Breast cancer recovery through the lens of real-world data
Recovering from breast cancer often involves managing a complex set of chronic conditions that persist long after primary treatment ends. However, a lot of clinical research still relies on snapshots from infrequent hospital visits, leaving critical gaps in understanding how patients function and feel in their everyday lives. Clinical research has the potential to be transformed by tapping into real-world data(opens in new window), health and lifestyle data collected for each patient from electronic health records (EHR), digital registries, smartphones and wearable devices.
A 360° view of patient recovery
The EU-funded REBECCA(opens in new window) project brings together 12 partners from seven European countries and aims to exploit the novel potential offered by the continuous collection of real-world data. To track patients’ functional, emotional and quality-of-life trajectories, the consortium developed a 360° Monitoring Platform, a suite of minimally obtrusive mobile tools offered to breast cancer survivors. The platform integrates EHRs and patient-reported outcome measures with passively collected digital biomarkers, including physical activity, heart-rate response, sleep patterns, eating habits and mobility trajectories. “The goal was to improve patient care by studying the long-term effects of primary and adjuvant cancer treatments,” explains project coordinator Anastasios Delopoulos. By aggregating and visualising these multimodal data streams, clinicians gain a far richer and more continuous understanding of patient recovery than is possible through routine follow-up visits alone. Instead of relying on appointments spaced months apart, healthcare professionals can observe recovery trends in near real time and receive automated alerts when abnormal patterns emerge.
Using AI to interpret real-world studies
The REBECCA platform has been deployed in six clinical studies across Norway, Spain, and Sweden, involving more than 650 participants. To interpret these rich datasets, REBECCA combined deep learning with AI-powered modelling. The models extract key information from real-world data and from the clinical trials, capturing key determinants of health and quality of life. Causal relationships among variables influencing breast cancer outcomes may also emerge. “By linking treatments, behaviours, and functional outcomes, the models provide actionable insights that support informed clinical decision-making,” emphasises Delopoulos. Early findings demonstrate that continuous monitoring can detect changes in activity, sleep or heart rate weeks before they would typically surface in routine care. These insights demonstrate the added value of real-world data in capturing aspects of recovery that matter most to patients but often remain invisible in standard clinical settings.
Building a foundation for future practice
Among REBECCA’s most significant achievements is the creation of a robust infrastructure for privacy-preserving, federated analysis of cross-country data. Just as important has been the project’s close collaboration with patient representatives, including the Swedish group AMAZONA, whose involvement helped ensure that the platform remains patient-centred and clinically relevant. As the project concludes, REBECCA leaves behind a unique multimodal dataset and a scalable platform designed for continued use. Next steps include deploying the infrastructure in additional oncology clinics, extending it to other cancers and chronic diseases, and refining causal models by integrating behavioural and biological data. Through these efforts, REBECCA aims to support the wider adoption of real-world data as a cornerstone of clinical research and patient management across Europe.
