In the Intensive Care Units (ICU) it is notorious the high number of data sources available. This situation brings more complexity to the way of how a professional makes a decision based on information provided by those data sources. Normally, the decisions are based on empirical knowledge and common sense. Often, they don't make use of the information provided by the ICU data sources, due to the difficulty in understanding them. To overcome these constraints an integrated and pervasive system called INTCare has been deployed. This paper is focused in presenting the system architecture and the knowledge obtained by each one of the decision modules: Patient Vital Signs, Critical Events, ICU Medical Scores and Ensemble Data Mining. This system is able to make hourly predictions in terms of organ failure and outcome. High values of sensitivity where reached, e.g. 97.95% for the cardiovascular system, 99.77% for the outcome. In addition, the system is prepared for tracking patients' critical events and for evaluating medical scores automatically and in real time.

No abstract

Decision making assumes a critical role in the Intensive Medicine. Data Mining is emerging in the clinical area to provide processes and technologies for transforming data into useful knowledge to support clinical decision makers. Appling clustering techniques to the data available on the patients admitted into Intensive Care Units and knowing which ones correspond to readmissions, it is possible to create meaningful clusters that will represent the base characteristics of readmitted patients. Thus, exploring common characteristics it is possible to prevent discharges that will result into readmissions and then improve the patient outcome and reduce costs. Moreover, readmitted patients present greater difficulty to be recovered. In this work it was followed the Stability and Workload Index for Transfer (SWIFT). A subset of variables from SWIFT was combined with the results from laboratory exams, namely the Lactic Acid and the Leucocytes values, in order to create clusters to identify, in the moment of discharge, patients that probably will be readmitted.

Nowadays, it is increasingly important to utilize intelligent systems to support the decision making process (DMP) in challenging areas such as Intensive Medicine. In Intensive Care Units (ICU), some of the biggest challenges relate both to the number and the different types of available data sources. Even though in such a setting the values for some variables are easy to collect, data collection is still performed manually in particular instances. In order to improve the DMP in ICU, a Pervasive Intelligent Decision Support System, called INTCare was deployed in the ICU of Centro Hospitalar do Porto in Portugal. This system altered the way information is collected and presented. Moreover, the tracking system deployed as a specific module of INTCare – Electronic Nursing Record (ENR) is made accessible anywhere and anytime. The system allows for the calculation of the critical events regarding five variables that are typically monitored in an ICU. Specifically, the INTCare tracking system characterizes a grid that shows the events by type and duration, empowers a warning system to alert the doctors and promotes intuitive graphics that allow care providers to follow the patient care journey. User acceptance was measured through a questionnaire designed in accordance with the Technology Acceptance Model (TAM) and results of implementing the INTCare tracking system, and its interface are reported.

Cardiac Arrhythmia (CA) is very dangerous and can significantly undermine patient condition. New tools are fundamental to forecast and to prevent possible critical situations. In order to help clinicians acting proactively, predictive data mining real-time models were induced using online-learning. As input variables were considered those acquired at the patient admission and complementary variables (vital signs, laboratory results, therapeutics) hourly collected. The results are very motivating; sensitivity near to 95% was obtained when using Support Vector Machines. The approach explored in this work reveals to be an interesting contribution to the healthcare in terms of predicting CA and a good direction to be further explored.