AMPEL is a clinical decision support system (CDSS) which provides a clinic-wide digital infrastructure at the University Clinic Leipzig that enables clinical AI applications in real-time. Its main task is to increase patient safety by detecting critical situations in routine healthcare. In this process, it can leverage diverse clinical data: laboratory data, diagnoses, medication, performed or planned surgery and more.
There is a wealth of academic AI models with life-saving potential. However, in practice, they do not find their way into real patient scenarios. Therefore, as the first German clinic, Leipzig University Hospital operates a CDSS for randomized controlled studies in routine healthcare. Once the study phase is successfully completed, AMPEL-models become instantly accessible to all patients.
AMPEL models are based on rule frameworks or machine learning (ML) algorithms. One of the 13 AMPEL models to date is the ML model for the early detection of sepsis based on routine laboratory data. The AMPEL team was able to demonstrate that even a few parameters from the complete blood count (RBC, WBC, PLT, MCV, HGB, Age, Sex) are sufficient to conduct a general sepsis screening. Furthermore, they found that traditional parameters such as Procalcitonin are significantly outperformed when combined with the base model. The early detection of sepsis with routine laboratory data is thus elevated to a new level through the application of ML methods. Integrated into an automated CDSS with real-time alerts, sepsis patients in the emergency department could be identified much earlier than before.
Beyond sepsis, AI algorithms can combine different biomarkers into new digital biomarkers, which have great potential to outperform traditional biomarkers. AMPEL provides the necessary framework for implementation and fast translation of these digital biomarkers in clinical routine.
From 2024 onward, the core of AMPEL will be expanded as an open-source initiative. The focus is on creating a non-profit AI infrastructure for healthcare with the highest standards of transparency, adaptability, and interoperability. As the first academic clinic-wide CDSS in Europe, AMPEL aims to set future standards and accelerate the research of digital assistants in hospitals (www.ampel-cdss.de).
