Network Meta-Regression for Ordinal Outcomes Under Different Links

• Activity Number: 506 - Categorical Data
• Type: Contributed
• Date/Time: Wednesday, July 31, 2019 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Yeongjin Gwon* and Mo May and Ming-Hui Chen and Zhiyi Chi and Juan Li and Amy Xia and Joseph G Ibrahim
• Companies: University of Nebraska Medical Center and Amgen Inc and University of Connecticut and University of Connecticut and Eli Lily and Company and Amgen Inc and UNC
• Keywords: Ordinal outcomes; Indirect comparison; Aggregate covariates; Random effects; Goodness-of-fit; Links

Abstract

In this presentation, we propose a network meta-regression approach for modeling ordinal outcomes under different links. Specifically, we develop regression models based on aggregate trial-level covariates for the underlying cut-off points of the ordinal outcomes as well as for the variances of the random effects to capture heterogeneity across trials. Our proposed models are particularly useful for indirect comparisons of multiple treatments that have not compared head-to-head within the network meta-analysis framework. Moreover, we introduce Pearson residuals and construct an invariant test statistic to evaluate goodness-of-fit in the setting of ordinal outcome meta-data. A case study demonstrating the usefulness of the proposed methodology is carried out using aggregate ordinal outcomes.