Correcting for Measurement Error in Self-Reported Dietary Data from a Longitudinal Lifestyle Intervention Trial Using an External Validation Study

• Activity Number: 635
• Type: Invited
• Date/Time: Thursday, August 4, 2016 : 8:30 AM to 10:20 AM
• Sponsor: Health Policy Statistics Section
• Author(s): Juned Siddique* and Laurence Freedman and Raymond Carroll and Trivellore Raghunathan and Elizabeth Stuart
• Companies: Northwestern University and Gertner Institute for Epidemiology and Health Policy Research and Texas A&M University and University of Michigan and Johns Hopkins Bloomberg School of Public Health
• Keywords: measurement error ; sensitivity analyses ; pattern-mixture model ; missing data ; diet

Abstract

Measurement error is considered to be an inevitable condition associated with self-reported diet. In lifestyle intervention trials, where the goal is to change a participant's weight or modify their eating behavior, self-reported diet is typically an outcome variable that is measured repeatedly throughout the intervention. Urinary recovery biomarkers for specific dietary components offer the opportunity to estimate and correct for measurement error. However, collecting 24-hour urine samples and obtaining these biomarkers is beyond the capacity of most intervention studies due to the cost and intensive nature of the protocols. We propose a statistical framework for correcting for measurement error in self-reported dietary data from longitudinal lifestyle intervention trials by combining the lifestyle intervention data with auxiliary data from biomarker validation studies where both self-reported and recovery biomarkers of dietary intake are available. Our approach facilitates the use of sensitivity analyses to address the influence of unverifiable assumptions regarding the measurement error process such as measurement error that is differential with respect to treatment and/or time.