Project

Personalized, Data-driven Prediction and Assessment of Infection Related Outcomes in Swiss ICUs (IICU)

Short Summary

Infectious diseases caused by bacteria, viruses, and fungi show a wide spectrum of clinical signs and symptoms and diverse clinical outcome. Some patients may show only mild fever, whereas others show a severe inflammation and even fatal course. This diversity is rooted in the complex interactions between the patient, the microorganism causing the infection, and the timing of treatment. Our project proposal aims to better understand the diversity of this interaction in patients requiring intensive care medicine. We will use a combination of clinical and laboratory data to improve personalized assessment, characterization, and outcome prediction in patients with infections. We will consider the clinical reasoning process of the attending physicians and the resulting patient assessment. This will significantly improve the quality of the collected data and generate a globally unique dataset for research in favor of critical ill patients with infections.

Goals

Our goals are to: 1. Define, standardize, document, and predict infection-related clinical presentation, course, and outcomes. 2. Increase accessibility of data including well-described clinical presentations, course, and outcomes. 3. Generate a public data repository to improve research in this important field 4. Develop new approaches to digital biomarker discovery and outcome prediction using artificial intelligence and machine learning. 5. Provide feedback to improve data quality control procedures. 6. Validate digital biomarkers and assess potential impact on treatment.

Significance

Our project will provide the baseline to (i) define, monitor, characterize and predict important infection-related outcomes, (ii) transfer and access data between involved centers using high quality data of clinical and microbiological phenotypes, (iii) to develop a new approach of clinician-to-data scientist exchange in order to analyze the data and generate novel types of digital biomarkers for risk assessments; and (iv) finally data-driven feedback to clinicians and microbiologists will help us to further improve the data quality, and optimize the clinical and diagnostic management of ICU patients with suspected or confirmed infections.

Background

Infections show a range of diverse phenotypes with variable impact on clinical course and outcomes. Our NDS proposal focuses on this heterogeneity within critically ill patients with infections using a combined data-driven approach for an improved personalized assessment, characterization, and outcome prediction on patients with infections.