The DynAIRx Project Protocol: Artificial Intelligence for dynamic prescribing optimisation and care integration in...

Area of Science:

• Health Informatics
• Artificial Intelligence in Healthcare
• Pharmacovigilance

Background:

• Structured Medication Reviews (SMRs) aim to optimize medicines for patients with multiple long-term conditions and polypharmacy, aligning with the NHS Long Term Plan.
• Challenges exist in gathering comprehensive patient data due to fragmented health records and a lack of guidance on identifying patients most in need of SMRs.

Purpose of the Study:

• To develop and pilot an interpretable AI approach for extracting longitudinal health and medication information from scattered clinical records.
• To predict risks of adverse outcomes and integrate this information into care records to inform SMRs.
• To co-design future medicines optimization decision support systems with end-users and patients.

Main Methods:

• The DynAIRx system will analyze structured clinical data from integrated care records (covering ~11 million people) and unstructured text using Natural Language Processing (NLP).
• AI models will be trained to identify patterns preceding adverse events, focusing on patients with mental and physical health issues, those with four or more conditions on ten or more drugs, and older, frail individuals.
• The approach will be piloted within primary care prescribing audit and feedback systems, incorporating AI-augmented visualization of care records.

Main Results:

• The study aims to create a learning system for medicines optimization by implementing and evaluating an AI-augmented visualization tool.
• The system is designed to identify individuals who would most benefit from SMRs, thereby improving the efficiency and effectiveness of medication reviews.
• Co-design with end-users and patients throughout the process ensures the system's practical applicability and user-centeredness.

Conclusions:

• AI-driven analysis of integrated health records can significantly enhance the identification and prioritization of patients for Structured Medication Reviews.
• This approach promises to improve medicines optimization for individuals with complex health needs and polypharmacy.
• The development of AI-augmented decision support systems has the potential to transform medication management within healthcare systems.