Area Under the Curve (AUC) Approach Using Last Observation Carried Forward (LOCF) vs. Mixed-Effects Model Repeated Measures (MMRM) in Analyzing Longitudinal Count Data

• Activity Number: 547
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Rakhi Kilaru*+
• Companies: PPD
• Keywords: AUC ; Longitudinal ; MMRM ; Ulcers ; Inflammatory ; Rheumatology

Abstract

A substantial percentage of the measurements of outcomes of interest are often missing in randomized clinical trials and introduces potential biases in the comparison of treatment arms. In rare circumstances, area under the curve analyses using LOCF imputation (Average Daily AUC) are used in exploratory analyses in the assessment of ulcer counts over time as a response measure particularly in rheumatology and inflammatory disorders. MMRM is a likelihood-based approach which models all actual observations jointly, with no attempt at imputation for missing values. In light of this specialized application of AUC analyses to summarize ulcer counts over time using the LOCF approach, our analyses will include evaluation of Average Daily AUC versus MMRM using simulated ulcer counts over time as a response measure with varying levels of missing data subject to LOCF imputation and summarize findings from the two approaches. Logarithmic transformation of calculated AUC values will serve as the dependent variable for the MMRM approach. Simulations will assume varying proportions of missing at random (MAR).