AI Meets ePA: Creating More Value from Health Data

Consider a common situation following a hospital stay. A discharge letter is available in the ePA, containing clinical findings, treatment information and a recommendation to visit an orthopaedic specialist for a follow-up appointment within three weeks. Once back home, questions often arise:

IBM demonstrates how AI and connected services can work together to address these challenges. AI can make discharge information easier to understand, simplify the search for appropriate follow-up care and support preparation for the next consultation. In this way, information stored in the ePA becomes actionable and relevant between clinical encounters.

Making Discharge Summaries
Easier to Understand

Discharge summaries contain essential information for ongoing treatment. However, they are primarily written for healthcare professionals, making it difficult for many patients to identify what is most relevant for their recovery and follow-up care. An AI-powered summary can make this information more accessible by explaining:

Translation into additional languages can further improve accessibility and understanding. The functionality developed by IBM runs directly on the insured person’s smartphone and can be integrated into the ePA application. Explanations are generated where the document is already stored, while the original discharge letter remains available as the primary source.

This enables patients to quickly understand what happened, what comes next and which questions they may want to discuss with their physician.

From Recommendation to Appointment

Understanding a recommendation is only the first step. Acting on it is equally important. In our example, the discharge summary recommends an orthopaedic follow-up appointment within three weeks. The challenge now is finding the right specialist and securing an appointment.

This removes unnecessary friction from the process. Patients no longer need to determine which specialty to search for, where to begin their search or which practice is suitable. A documented recommendation can become a direct pathway to follow-up care. From a technical perspective, such integration is already feasible within an ePA application.

Preparing for the Next Consultation

Once an appointment has been booked, the healthcare provider may ask the patient to complete a medical history questionnaire covering symptoms, pre-existing conditions and medication. Much of this information has already been captured elsewhere and is available in the ePA.

In the envisioned workflow, the practice sends the questionnaire as a PDF via TI Messenger, the communications service of Germany’s telematics infrastructure. AI analyses the document and pre-populates relevant fields using information from the ePA. Any missing details are then completed through an interactive dialogue with the patient.

Patients can add current symptoms, review imported information and make corrections where necessary. After validation, the completed medical history is transmitted to the practice in a structured format. Questions collected by the patient in advance can also be shared.

This reduces repetitive manual data entry while providing healthcare professionals with a prepared overview before the consultation. Transparency remains essential:

Unlocking the Value of
Paper-based Records

Many aspects of a person’s health history still exist exclusively in paper form. Vaccination records are a common example, often spanning many years and containing handwritten entries or attached labels.

Germany’s electronic vaccination record is already on the ePA roadmap. AI can complement this development by simplifying the digitisation of existing paper documentation. The concept is straightforward:

This creates a bridge between historical paper records and digital healthcare systems. Following an injury, for example, a medical practice could quickly verify whether a current tetanus vaccination is documented and whether a booster may be required. Clinical decisions continue to be based on verified medical information.

Identifying Duplicate Medication

Medication management represents another area where AI-supported services can provide tangible value.

This approach has the potential to improve patient safety while also avoiding unnecessary medication costs. However, its effectiveness depends on maintaining an accurate and up-to-date medication record, including over-the-counter products and medications that are no longer being taken.

The Next Step Matters

The future development of the ePA is not only about storing information, but about providing the right information in the right format at the right time. The use cases outlined here demonstrate how healthcare information can be combined with practical support:

Realising this vision requires more than isolated solutions. The greatest value emerges when services work together seamlessly. Understanding a discharge letter becomes even more valuable when recommended follow-up care can be arranged directly and relevant information is available at the next clinical encounter.

Achieving this level of support requires shared standards, interoperable interfaces and a connected ecosystem. Individual showcases help illustrate what is possible. An interoperable architecture provides the foundation for making these capabilities available throughout the entire patient journey.

Technical and regulatory questions remain, depending on the specific application. Ultimately, however, success will be measured by everyday outcomes:

At precisely these transition points, health information can evolve into continuous, patient-centred support.

The approaches and showcases presented were developed collaboratively by Benjamin Freudenreich, Katina Sostmann, Lukas Bauer, Philip Niemeyer, Stefany Santos-Greiner and Tobias Niklas Sasse. Bringing together expertise in digital health, AI, design, product development and IT architecture, the team is working to evolve the ePA into a practical companion that supports insured individuals throughout their healthcare journey.

Share