Real-world data: better patient monitoring between visits

The medical record documents diagnoses, tests, prescriptions and procedures. Patient-reported questionnaires can track pain, quality of life, fatigue or a specific symptom. Connected devices add repeated measurements such as activity, heart rate or sleep, depending on their capabilities and intended use. Finally, messages, calls or teleconsultations can document a change in the patient’s situation.

Every source has its limits. Self-reported data depends on understanding and context. A connected measurement depends on the device, whether it is actually worn and the conditions of acquisition. Administrative data reflects the care that was billed, not necessarily the full clinical picture.

Before activating any data collection, the care team should answer four questions: which decision will it inform, which variable is needed, what frequency is relevant, and who acts when an alert is raised? This approach avoids overloaded dashboards and signals that nobody owns.

Useful monitoring defines thresholds, exceptions and timeframes. A minor variation can simply be archived. A persistent trend may justify an additional questionnaire. A critical signal can trigger a specific procedure. These rules must be adapted to the population and reviewed in light of real-world experience.

France’s National Authority for Health (HAS) stresses the methodological importance of real-world studies and the sponsor’s responsibility for the data produced. The European Medicines Agency now uses real-world data to address questions of safety, utilization, epidemiology and effectiveness, while highlighting the challenges of source quality and methodology.

The same caution applies to routine care: identifiable provenance, timestamps, units, missing-data rates, device changes and clinical context. A long but unstable series can be less useful than a simple, reliable measurement.

The most immediate risk is organizational. If the system flags too many irrelevant anomalies, teams end up ignoring alerts. Design must therefore involve clinical users from the start, define priority levels and measure the share of alerts that are truly actionable.

Useful indicators are not only technical. Teams should track response time, the number of alerts with no action taken, appointments triggered, missing data, patient satisfaction and the workload perceived by staff. The safety of a system also depends on how well it fits into real-world clinical work.

Building the full alert chain, from trigger to documented action, is covered in our guide to remote patient monitoring.

Health data benefits from enhanced protection. The purpose, legal basis, recipients, retention period and individuals’ rights must all be defined. The European Health Data Space, which entered into force in March 2025, organizes a gradual transition toward greater access and interoperability, without removing security and governance obligations.

A digital twin can serve as a representation layer: it connects the medical record, patient-reported data and certain repeated measurements to give a more coherent view of how the patient’s health evolves. Its relevance, however, depends on the models activated and how they were validated.

MyTwin for clinicians offers a journey in which medical record data and data generated between visits feed continuous monitoring. MyTwin provides recommendations to the professional, who stays in control of the modules activated. To understand the patient side of the experience, see MyTwin for patients, as well as stories from patients and professionals.