 1 health domain were 1.79 (1.55 to 2.06) in Eastern Europe and 0.76 (0.67 to 0.86) in Asia. CONCLUSIONS: There is substantial variation across regions in reporting on functional health problems, which cannot be explained by differences in demographic variables, clinical risk factors, or rates of complications. This suggests that commonly used health status instruments may have important problems in comparability across settings.