Computational modelling for the optimisation of pharmacological therapies and prediction of drug-drug interactions

25^th September 2019, 11:00
Marco Siccardi
Translational Medicine, University of Liverpool

Abstract

Individuals can have different diseases at the same time and therefore can be treated with multiple drugs. Drugs given at the same time to patients can influence each other distribution in the body and/or result in side effects. This process is called drug-drug interactions (DDIs) and many patients (ranging from 15% to 60%) have at least one drug-drug interactions with potential relevant side effects.

Unfortunately, due to the extremely high number of potential combinations of drugs is practically impossible to understand how to properly manage DDIs through clinical studies in and many DDIs are unstudied or cannot be ethically studied in patients. Consequently, when clinicians have to prescribe complex therapies, they rely often on opinions of experts instead of solid evidence.
Computational modelling can represent a very powerful approach to improve our understanding of the mechanisms underpinning DDIs and machine learning approaches can support the prediction of DDI magnitude and the identification of drug combinations resulting in higher risk for patients.

A comprehensive integration of pharmacological quantitative data into computational predictive modelling has the potential to generate impactful applications for a better management of existing therapies as well as a more rationale characterisation of risks for future drugs.