Effects of an online information tool on post-traumatic stress disorder in relatives of intensive care unit patients: a multicenter double-blind, randomized, placebo-controlled trial (ICU-Families-Study).

PURPOSE: Intensive care unit (ICU) hospitalization is challenging for the family members of the patients. Most family members report some level of anxiety and depression, sometimes even resulting in post-traumatic stress disorder (PTSD). An association has been reported between lack of information and PTSD. This study had three aims: to quantify the psychological burden of family members of critically ill patients, to explore whether a website with specific information could reduce PTSD symptoms, and to ascertain whether a website with information about intensive care would be used. METHOD: A multicenter double-blind, randomized, placebo-controlled trial was carried out in Austria and Switzerland. RESULTS: In total, 89 members of families of critically ill patients (mean age 47.3 ± 12.9 years, female n = 59, 66.3%) were included in the study. 46 relatives were allocated to the intervention website and 43 to the control website. Baseline Impact of Event Scale (IES) score was 27.5 ± 12.7. Overall, 50% showed clinically relevant PTSD symptoms at baseline. Mean IES score for the primary endpoint (~ 30 days after inclusion, T1) was 24 ± 15.8 (intervention 23.9 ± 17.9 vs. control 24.1 ± 13.5, p = 0.892). Hospital Anxiety and Depression Scale (HADS - Deutsch (D)) score at T1 was 12.2 ± 6.1 (min. 3, max. 31) and did not differ between groups. Use of the website differed between the groups (intervention min. 1, max. 14 vs. min. 1, max. 3; total 1386 "clicks" on the website, intervention 1021 vs. control 365). Recruitment was prematurely stopped in February 2020 due to coronavirus disease 2019 (COVID-19). CONCLUSION: Family members of critically ill patients often have significant PTSD symptoms and online information on critical illness did not result in reduced PTSD symptoms.

Prioritizing information topics for relatives of critically ill patients : Cross-sectional survey among intensive care unit relatives and professionals.

A patient's admission to an intensive care unit (ICU) has a significant impact on family members and other relatives. In order for them to be able to cope with such a stressful situation, the availability of appropriate understandable and accessible information is crucial. The information asymmetry between relatives and medical professionals may adversely affect satisfaction of relatives and their risk of subsequent anxiety, depression and stress symptoms. The aim of this study was therefore to understand which topics are most important to the relatives of ICU patients and to quantify the perceptions of medical professionals regarding the information needs of relatives. A cross-sectional survey was conducted in 2015. The survey had 42 questions, such as 'diagnosis', 'treatment', 'comfort', 'family' and 'end of life'. In total, the survey was handed out to four different groups. A total of 336 persons answered the survey (26 relatives, 28 ICU physicians, 202 ICU nurses and 80 ICU medical professionals in a closed Facebook© group [Facebook, Menlo Park, California, USA]). Relatives ranked the five most important topics as follows: 'recent events (crisis)', 'my participation', 'contamination in hospital', 'physical pain', and 'probability'. Several significant differences (p<0.001) were detected, for example for the topics fever, medication, recent events (crisis), appointments, relapse, and investigations. Even the topic with the lowest ranking (religion) had a score of 3.15 (min. 1.00, max. 5.00) among relatives. The ICU professionals appear to have divergent opinions regarding the most important topics for ICU relatives as compared to relatives themselves.

Training echo state networks for rotation-invariant bone marrow cell classification.

The main principle of diagnostic pathology is the reliable interpretation of individual cells in context of the tissue architecture. Especially a confident examination of bone marrow specimen is dependent on a valid classification of myeloid cells. In this work, we propose a novel rotation-invariant learning scheme for multi-class echo state networks (ESNs), which achieves very high performance in automated bone marrow cell classification. Based on representing static images as temporal sequence of rotations, we show how ESNs robustly recognize cells of arbitrary rotations by taking advantage of their short-term memory capacity. The performance of our approach is compared to a classification random forest that learns rotation-invariance in a conventional way by exhaustively training on multiple rotations of individual samples. The methods were evaluated on a human bone marrow image database consisting of granulopoietic and erythropoietic cells in different maturation stages. Our ESN approach to cell classification does not rely on segmentation of cells or manual feature extraction and can therefore directly be applied to image data.

A Smartwatch-Based Assistance System for the Elderly Performing Fall Detection, Unusual Inactivity Recognition and Medication Reminding.

The growing number of elderly people in our society makes it increasingly important to help them live an independent and self-determined life up until a high age. A smartwatch-based assistance system should be implemented that is capable of automatically detecting emergencies and helping elderly people to adhere to their medical therapy. Using the acceleration data of a widely available smartwatch, we implemented fall detection and inactivity recognition based on a smartphone connected via Bluetooth. The resulting system is capable of performing fall detection, inactivity recognition, issuing medication reminders and alerting relatives upon manual activation. Though some challenges, like the dependence on a smartphone remain, the resulting system is a promising approach to help elderly people as well as their relatives to live independently and with a feeling of safety.

Development of a generic monitoring application by the example of coronary artery disease.

The demographic change, the social structure and the development of new and complex therapies for diseases are responsible for a permanent rise of health expenditure. To guarantee affordable health care, AAL (Ambient Assisted Living) is one possibility. Our objective is to show that it is possible to create remote support for caregivers by physicians even with widely available and versatile hardware. A single board computer currently only equipped with 6 different sensors to measure blood pressure, glucose level, pulse, oxygen saturation, temperature and ECG is used as a showcase for patients with e.g. coronary artery disease. We created a user-friendly local application that collects and stores all data and transmits them to a remote server as soon as an Internet connection becomes available. Additionally, a web-application gives access to remote physicians and nurses to support caregivers and patients. Patients keep the overview of their vital signs, caregivers, nurses and physician can be alerted on demand or check the patient's data anytime to give advice. We conclude, that projects like this are showcases for the usage of technology that can foster AAL due to the focus on widely available and cheap, versatile equipment.

Development and evaluation of a web-based application for digital findings and documentation in physiotherapy education.

Findings in physiotherapy have standardized approaches in treatment, but there is also a significant margin of differences in how to implement these standards. Clinical decisions require experience and continuous learning processes to consolidate personal values and opinions and studies suggest that lecturers can influence students positively. Recently, the study course of Physiotherapy at the University of Applied Science in Graz has offered a paper based finding document. This document supported decisions through the adaption of the clinical reasoning process. The document was the starting point for our learning application called "EasyAssess", a Java based web-application for a digital findings documentation. A central point of our work was to ensure efficiency, effectiveness and usability of the web-application through usability tests utilized by both students and lecturers. Results show that our application fulfills the previously defined requirements and can be efficiently used in daily routine largely because of its simple user interface and its modest design. Due to the close cooperation with the study course Physiotherapy, the application has incorporated the various needs of the target audiences and confirmed the usefulness of our application.

Visualization of CDA laboratory reports and long term trends as a possible EHR application for patients and physicians.

To increase the patient's acceptance of electronic health records and understanding for their laboratory findings a web application was developed which presents all parameters and possible deviations of standard values in a clear way and visualizes the time based trend of all recorded parameters graphically. Documents corresponding to the Clinical document architecture (CDA) R2 laboratory reports standard and a rapid prototyping framework called Groovy on Grails were used. This work shows, that it is possible to create a useful, standards based tool for patients and physicians with comparatively few resources - an application that could be in similar form a part of an electronic Health Record (EHR) system like the Austrian electronic Health Record (ELGA).

Improving emergency medical dispatching with emphasis on mass-casualty incidents.

Dispatching ambulances is a demanding and stressful task for dispatchers. This is especially true in case of mass-casualty incidents. Therefore, the aim of this work was to investigate if and to what extent the dispatch operation of the Red Cross Salzburg can be optimized on such occasions with a computerized system. The basic problem of a dynamic multi-vehicle Dial-a-Ride Problem with time windows was enhanced according to the requirements of the Red Cross Salzburg. The general objective was to minimize the total mileage covered by ambulances and the waiting time of patients. Furthermore, in case of emergencies suitable adaptions to a plan should be carried out automatically. Consequently, the problem is solved by using the Adaptive Large Neighborhood Search. Evaluation results indicate that the system outperforms a human dispatcher by between 2.5% and 36% within 1 minute of runtime concerning total costs. Moreover, the system's response time in case that a plan has to be updated is less than 1 minute on average.

A learning rule for very simple universal approximators consisting of a single layer of perceptrons.

One may argue that the simplest type of neural networks beyond a single perceptron is an array of several perceptrons in parallel. In spite of their simplicity, such circuits can compute any Boolean function if one views the majority of the binary perceptron outputs as the binary output of the parallel perceptron, and they are universal approximators for arbitrary continuous functions with values in [0,1] if one views the fraction of perceptrons that output 1 as the analog output of the parallel perceptron. Note that in contrast to the familiar model of a "multi-layer perceptron" the parallel perceptron that we consider here has just binary values as outputs of gates on the hidden layer. For a long time one has thought that there exists no competitive learning algorithm for these extremely simple neural networks, which also came to be known as committee machines. It is commonly assumed that one has to replace the hard threshold gates on the hidden layer by sigmoidal gates (or RBF-gates) and that one has to tune the weights on at least two successive layers in order to achieve satisfactory learning results for any class of neural networks that yield universal approximators. We show that this assumption is not true, by exhibiting a simple learning algorithm for parallel perceptrons - the parallel delta rule (p-delta rule). In contrast to backprop for multi-layer perceptrons, the p-delta rule only has to tune a single layer of weights, and it does not require the computation and communication of analog values with high precision. Reduced communication also distinguishes our new learning rule from other learning rules for parallel perceptrons such as MADALINE. Obviously these features make the p-delta rule attractive as a biologically more realistic alternative to backprop in biological neural circuits, but also for implementations in special purpose hardware. We show that the p-delta rule also implements gradient descent-with regard to a suitable error measure-although it does not require to compute derivatives. Furthermore it is shown through experiments on common real-world benchmark datasets that its performance is competitive with that of other learning approaches from neural networks and machine learning. It has recently been shown [Anthony, M. (2007). On the generalization error of fixed combinations of classifiers. Journal of Computer and System Sciences 73(5), 725-734; Anthony, M. (2004). On learning a function of perceptrons. In Proceedings of the 2004 IEEE international joint conference on neural networks (pp. 967-972): Vol. 2] that one can also prove quite satisfactory bounds for the generalization error of this new learning rule.