Heart failure is one of the leading causes of death in the Netherlands, yet recognising the final phase of life remains challenging. Patients may experience fluctuating symptoms and periods of temporary stability, which can delay discussions about comfort care and palliative support.

The project led by Alicia Uijl focuses on developing an AI-based prediction model that estimates when a patient with heart failure may be entering the last year of life. The model analyses data from electronic health records, including clinical measurements, medication use and diagnostic results, to identify patterns that are difficult to recognise through clinical judgement alone. By providing timely signals, the tool can support clinicians in initiating earlier conversations about care preferences and goals.

Alongside technical development, the research places strong emphasis on patient, caregiver and clinician perspectives. Through interviews, surveys and focus groups, the team examines experiences with decision-making around comfort care and attitudes towards the use of AI in this sensitive context. This participatory approach helps ensure that the technology supports both clinical practice and patient values.

Trust and transparency are key principles in the project. The AI model is designed to be explainable, enabling clinicians to understand which factors influence predictions. This reduces reliance on opaque systems and supports responsible use in healthcare.

To safeguard patient privacy, federated learning techniques are used, allowing the model to learn from data across multiple hospitals without sharing raw patient data. Development takes place in stages, starting with basic patient characteristics and gradually incorporating more complex clinical information. Validation includes simulated clinical scenarios in which healthcare professionals assess how AI-based insights could influence care decisions.

The long-term aim is to integrate the AI tool into electronic health record systems in both hospital and primary care settings. By supporting timely comfort care decisions, the project illustrates how AI, when developed responsibly and under clinical leadership, can strengthen healthcare and improve quality of life for patients with heart failure.