Objective: To present an approach of clustering physical interventions for Parkinson’s disease (PD) patients within a network meta-analysis (NMA) that allows the integration of interventions that are highly diverse with respect to e.g. training modality, environment, and use of devices.

Background: Physical interventions are known to be effective in the management of PD improving e.g. motor functions and quality of life. In this broad field, it is still not clear which intervention types (e.g. gait training, Tai Chi, cycling) are most effective in PD. NMA have become a popular method to address this question. However, clustering these interventions when conducting NMA can be challenging due to their high complexity, overlapping intervention components, and missing information on the specific intervention content.

Method: We conducted a systematic literature search for randomized controlled trials (RCTs) of physical interventions for PD patients and clustered eligible trials using a pre-existing approach. We took an approach that had been developed to categorize physical interventions for falls prevention trials (ProFaNE taxonomy; Lamp 2007) as a basis, adapted the original categories and added new categories to integrate eligible physical interventions that could not be matched clearly to the original categories.

Results: The original ProFaNE taxonomy specified five categories for structured exercise: Gait, balance and functional training; strength/resistance; flexibility; 3D-exercises (e.g. Tai Chi, dance); endurance. We separated the original 3D category into mind-body and dance which we considered distinct interventions, and added water-based training as a third 3D category in order to integrate aquatic interventions. For the integration of physical interventions that were delivered using a virtual reality (VR) device which we did not find covered in any of the existing categories, we added a VR category.

Conclusion: Our adaptation of the pre-existing taxonomy allows to cluster a wide range of physical interventions in NMA. Therefore, a more realistic picture of current physical interventions can be represented in analyses comparing several treatments. Our operationalization of each cluster may help trial investigators to describe their interventions more precisely. The adapted system may be used when synthesizing evidence on physical interventions in other diseases.

References: Lamb S, Becker C, Gillespie L, Smith J, Finnegan S, Potter R, Pfeiffer K on behalf of Taxonomy Investigators and the ProFaNE Group (2007). Taxonomy to describe and conceptualise fall prevention interventions (Manual). Available at www.profane.co.

« Back to MDS Virtual Congress 2020

MDS Abstracts - https://www.mdsabstracts.org/abstract/clustering-physical-interventions-for-patients-with-parkinsons-disease-in-network-meta-analyses/