Consolidating and Advancing the Methodology to Address Missing Participant Outcome Data in Conventional and Network Meta-analysis of Healthcare Interventions

• Applicant: Privatdozentin Dr. Loukia Spineli
• Subject Area: Epidemiology and Medical Biometry/Statistics
• Term: from 2017 to 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 339420617

Despite the advent of empirical and methodological publications for proper reporting and handling of aggregate missing (participant) outcome data (MOD), researchers insist on employing suboptimal strategies for MOD. Occasioned by the current status quo, we performed a multifaceted research regarding the handling of MOD in network meta-analysis (NMA) that revealed the same reporting shortcomings with conventional systematic reviews. In addition, our research first demystified the properties of several strategies to address aggregated MOD. Based on these findings, we are now aiming to enrich and further advance the methodology for MOD in systematic reviews. Our research agenda includes an extensive refinement of current modelling strategies for continuous MOD that will offer, in addition, the possibility to learn about the missingness mechanisms in the collected trials. Furthermore, we will develop a user-appealing graphical approach to present the results from conventional and network meta-analysis. This graphical approach aims to accommodate several clinically plausible scenarios about the missingness mechanisms while handling MOD properly. To illustrate the proposed methodology for NMA, we will consider our collection of networks from our previous work, whereas for conventional meta-analysis, we will use our collection of Cochrane systematic reviews in mental health. With the intention to initiate and establish a paradigm shift on addressing MOD properly in systematic reviews, we will create an R-package that will provide a complete, up-to-date synthesis toolkit to handle MOD in conventional and network meta-analysis. We consider this software environment to be ideal for delivering all functions that have been (and will be) developed within this project.

• DFG Programme: Research Grants