Monitor mit medizinischen Daten


Data-driven Priorization of Intensive Care patients for transfer |

Medical Need

Ideally, ICUs would have a bed and the necessary personnel and equipment available whenever a critical patient arrives. In reality, ICU capacity is restricted and maximum utilization of beds and personnel desirable. At the same time, fast admission is key to the successful treatment of critical patients. Since there is no option of queuing, the physician in charge has to choose suitable patients for transfer to general ward to free up beds. It is a tough decision with potentially life-risking consequences for the transferred patient. Unfortunately, it often happens in periods of high acute work-load, stress or fatigue often involving mutual agreements with other units. Decision support in this area does not only assist the clinician but also serves the welfare of the patient.

PrioriCare provides a decision support tool to identify eligible transfer patients from ICU by using patient-related data as well as operational clinical information. The software evaluates each patient along with their therapeutic characteristics and provides the physician with an aggregate easy-to-interpret score to measure the eligibility of a candidate.

Intensive care unit, transfer planning and control, predictive modelling, business analytics

Project team members |

Clinician |

TU Dresden, Faculty of Medicine
University Hospital Dresden, Department of Anesthesiology and Critical Care Medicine

TU Dresden, Faculty of Medicine
University Hospital Dresden, Surgical Intensive Care Unit

Project team members |

High-tech |

TU Dresden, Faculty of Business and Economics
Industrial Management

Abstract |

Intensive care unit (ICU) capacity is crucial for the treatment of the most severely sick patients. Due to the common lack of ICU bed availability, the ICU physician in charge must triage which patient can be sent to the regular ward when new critical patients arrive. The choice of an ideal candidate for transfer is subject to a variety of medical and operational factors compounded by uncertainty regarding the recovery process of each individual. Transfers are excluded by a number of knock-out-criteria e.g. an ongoing mechanical ventilation or catecholamine therapy. Further restrictions are imposed by some guidelines and threshold levels for vital parameters or indicators such as Simplified Acute Physiology Score (SAPS) or Sequential Organ Failure Assessment (SOFA). However, the decision remains somewhat subjective as it is influenced by several other factors and thus greatly relies on the personal expertise of the physician. In addition, it may come at times of high workload or fatigue.
This project aims to develop a software-based decision support system that supports clinicians in identifying patients who are most suitable for transfer to a lower ward. Our approach is to analyze historical ICU data to estimate the effects of a potential transfer on patient outcome in the presence of specific combinations of vital signs and operational factors. In an acute situation, the software tool applies these estimates to the current set of ICU patients along with their actual therapeutic characteristics in order to identify eligible candidates for transfer. Alternatively, it provides the physician with an aggregate easy-to-interpret eligibility score for each patient.

Skip to content