Development Process of a Clinical Decision Support System for Empiric Antibiotic Therapies in Patients With Sepsis: Case Study

Clinical Snapshot

Key Findings

• P Value: Not reported

• Effect Size: Not reported - qualitative assessment of 'mainly weak' performance

• Primary Outcome: Weak predictive ability for empiric antibiotic therapy selection

• Nnt Or Sensitivity: Not suitable for clinical practice implementation

• Confidence Interval: Not reported

Clinical Application

Low feasibility with current model performance; requires enhanced data quality and quantity Limited applicability to Australian ICUs due to single German hospital derivation; would require local validation and adaptation to Australian antimicrobial guidelines and resistance patterns Currently not applicable - models require significant improvement before clinical use

Abstract

BACKGROUND: Antibiotic therapies are the main treatment for bacterial infections, but growing antibiotic resistance is a major global health threat, severely impacting patients with sepsis. Rapid selection of the most effective antibiotic therapy is critical for survival and for preventing further resistance. Physicians must consider numerous factors for proper empiric treatment selection. A clinical decision support system (CDSS) aims to support physicians in this process, facilitating rapid and targeted therapy. OBJECTIVE: The purpose of this work is to explore the extent to which the realization of a CDSS is possible based on the data available to us and to document insights gained during the development of a foundational model designed to assist physicians in determining empiric treatment options for patients with sepsis. In this regard, rather than aiming to develop a CDSS for clinical application, we highlight the importance of close interprofessional collaboration between scientists from various disciplines and analyze the effects of data quality and quantity on the performance of our statistical models. METHODS: Empirical scientists conducted interviews with medical practitioners to acquire the medical knowledge required to develop sound statistical models. We developed and applied 2-step cross-sectional, as well as time-series classification models, to carefully preprocessed data of patients with sepsis admitted to the intensive care unit of a German hospital. RESULTS: We identified several factors as crucial information for valid decisions on empiric therapy for treating patients with sepsis. These include the patients' core data, especially the infection focus. To prevent further resistance, individual risk factors such as travel history and professional background should be considered. The evaluation of a therapy's effectiveness is mainly based on the patient's general condition and blood values such as procalcitonin and interleukin 6. One key factor in the acceptance of a CDSS is the explainability of the results produced by the applied methods. Our models demonstrated mainly weak predictive ability for all considered empiric antibiotic therapies. However, they are not yet suitable for use in clinical practice, especially as they are based on prescribing habits rather than on optimal treatment decisions. CONCLUSIONS: This work highlights the importance of interprofessional collaboration between medical experts and model developers, ensuring that data quality and clinical relevance are central to the process. It emphasizes the urgent need for high-quality, comprehensive data to overcome challenges such as data discontinuity and improve model performance, particularly through enhanced digitization in health care. This feasibility study will facilitate future efforts to develop a CDSS for treating patients with sepsis and to translate it into clinical use.

References

1. 1.Schmiegel, S., Marchi, H., Hege, P., Elkenkamp, S., Duevel, J., Düsing, C., Greiner, W., Scholz, S. S., Witzke, D., Tebbe, J. J., Wehmeier, M., Kaup, O., Borgstedt, R., Rehberg, S., Cimiano, P., & Fuchs, C. (2026). Development Process of a Clinical Decision Support System for Empiric Antibiotic Therapies in Patients With Sepsis: Case Study. JMIR Medical Informatics. https://doi.org/10.2196/79929