Authors: Doherty, A. S.1, Essilini, A. 1, Gomez-Lemus, J. 1, Muddiman, R. 1, & Moriarty, F. 1 

Affiliations: 1 School of Pharmacy & Biomolecular Sciences, RCSI University of Medicine and Health Sciences, Dublin

Background/ Introduction: Deprescribing is the planned, supervised process of dose reduction or cessation of medication that may cause harm or are no longer appropriate. Embedding deprescribing into routine clinical practice has been hampered by the lack of clear evidence-based guidance on when and how to deprescribe safely. The DIAMOND programme aims to advance deprescribing by developing advanced analytical methods for application to real-word data.

Material & Methods: The DIAMOND programme comprises several inter-related work packages. Work Package 1 (WP1) characterises deprescribing via a scoping review of deprescribing definitions that have been applied in studies that utilise routinely collected data. These definitions will be applied to primary care databases to examine factors associated with deprescribing. In Work Package 2 (WP2), simulation studies will test and refine the application of the target trial emulation framework to deprescribing. Two simulation studies will compare discontinuation to continuation, for medications that can be stopped immediately and those which require tapering. Work Package 3 (WP3) focuses on serotonergic burden. An overview of reviews summarises evidence on medications associated with serotonin syndrome. Disproportionality analysis of adverse drug reaction (ADR) reports captured within pharmacovigilance databases will detect ADR signals relevant to serotonergic burden. A serotonin syndrome risk index will subsequently be developed using expert consensus methods before application to primary care data.

Results: All work packages are ongoing. The scoping review of deprescribing definitions (WP1) identified 727 studies for inclusion. Discontinuation was most commonly operationalised using gap-based definitions. The simulation study targeting an intention-to-treat estimand (WP2) indicates that static treatment strategies can be estimated with minimal bias using semi-parametric G-methods. The overview of reviews summarising medications associated with serotonin syndrome (WP3) included 33 studies, with antidepressants and analgesics commonly implicated drug classes.

Conclusion: These three work packages demonstrate the broad potential of real-world data to generate high-quality evidence that supports deprescribing.