Use of artificial intelligence to evaluate effectiveness of polypharmacy in patients with depression-related multimorbidity

Lead Supervisor
Dr Alex Dregan
Senior Lecturer
Psychological Medicine, Institute of Psychiatry, Psychology and Neuroscience (IoPPN)
King’s College London
alexandru.dregan@kcl.ac.uk

Co-supervisor
Dr Jayati Das-Munshi

Project Details

Polypharmacy is associated with multimorbidity, but the complex relationship and interplay between polypharmacy, multimorbidity and frailty is less well understood. Partly to blame for this state of affairs are inconsistent definition of polypharmacy across different studies and datasets. Most studies have used numerical definition of polypharmacy, such as the concurrent prescription of 5 or more drugs. Polypharmacy may arise from multimorbidity, but may also accelerated the progression of multimorbidity via multiple pathways, including drug-drug interactions, drug-disease interactions, or adverse drug events. Currently, there is an ongoing debate between appropriate and inappropriate polypharmacy, and this differentiation requires more sophisticated modelling to support clinicians. The complex relationship between polypharmacy, multimorbidity and frailty calls for a systems approach to multimorbidity, that better capture and model the inherent relationships between prescribing practices, multimorbidity, and social factors, facilitating optimal prescribing for individual patients.

Objective
The primary goal of the proposed research is to characterise the complex and dynamic inter-relationships between patterns of prescribing and combination of drugs classes in polypharmacy, inequalities and clusters of diseases among patients with depression-related multimorbidity across the life-course. This evidence will help identify promising avenues for interventions that could improve patient-reported outcomes and reduce health inequalities. To achieve this goal, the proposed project will use AI methods capable to unpack the complex relationship between drug combinations, multimorbidity and functional status across the life-course. Using AI-based data mining and prognostic models, the project will build longitudinal patterns of clusters of polypharmacy. link these patterns to multimorbidity progression and frailty outcome across the life-course.

Datasets

Clinical Practice Research Datalink and Mental Health Services Dataset, obtained all necessary approvals

Keywords

Multimorbidity, machine learning