Bram Gadeyne

COSARA: Integrated Service Platform for Infection Surveillance and Antibiotic Management in the ICU

The Intensive Care Unit is a data intensive environment where large volumes of patient monitoring and observational data are daily generated. Today, there is a lack of an integrated clinical platform for automated decision support and analysis. Despite the potential of electronic records for infection surveillance and antibiotic management, different parts of the clinical data are stored across databases in their own formats with specific parameters, making access to all data a complex and time-consuming challenge. Moreover, the motivation behind physicians’ therapy decisions is currently not captured in existing information systems. The COSARA research project offers automated data integration and services for infection control and antibiotic management for Ghent University Hospital. The platform not only gathers and integrates all relevant data, it also presents the information visually at the point of care. In this paper, we describe the design and value of COSARA for clinical treatment and infectious diseases monitoring. On the one hand, this platform can facilitate daily bedside follow-up of infections, antibiotic therapies and clinical decisions for the individual patient, while on the other hand, the platform serves as management view for infection surveillance and care quality improvement within the complete ICU ward. It is shown that COSARA is valuable for registration, real-time presentation and management of infection-related and antibiotics data.

Validity analysis of a unique infection surveillance system in the intensive care unit by analysis of a data warehouse built through a workflow-integrated software application

BACKGROUND An electronic decision support programme was developed within the intensive care unit (ICU) that provides an overview of all infection-related patient data, and allows ICU physicians to add clinical information during patient rounds, resulting in prospective compilation of a database. AIM To assess the validity of computer-assisted surveillance (CAS) of ICU-acquired infection performed by analysis of this database. METHODS CAS was compared with prospective paper-based surveillance (PBS) for ICU-acquired respiratory tract infection (RTI), bloodstream infection (BSI) and urinary tract infection (UTI) over four months at a 36-bed medical and surgical ICU. An independent panel reviewed the data in the case of discrepancy between CAS and PBS. FINDINGS PBS identified 89 ICU-acquired infections (13 BSI, 18 UTI, 58 RTI) and CAS identified 90 ICU-acquired infections (14 BSI, 17 UTI, 59 RTI) in 876 ICU admissions. There was agreement between CAS and PBS on 13 BSI (100 %), 14 UTI (77.8 %) and 42 RTI (72.4 %). Overall, there was agreement on 69 infections (77.5%), resulting in a kappa score of 0.74. Discrepancy between PBS and CAS was the result of capture error in 11 and 14 infections, respectively. Interobserver disagreement on probability (13 RTI) and focus (two RTI, one UTI) occurred for 16 episodes. The time required to collect information using CAS is less than 30% of the time required when using PBS. CONCLUSION CAS for ICU-acquired infection by analysis of a database built through daily workflow is a feasible surveillance method and has good agreement with PBS. Discrepancy between CAS and PBS is largely due to interobserver variability.

Predictive modelling of survival and length of stay in critically ill patients using sequential organ failure scores

INTRODUCTION The length of stay of critically ill patients in the intensive care unit (ICU) is an indication of patient ICU resource usage and varies considerably. Planning of postoperative ICU admissions is important as ICUs often have no nonoccupied beds available. PROBLEM STATEMENT Estimation of the ICU bed availability for the next coming days is entirely based on clinical judgement by intensivists and therefore too inaccurate. For this reason, predictive models have much potential for improving planning for ICU patient admission. OBJECTIVE Our goal is to develop and optimize models for patient survival and ICU length of stay (LOS) based on monitored ICU patient data. Furthermore, these models are compared on their use of sequential organ failure (SOFA) scores as well as underlying raw data as input features. METHODOLOGY Different machine learning techniques are trained, using a 14,480 patient dataset, both on SOFA scores as well as their underlying raw data values from the first five days after admission, in order to predict (i) the patient LOS, and (ii) the patient mortality. Furthermore, to help physicians in assessing the prediction credibility, a probabilistic model is tailored to the output of our best-performing model, assigning a belief to each patient status prediction. A two-by-two grid is built, using the classification outputs of the mortality and prolonged stay predictors to improve the patient LOS regression models. RESULTS For predicting patient mortality and a prolonged stay, the best performing model is a support vector machine (SVM) with GA,D=65.9% (area under the curve (AUC) of 0.77) and GS,L=73.2% (AUC of 0.82). In terms of LOS regression, the best performing model is support vector regression, achieving a mean absolute error of 1.79 days and a median absolute error of 1.22 days for those patients surviving a nonprolonged stay. CONCLUSION Using a classification grid based on the predicted patient mortality and prolonged stay, allows more accurate modeling of the patient LOS. The detailed models allow to support the decisions made by physicians in an ICU setting.

Random Survival Forests for Predicting the Bed Occupancy in the Intensive Care Unit

Predicting the bed occupancy of an intensive care unit (ICU) is a daunting task. The uncertainty associated with the prognosis of critically ill patients and the random arrival of new patients can lead to capacity problems and the need for reactive measures. In this paper, we work towards a predictive model based on Random Survival Forests which can assist physicians in estimating the bed occupancy. As input data, we make use of the Sequential Organ Failure Assessment (SOFA) score collected and calculated from 4098 patients at two ICU units of Ghent University Hospital over a time period of four years. We compare the performance of our system with a baseline performance and a standard Random Forest regression approach. Our results indicate that Random Survival Forests can effectively be used to assist in the occupancy prediction problem. Furthermore, we show that a group based approach, such as Random Survival Forests, performs better compared to a setting in which the length of stay of a patient is individually assessed.

De-Escalating Anti-Pseudomonal β-Lactams

De-escalation of broad-spectrum antibiotics such as anti-pseudomonal beta-lactams is recommended to reduce antimicrobial selection pressure.