The German Federal Ministry of Health is funding the project for the development of smart emergency algorithms through explainable AI procedures (ENSURE), which was developed in cooperation between the Interdisciplinary Emergency Department and the Institute for Medical Informatics. The funding period covers 3 years and is funded with about 1.4 million Euros. The project started on 01.10.2020 and runs until 30.09.2023, with the aim of improving timely, targeted diagnostics and initial therapy in emergency care.

In Germany, preclinical and clinical emergency care has developed into a very demanding professional field of action in recent years. The complex medical challenge consists primarily in the time-critical treatment of patients of all ages and a wide variety of illnesses and injuries, with a spectrum ranging from outpatient treatment to shock room care for the most seriously injured and patients requiring intensive care. In addition, time and cost pressures, multiple interface areas and increasing case numbers place the highest demands on the professional qualifications and skills of the medical staff. However, in this complex, high-risk setting, residents and newcomers to the profession from various disciplines are also deployed on a rotational basis. In the absence of supervision, they often have to resort to web-based, usually non-evidence-based knowledge platforms in case of emergency. The aim of the planned collaborative project ENSURE is to develop the prototype of an on-demand, clinical decision support in terms of smart emergency algorithms, which integrates both evidence-based explicit knowledge (rule-based system) and AI methods (machine learning system). Within the framework of a clinical pilot study, the prototype will be tested in three model hospitals and evaluated with regard to defined process and quality indicators in the emergency department. Ultimately, ENSURE is intended to provide smart emergency algorithms that support physicians' competencies throughout the entire emergency medical treatment chain and thus improve the quality of processes and outcomes in emergency care.

In the planned joint project ENSURE, a prototype of an on-demand clinical decision support system of the same name is to be developed and tested in the sense of smart emergency algorithms for prompt, targeted diagnostics and initial therapy in emergency care. The aim is to use ENSURE to support the action competence of medical staff in the emergency medical treatment chain from prehospital to hospital and thus improve the process and outcome quality in emergency care. Such a data-driven tool is associated with requirements regarding the transparency and traceability of the decision criteria as well as the quantity and quality of the data. The eponymous decision support system ENSURE comprises two decision support systems: a rule-based system as well as an explainable AI system, which support the diagnosis in the emergency room as well as suggest measures for initial therapy. The data used for training the ML procedures are merged in a data integration center in a quality-checked manner and represented in a standardized way. ENSURE integrates components for targeted search in the knowledge base as well as for visualization of decision criteria. The subsequent use of the data quality-checked within the project and the developed software shall be guaranteed.

The evaluation is carried out by means of a prospective, multi-center pilot study at three model hospitals (Universitätsmedizin Berlin - Charité, Klinikum Fürth and UMG).  In the first phase, a detailed as-is analysis of the current health care reality will be conducted. For this purpose, important quality indicators and key figures of the emergency department will be collected.  In addition, knowledge and case simulation tests will be conducted to evaluate the current competence of medical staff in defined emergency situations. In phase 2, the prototype will be tested. The effects of the application on the competence of the target group as well as the process and result quality of the emergency care will be examined.

Network partners

• Heilbronn University, GECKO Institute Medicine, Prof. Dr. Martin Haag
• German Research Center for Artificial Intelligence - Educational Technology Lab, Dipl.-Inform. Michael Dietrich
• Otto von Guericke Universität Magdeburg, Medical Faculty, Prof. Dr. Felix Walcher

Model Clinics

• Charité - Universitätsmedizin Berlin, Prof. Dr. Martin Möckel
• Fürth Hospital, Central Emergency Room, Prof. Dr. Harald Dormann
• Universitätsmedizin Göttingen – Interdisciplinary emergency room, Prof. Dr. Sabine Blaschke (Consortium management in cooperation with Prof. Dr. Dagmar Krefting, Head of Department of Medical Informatics)

Associated partners

• DGINA Deutsche Gesellschaft Interdisziplinäre Notfall- und Akutmedizin e.V.
• DIVI Deutsche Interdisziplinäre Vereinigung für Intensiv- und Notfallmedizin
• Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (GMDS) e.V.

Consortium management

• Prof. Dr. Sabine Blaschke (Central Emergency Room)
• Prof. Dr. Dagmar Krefting (Department of Medical Informatics)

Project coordination

• Kerstin Pischek-Koch (Department of Medical Informatics)