Matching-Adjusted Indirect Comparison Using Bootstrapping for Comparative Effectiveness: Individual Patient Data vs. Published Aggregated Data

• Activity Number: 522
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Daniel Parks*+ and Xiwu Lin and Alan J Menius and Kwan R. Lee
• Companies: GlaxoSmithKline and GlaxoSmithKline and GlaxoSmithKline and GlaxoSmithKline
• Keywords: Indirect treatment Comparison

Abstract

The development of methods for indirect comparison of clinical outcomes is one of the most important areas in comparative effectiveness research and health technology assessment. Indirect comparisons enable estimation of the relative efficacy and safety of therapies, where competing treatments are often assessed against placebo. Indirect cross-trial treatment comparisons may be possible, but differences in patient characteristics across trials may provide an undesired bias, which cannot be adjusted using aggregate data alone. Signorovitch et al. (2010) proposed a method for adjusting average patient characteristics in trials with individual patient data to match those reported for trials without individual data, thus enabling indirect comparisons in cases where one has access to individual patient data for a particular treatment even if only aggregate data is available for comparator treatments. In this work, we modify the method of Signorovitch using different weights and bootstrapping. We will show comparisons of the weights behavior and the bias reduction of our method with that of Signorovitch through simulations, and we will provide further comparisons using real data.