Comparison of qSOFA score, SOFA score, and SIRS criteria for the prediction of infection and mortality among surgical intermediate and intensive care patients.

BACKGROUND: It is crucial to rapidly identify sepsis so that adequate treatment may be initiated. Accordingly, the Sequential Organ Failure Assessment (SOFA) and the quick SOFA (qSOFA) scores are used to evaluate intensive care unit (ICU) and non-ICU patients, respectively. As demand for ICU beds rises, the intermediate care unit (IMCU) carries greater importance as a bridge between the ICU and the regular ward. This study aimed to examine the ability of SOFA and qSOFA scores to predict suspected infection and mortality in IMCU patients. METHODS: Retrospective data analysis included 13,780 surgical patients treated at the IMCU, ICU, or both between January 01, 2012, and September 30, 2018. Patients were screened for suspected infection (i.e., the commencement of broad-spectrum antibiotics) and then evaluated for the SOFA score, qSOFA score, and the 1992 defined systemic inflammatory response syndrome (SIRS) criteria. RESULTS: Suspected infection was detected in 1306 (18.3%) of IMCU, 1365 (35.5%) of ICU, and 1734 (62.0%) of IMCU/ICU encounters. Overall, 458 (3.3%) patients died (IMCU 45 [0.6%]; ICU 250 [6.5%]; IMCU/ICU 163 [5.8%]). All investigated scores failed to predict suspected infection independently of the analyzed subgroup. Regarding mortality prediction, the qSOFA score performed sufficiently within the IMCU cohort (AUCROC SIRS 0.72 [0.71-0.72]; SOFA 0.52 [0.51-0.53]; qSOFA 0.82 [0.79-0.84]), while the SOFA score was predictive in patients of the IMCU/ICU cohort (AUCROC SIRS 0.54 [0.53-0.54]; SOFA 0.73 [0.70-0.77]; qSOFA 0.59 [0.58-0.59]). CONCLUSIONS: None of the assessed scores was sufficiently able to predict suspected infection in surgical ICU or IMCU patients. While the qSOFA score is appropriate for mortality prediction in IMCU patients, SOFA score prediction quality is increased in critically ill patients.

Generating Surveillance Data for Nosocomial Infections from Routine Charting in Intensive Care Units.

The University Hospital of Giessen and Marburg (UKGM), Giessen, Germany participates in a country wide nosocomial infections surveillance project on intensive care units (ICU) called KISS. KISS data must be gathered daily and is comprised of detailed clinical data like patient movements, ward utilization, presence of central venous and urinary catheters and types of artificial respiration. This happens for all of the 10 ICU's at UKGM Giessen and proved to be very tedious and time consuming for the institute for hospital hygiene. The goal of this project was the automated generation of the KISS reports from routinely collected data in the patient data management system (PDMS). The results show that this is largely feasible without changing the documentation habits of physicians and nurses.

StudyAlert: From eCharts to Modern Messengers.

We developed a tool to detect patients possibly elegible for clinical studies by analysing the HL7-message-stream of the patient management system and notifying study investigators by email or a common messenger service via secured communication channels.

Effects of inverse ratio ventilation with PEEP on kidney function.

INTRODUCTION: Positive pressure ventilation influences hemodynamics and impairs renal function. The effects of inverse ratio ventilation (IRV) on hormonal response and renal function need to be investigated because this represents a highly invasive form of positive pressure ventilation. MATERIALS AND METHODS: Ten male patients were treated postoperatively for 60 min using five different ventilation modes. At the end of these periods, hemodynamics, urine production, fractional excretion of sodium (FESo), as well as the hormones, atrial natriuretic peptide (ANP), renin, angiotensin II, aldosterone, and antidiuretic hormone (ADH), were measured in plasma. RESULTS: Central venous pressure (CVP), ADH, and renin with a positive end-expiratory pressure (PEEP) of 15 cm H(2)O and an inspiration/expiration ratio (I:E) of 1:2 revealed significant increases from baseline, whereas CVP, renin, and FESo showed an increase with an PEEP of 15 cm H(2)O and an I:E ratio of 2:1. Urine production significantly decreased with an PEEP of 15 cm H(2)O and an I:E ratio of 2:1. CONCLUSIONS: IRV with PEEP induced clear effects on hemodynamics and hormonal responses (renin) as well as a decrease in urine production in patients with healthy renal and pulmonary systems after an operation. However, all parameters apart from FESo and renin remained within the normal range. Whether pathological values are also observed after longer periods of positive pressure ventilation shall be the subject of other studies.

Online guideline assist in intensive care medicine--is the login-authentication a sufficient trigger for reminders?

INTRODUCTION: Rising cost pressure due to the implementation of the DRG-System and quality assurance lead to an increased use of therapy standards and standard operating procedures (SOPs) in intensive care medicine. The intention of the German Scientific Society supported project "OLGA" (Online Guideline Assist) is to develop a prototype of a knowledge based system supporting physicians of an intensive care unit in recognizing the indication for and selecting a specific guideline or SOP. While the response of the prototype on user entries can be displayed as a signal on the used workstation itself, the location and time for a reminder of scheduled or missed procedures or reactions to imported information is a difficult issue. One possible approach to this task is the display of non acknowledged reminders or recommendations while logging on to a system. The objective of this study is to analyse user behaviour of the physicians working on the surgical intensive care unit to decide whether the login authentication is a sufficient trigger for clinical reminding. METHODS: The surgical intensive care unit examined in this study comprises 14 beds. Medical care is provided by physicians working in shifts 24 hours a day, 7 days a week, with two anaesthetists at a time and an additional senior consultant during daytime. The entire documentation (examinations, medication, orders, care) is performed using the patient data management system ICUData. The authentication process of the physicians was logged and analysed. RESULTS: Throughout the observation period from December 13th 2005 to January 11th 2006 3563 physician logins were counted in total. The mean span between logins was in 11.3 minutes (SD 14.4), the median 7 minutes. The 75% centile was 14 minutes, the 95% centile 38 min. Intervals greater than 60 minutes occurred in 75%, and greater than 90 minutes in 25% of the days. DISCUSSION: It seems reasonable that reminders sent during authentication are able to enforce workflow compliance. It is possible to send notifications caused by external events to the physician depending on the importance of the event. Serious events with high urgency should be reliably passed using wireless pager or handheld technology. It seems that after the implementation of the prototype guideline assist further investigation is needed to monitor changes in authentication behaviour and reactions to the guideline advisory. This is also required to investigate the influence of unit's size, medical specialty and actual ward workload.

An integrated data-warehouse-concept for clinical and biological information.

The development of medical research networks within the framework of translational research has fostered interest in the integration of clinical and biological research data in a common database. The building of one single database integrating clinical data and biological research data requires a concept which enables scientists to retrieve information and to connect known facts to new findings. Clinical parameters are collected by a Patient Data Management System and viewed in a database which also includes genomic data. This database is designed as an Entity Attribute Value model, which implicates the development of a data warehouse concept. For the realization of this project, various requirements have to be taken into account which has to be fulfilled sufficiently in order to align with international standards. Data security and protection of data privacy are most important parts of the data warehouse concept. It has to be clear how patient pseudonymization has to be carried out in order to be within the scope of data security law. To be able to evaluate the data stored in a database consisting of clinical data collected by a Patient Data Management System and genomic research data easily, a data warehouse concept based on an Entity Attribute Value datamodel has been developed.