Plausible 'Nuisance' Contributor to Decreasing Deleterious Association of Overweight and Obese on Mortality Rate Over Calendar Time

• Activity Number: 312
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Tapan Mehta* and Nicholas Pajewski and Scott Keith and Kevin Fontaine and David B. Allison+
• Companies: The University of Alabama at Birmingham and The University of Alabama at Birmingham and Thomas Jefferson University and The Johns Hopkins University School of Medicine and The University of Alabama at Birmingham
• Keywords: BMI - Body Mass Nndex ; NHANES -National Health and Nutrition Examination Surveys ; HR - hazard ratio ; MR - mortality rate

Abstract

NHANES analyses have suggested that the harmful effects of obesity may have decreased over calendar time. We hypothesized that statistical artifact introduced by the influence of a shifting BMI distribution and the application of fixed BMI categories on the association of obesity with MR may partly or entirely explain the reduction. NHANES I with linked mortality were used to optimally fit a Cox model with BMI as a continuous predictor in presence of covariates to derive a 'true' model. Coefficients estimated from the 'true' model were used to simulate mortality for NHANES III. Hence null was set such that BMI-MR relation in NHANES I and III is same. Then in NHANES I and III HRs were computed for 5 BMI categories. We tested whether the HR estimates for BMI categories between NHANES I and III were equal. Results indicated that some of the apparent BMI-MR reduction is due to the artifact.