This study provides health economists, researchers and policy-makers with a reference for health utilities of various chronic conditions, different age groups, gender and comorbidities.

long-term residents of hospitals or residential care facilities. The sub-sample used in this analysis consisted of 38 151 respondents (52.4% male) between the ages of 20 and 64 y, excluding pregnant women. Health Utilities Index-Mark III (HUI3) scores were used to de®ne normal weight (body mass index (BMI) 19 ± 24.9 kgam 2 ), overweight (BMI 25 ± 29.9 kgam 2 ), obese (BMI 30 ± 34.9 kgam 2 ), and morbidly obese (BMI ! 35 kgam 2 ) individuals. HUI3 scores were age-and gender-standardized. RESULTS: The overall prevalence of obesity (BMI ! 30 kgam 2 ) in this Canadian population was 13.3%. The average difference in HUI3 scores between normal weight and morbidly obese respondents was 0.04 (P`0.001). Statistically signi®cant (P`0.05) differences across BMI categories were found in each of the eight component attributes of the HUI3. The attributes with the most substantial difference between normal and obese patients were cognition, mobility and pain. All demonstrated a ! 2-fold increase in the proportion of individuals in poorer classi®cations of health when normal weight respondents were compared with the morbidly obese. The magnitude of the decrement in utility ratings associated with obesity was comparable with other chronic non-cardiovascular conditions such as migraine or colitis. CONCLUSION:The results indicate that changes in self-rated health status appear to be due to signi®cant changes across several relevant domain attributes. Obesity has a signi®cant impact on both quality of life and health.

No abstract