Marginal Analysis of Linear Model with Application to Testing for Resting Energy Expenditure Model

• Activity Number: 595
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Junyi Zhang*+ and ZiMian Wang and Zhiliang Ying
• Companies: Columbia University and St. Luke's-Roosevelt Hospital and Columbia University
• Keywords: Multiple linear regression ; resting energy expenditure ; metabolic rates ; marginal analysis ; colinearity ; Elia's coefficients

Abstract

One of the primary areas in human energy metabolism research is to explore organ specific metabolic rates under resting condition. The whole-body resting energy expenditure (REE) is totality of organs/tissues energy expenditures, which can be evaluated individually as product of each organ/tissue mass and its corresponding metabolic rate. Elia postulated the well-known set of rate coefficients through an empirical method and derived a multiple linear regression. We consider testing validity of Elia's coefficients. Due to the high correlation among organ/tissue masses, conventional approach is unstable and inefficient. We propose a marginal approach by considering one organ/tissue at a time, thereby bypassing the difficulties caused by the near-ill conditioned design matrix. We apply the proposed approach to a data set and discuss results and new findings for older and younger adults.