AI helps nurses document wounds faster

LUMC and KickstartAI take next step towards faster and more complete wound documentation following initial results

LEIDEN, 23 September 2026 - A new AI assistant developed by Leiden University Medical Center (LUMC) and KickstartAI can now automatically prepare several elements of a wound report based on a photograph. An initial pilot shows, among other things, that the technology can already measure larger wounds with relatively high accuracy and provide nurses with a suggested report. This brings faster and more consistently complete wound documentation one step closer. This is particularly relevant because completing a full wound report currently takes an average of 10.7 minutes and, partly as a result, only 10.7% of relevant patient encounters are fully documented.

The AI assistant analyses a photograph of a wound and then generates suggestions for different parts of the report, including the size of the wound and the visible tissue. The nurse reviews the output, adjusts it where necessary and always remains responsible for the final patient record.

Consistent documentation is particularly important in wound care. Healthcare professionals need to be able to see whether a wound is becoming smaller, how the tissue is changing and whether healing is progressing in the right direction. When this information is recorded consistently, the development of a wound can be tracked more effectively over time and different healthcare professionals have access to the same baseline information.

Initial results

During the pilot, three experienced wound care nurses assessed 150 existing wound cases. For wounds measuring five centimetres or more, the average deviation of the automated measurements was 6.8% for wound length and 11.4% for wound width.

The test therefore shows that AI can already prepare several parts of the wound documentation process. At the same time, it highlighted areas where further improvement is needed, including the automatic identification of precise wound boundaries. The nurses’ assessments and corrections are generating new expert-validated data that can be used to further train the models.

“The most important result is that we can now see that AI is genuinely capable of preparing parts of the time-consuming documentation work,” says Alexander van Someren, Product Manager at LUMC Cairelab.  “If we can improve this further, it will become easier for nurses to document wounds more frequently and in a consistent way.”

From model to real-world application

The TAIME project originated from the 2025 National AI Challenge, an initiative by KickstartAI and the AI Coalition for the Netherlands (AIC4NL). Cairelab, LUMC’s AI center of excellence for the responsible implementation of artificial intelligence in healthcare, is also involved in the project. Cairelab focuses on translating promising AI technologies into responsible use in everyday clinical practice.

The next step is a pilot with real patients. This will assess how much time the application can actually save in day-to-day practice and whether it enables nurses to complete wound reports more frequently. For this phase, the models will be further improved and the application will be prepared for integration into everyday wound care workflows.

For this next phase, LUMC is also looking for collaboration partners. These include other hospitals that can share wound data to help make the models more robust, as well as medical software providers and startups interested in contributing to the further development and implementation of TAIME.

“For us, AI only really matters when it moves from experimentation into everyday practice and solves a concrete problem. With TAIME, we are not only building a practical application for wound care, but also learning what it takes to truly integrate AI into the work of healthcare professionals. We want to share those lessons so that other healthcare organizations can also move more quickly from pilot to practice,” says Erick Webbe, CEO of KickstartAI.