Overcoming challenges in network meta-analyses of non-pharmacological trials: Experiences from a Cochrane review on physical exercise in Parkinson´s disease
2Medical Psychology | Neuropsychology and Gender Studies & Centre for Neuropsychological Diagnostics and Intervention (CeNDI), Faculty of Medicine and University Hospital, University of Cologne, Cologne, Germany
3Ambulatory Assessment in Psychology, Department of Psychology, Carl von Ossietzky University Oldenburg, Oldenburg, Germany
4Department of Neurology, University Hospital Marburg, Marburg; Department of Neurology, Knappschaftskrankenhaus Bottrop GmbH, Bottrop, Germany
5Institute of Medical Sociology, Health Services Research and Rehabilitation Science, Faculty of Human Sciences and Faculty of Medicine and University Hospital, University of Cologne, Cologne, Germany
6MAGIC Evidence Ecosystem Foundation, Oslo, Norway
7Institute of Health and Society, Faculty of Medicine, University of Oslo, Oslo, Norway
While network meta-analysis (NMA) allows to compare multiple interventions in a single analysis by combining direct and indirect evidence, this technique introduces several challenges to authors of systematic reviews (SR) especially in the field of non-pharmacological trials, e.g., in clustering interventions, analyzing poorly reported safety outcomes, and presenting key results.
Objectives: To present solutions to challenges specific to NMAs of non-pharmacological trials implemented in a Cochrane review on physical exercise in Parkinson’s disease (PD).
Methods: We clustered physical exercise in N=156 eligible studies adapting the ProFaNE taxonomy developed to categorize falls prevention trials for the elderly (Lamb 2011) in order to compare the effects of various exercise types in an NMA. While we rated the confidence in the evidence using CINeMA (Nikolakopoulou 2020) for our primary efficacy outcomes, we summarized poorly reported safety data narratively and rated our confidence in the evidence using GRADE (Schünemann 2022). We additionally created an interactive Summary of Findings (iSoF) table using the MATCH-IT tool (https://magicevidence.org/match-it) to present key results beyond the information included in conventional SoF tables.
Results: Our approach to cluster interventions allowed to compare the effects of a wide range of exercise types representing a realistic picture of current treatment options in PD in an NMA. Combining quantitative and narrative analyses, we systematically synthesized both efficacy and safety data to the most informative degree possible given the available evidence. Integrating an iSoF table enabled us to present key results including available data, confidence in the evidence and intervention effects for multiple outcomes and multiple interventions in a clear and consumer-friendly way.
Conclusions: We provide solutions to several challenges faced when conducting a Cochrane review with NMA on physical exercise in PD. Our approach may guide authors of SRs with NMAs of non-pharmacological trials in clustering interventions, analyzing poorly reported safety outcomes and presenting key results.
Patient, public and/or healthcare consumer involvement: Our work supports authors who conduct SRs in dealing with challenges specific to NMAs of non-pharmacological trials including consumer-friendly presentation of key results. It may therefore contribute to producing robust evidence syntheses that are relevant and informative for consumers.