Sunday, October 18, 2009
Grand Ballroom, Salons 1 & 2 (Renaissance Hollywood Hotel)
Tanya G.K. Bentley, PhD1, Mari Palta, PhD2, Adam Paulsen2, David Feeny, PhD3, Dasha Cherepanov, PhD2, Nancy Cross Dunham, PhD2, Robert M. Kaplan, PhD4 and Dennis G. Fryback, PhD5, (1)Partnership for Health Analytic Research, LLC, Beverly Hills, CA, (2)University of Wisconsin, Madison, WI, (3)Health Utilities, Inc., and Kaiser Permanente Northwest Center for Health Research, Portland, OR, (4)University of California Los Angeles, Los Angeles, CA, (5)University of Wisconsin - Madison, Madison, WI
Purpose: We assessed the relationship between health-related quality of life (HRQoL) and self-reported body mass index (BMI) using nine distinct measures of health status and HRQoL.
Methods: Data were from the National Health Measurement Study (NHMS), a cross-sectional telephone survey of 3,844 non-institutionalized U.S. adults aged 35-89 years, administering all HRQoL measures to each participant. BMI was calculated from self-reported height and weight. We estimated mean HRQoL scores using weighted least squares for normal, overweight, and obese categories of BMI (18.5-24.9 kg/m2; 25-29.9; and ≥30, respectively) with respect to age, race, education and smoking for six standard indices of HRQoL (EQ-5D; SF-6D; HUI3; HUI2; QWB-SA; and HALex) and three summary measures of health status (Theta – an item response theory-derived composite measure – and SF-36 physical and mental component scores (PCS and MCS)).
Results: Among 3,712 U.S. adults with complete BMI data, eight of out nine measures captured significantly lower unadjusted mean HRQoL (p<0.0001) with increasing BMI category, with differences between obese and normal weight 5.2 for PCS, 0.36 for Theta and ranging from 0.10 for HALex to 0.04 for SF-6D. Differences between overweight and normal individuals were 1.9 for PCS, 0.11 for Theta, and ranged from 0.0 on SF-6D to 0.03 on both QWB and HALex. There was no significant difference in mean MCS scores between BMI groups (p=0.46), except that obese women had worse MCS than normal weight women (difference=1.2, p=0.0134). While other race and gender patterns appeared different based on effect sizes, these differences were not statistically significant with a few exceptions. There was an indication that blacks were less affected by overweight than other races (p=0.034), particularly on the HUI2, HUI3, and EQ-5D (normal-overweight differences of 0.04 on each). Adjustments did not change results.
Conclusions: Among U.S. adults, six commonly-used HRQoL indices and two summary measures – excluding the MCS – of health status detected significantly lower HRQoL associated with obesity and overweight as compared with normal BMI, although preference score indexes did differ in their sensitivity to BMI. The association appeared to be primarily driven by physical but not mental health status for all groups except for women.
Candidate for the Lee B. Lusted Student Prize Competition
