Wir überwinden Grenzen

Artificial intelligence can significantly simplify medical documentation / Best model delivers 93.1 percent usable documents / Potential to reduce the workload of medical professionals / Already in regular use

Researchers at the Medical Center - University of Freiburg have investigated the suitability of large language models (LLMs) for the creation of medical documentation in the German healthcare system. A comparison of four language models showed that the best results were achieved with a non-commercial model. Here, 93.1 percent of the reports could be used with only minimal adaptations. The study, which was published in the journal JMIR Medical Informatics on August 28, 2024, underlines the potential of such models to simplify documentation processes and thus reduce the workload of physicians and nursing staff. The AI software is already being used to some extent in regular operations at the Medical Center - University of Freiburg.

"Our results show that models specially trained for the German language can provide valuable support in the creation of medical reports. This could significantly simplify workflows in everyday clinical practice," saysstudy leader Dr. Christian Haverkamp, acting director of the Institute of Digitalization in Medicineat the Medical Center - University of Freiburg.

"The AI doctor's letter is an excellent example of how much potential AI applications have in medicine. For such solutions, we need bright minds who are willing to experiment and develop new things. I am delighted that we have created an environment at the Medical Center - University of Freiburg that strongly promotes these activities," says Prof. Dr. Frederik Wenz, Chief Medical Officer of the Medical Center - University of Freiburg.

Advanced technologies for clinical practice

In the study, first author Felix Heilmeyer, research associate at the Institute of Digitalization in Medicine, used a total of 90,000 real clinical documents from the Medical Center - University of Freiburg's Department of Ophthalmology to train the models. Several language models generated doctor's letters on this basis, which were then evaluated by medical professionals. The evaluations showed that 93.1 percent of the documents generated by the BLOOM-CLP-German model were suitable for clinical use after minor corrections. Start of formEnd of form In addition to the choice of the appropriate model, the researchers emphasize that its adaptation to the respective language area plays a decisive role in its suitability for practical use.

Three hours of documentation a day

According to surveys, physicians spend almost three hours a day on documentation activities. "The automation of medical documentation has the potential to save physicians valuable time that they can devote directly to their patients. At the Medical Center - University of Freiburg's Department of Ophthalmology, we already use an AI tool for writing doctors' letters on a regular basis," says Prof. Dr. Daniel Böhringer, initiator of the study and senior physician at the Medical Center - University of Freiburg's Department of Ophthalmology.

The biggest challenge of the study was to ensure that the documents generated by the language model met the high standards of medical documentation in German. "Adapting to specific medical terminology and the structure of clinical reports was a particularly demanding task, as the model had to deliver texts that were both precise and comprehensible," says Böhringer.

Original title of the publication: Viability of Open Large Language Models for Clinical Documentation in German Health Care
DOI: 10.2196/59617
Link to the study:JMIR Medical Informatics