Drug-drug interactions
Drug-drug interactions
Led by Dr Stella Talic and Dr Harvey Koh
Drug–drug interactions (DDIs) are an important clinical and public health consideration, and are emerging as a specialised focus area within pharmacoepidemiology. DDIs are becoming more common especially within the context of polypharmacy and multimorbidity.
A DDI occurs when one medication alters the effect of another, most often resulting in adverse or harmful outcomes. While the majority of interactions produce adverse responses, some combinations can lead to drug synergy, where therapeutic effects are enhanced in beneficial ways. Many DDIs are identified by post market surveillance, for which, further investigations are warranted in upticks of adverse events.
Our team focuses on combining pharmacoepidemiological approaches to identify DDIs and the expansive field of computer science, to develop preventative approaches and novel methods to identify and classify DDIs. By redefining DDI classification and identification to create a foundational dataset of DDIs and building upon the expansive knowledge of mechanistic pathways, we aim to develop prescribing tools to identify DDIs via predictive modelling.
Our research aims to contribute to the global practices in DDI identification and classification to assist clinicians with management and prevention of DDIs.
Current projects
Our research program is advancing methods to improve the detection, classification, and reporting of DDIs:
- Consensus Study on DDI severity Classification
We are conducting an international, consensus-based study to determine how DDIs should be systematically classified for clinical and regulatory use.
- Checklist for Reporting Standards
We are developing a comprehensive checklist to improve how DDIs are identified, described, and reported across clinical research and regulatory documentation.
- Agreement Study Across Regulators
To evaluate consistency in reporting, we are examining how leading regulatory bodies present DDI information in drug monographs and product characteristics.
- Systematic Review of DDI checkers and their reliability and validity
Our team is conducting a critical review of existing DDI prediction and extraction models, assessing their strengths, limitations, and clinical applicability.
- Clinically Relevant DDI Dataset
We are building a dedicated DDI dataset by using the previously mentioned checklist for classifying DDIs and novel statistical methods.
- Deep Learning Prediction of DDIs
Our team is applying advanced prediction models using resources such as UK Biobank data, to estimate biomarker changes associated with DDIs and identify new DDIs through encoding mechanism pathways.