Four Years of Research Towards Safer, More Personalised Medication

SafePolyMed concluded in August 2026. Bringing together expertise in clinical pharmacy, genetics, artificial intelligence, digital health and patient advocacy, the project developed innovative approaches to understanding medication-related risks and supporting safer treatment. From AI-based risk prediction to patient-centred digital tools, its results lay important foundations for future research and more personalised healthcare.

Why medication safety matters

Taking several medicines at the same time is a reality for many people, particularly those living with chronic conditions. While each medicine may be necessary, combining them can increase the risk of unwanted side effects. Medicines can interact with one another, and a person's genetic makeup can also influence how their body processes and responds to treatment.

SafePolyMed set out to address these challenges by looking at the bigger picture: How can we better predict who is at risk of adverse drug reactions? How can healthcare professionals make more informed treatment decisions? And how can patients become more actively involved in managing their medication? Since its launch in June 2022, the EU-funded project has worked towards answering these questions by combining large-scale health data, advanced computer models and the experiences of patients themselves.

Using artificial intelligence to better understand medication risks

One of SafePolyMed's major achievements is the development of new methods to identify patients who may be more likely to experience adverse drug reactions. Researchers analysed large datasets containing information about medication use, medical conditions and genetic variations. Using AI and machine learning, they developed models that can identify patterns associated with medication-related risks.

A particularly promising model was trained using data from European biobanks and subsequently tested on an independent dataset. It demonstrated encouraging performance in distinguishing patients at higher risk of adverse drug events. Importantly, the researchers also developed ways to analyse sensitive health data across institutions without transferring individual patient records between them. This approach, known as federated learning, offers opportunities for future international research while protecting patient privacy. The results provide a foundation for more personalised risk assessment. Further refinement and clinical validation is necessary before these models can support routine healthcare decisions.

The Medication Management Center

One of SafePolyMed's results is the Medication Management Center (MMC), a digital platform designed to make complex medication information easier to access and understand. For people taking multiple medicines, keeping track of prescriptions, possible interactions and side effects can be challenging. The MMC addresses this by bringing different types of information together. It enables users to maintain an overview of their medication, access information about potential interactions, consider relevant genetic information and record symptoms or health complaints over time. For healthcare professionals, it offers an integrated view of medication-related information that could support discussions with patients.

The MMC was developed with input from patient representatives, whose feedback helped improve its design and usability. Although the MMC is not yet available for routine clinical use, it represents an important step towards digital tools that help patients and healthcare professionals better understand medication-related risks. Learn more about the Medication Management Center

Towards the right dose for the right patient

SafePolyMed also made substantial progress in precision dosing: adapting medication doses to individual patient characteristics rather than relying solely on standard dosing recommendations. Researchers developed mathematical models that simulate how medicines move through the body and how their concentrations may change depending on other medications, genetic differences and individual characteristics. The models were integrated into a web-based Decision Support System (DSS), allowing users to explore complex medication scenarios without requiring specialist expertise in mathematical modelling. By the end of the project, the system could perform dose calculations for eight medicines across more than 6,000 combinations of drug-drug and drug-gene interactions.

This technology demonstrates how advanced modelling could eventually support more personalised treatment decisions. Further testing, validation and regulatory assessment are the next steps to move forward with the DSS.

Putting patients at the centre of medication safety

An important principle throughout SafePolyMed was that patients should not simply receive information about their treatment but also contribute their own experiences and perspectives. Working closely with patient organisations, the consortium developed a set of patient-reported outcome measures (PROMs) for medication safety. These are structured questions that help patients describe symptoms and possible side effects in a consistent way. Patients and healthcare professionals contributed to selecting the most relevant symptoms through a structured consultation process. The resulting set of measures was integrated into the MMC, allowing patients to record information that might otherwise be missed between healthcare appointments.

Beyond developing digital tools, SafePolyMed also invested in education. Together with patient organisations LeukaNET and Patvocates, the project developed a Safety in Polypharmacy training programme covering medication interactions, genetics, side effects, patient rights and participation in healthcare decisions.

Testing the approach in real life: The EmPaSafe study

The EmPaSafe clinical study examined how the Medication Management Center could be used by patients taking multiple medicines. The study enrolled 132 participants across five clinical sites in four European countries: Germany, the Netherlands, Slovenia and Greece. Participants used the MMC, while researchers collected information about their experiences, medication-related knowledge and perceptions of the tool. The findings offer valuable lessons for the further development of patient-centred digital health solutions.

Interviews showed how participants experienced the platform and highlight both the potential and the challenges of digital patient empowerment. Providing information through technology is an important first step, but effective implementation also requires intuitive design, appropriate support and integration into everyday healthcare. The EmPaSafe results will help researchers refine the tools and design future studies to better understand how digital medication management can benefit patients. Keep an eye on our list of publications to know when the EmPaSafe results and further papers currently underway are finally published.

Turning research results into future healthcare solutions

With SafePolyMed officially completed in August 2026, the focus now shifts towards making its results accessible and supporting their further development and uptake.

• Further validation: Testing and refining the AI risk prediction models, MMC and precision dosing tools in additional settings and patient populations.
• Preparing for clinical implementation: Addressing regulatory requirements, software maintenance, interoperability and the evidence needed before digital tools can be used routinely in healthcare.
• Making research results accessible: Publishing scientific findings, sharing models and educational resources to enable other researchers to build on the project's work.
• Exploring future applications: Investigating opportunities to integrate SafePolyMed technologies into existing healthcare systems and to support pharmaceutical research and development.
• Building new collaborations: Pursuing follow-up research and funding opportunities with academic, clinical and commercial partners.

SafePolyMed demonstrated how combining genetic information, real-world health data, advanced modelling and patient perspectives can open new possibilities for understanding medication-related risks. While further research and development are needed before many of these innovations can become part of everyday healthcare, SafePolyMed has established a valuable foundation for future progress.