We investigate the interplay of rotor-angle and voltage stability in electric power systems. To this end, we carry out a local stability analysis of the third-order model which entails the classical power-swing equations and the voltage dynamics. We provide necessary and sufficient stability conditions and investigate different routes to instability. For the special case of a two-bus system, we analytically derive a global stability map.

IntroductionThe acute respiratory distress syndrome (ARDS) is a highly relevant entity in critical care with mortality rates of 40%. Despite extensive scientific efforts, outcome-relevant therapeutic measures are still insufficiently practised at the bedside. Thus, there is a clear need to adhere to early diagnosis and sufficient therapy in ARDS, assuring lower mortality and multiple organ failure.Methods and analysisIn this quality improvement strategy (QIS), a decision support system as a mobile application (ASIC app), which uses available clinical real-time data, is implemented to support physicians in timely diagnosis and improvement of adherence to established guidelines in the treatment of ARDS. ASIC is conducted on 31 intensive care units (ICUs) at 8 German university hospitals. It is designed as a multicentre stepped-wedge cluster randomised QIS. ICUs are combined into 12 clusters which are randomised in 12 steps. After preparation (18 months) and a control phase of 8 months for all clusters, the first cluster enters a roll-in phase (3 months) that is followed by the actual QIS phase. The remaining clusters follow in month wise steps. The coprimary key performance indicators (KPIs) consist of the ARDS diagnostic rate and guideline adherence regarding lung-protective ventilation. Secondary KPIs include the prevalence of organ dysfunction within 28 days after diagnosis or ICU discharge, the treatment duration on ICU and the hospital mortality. Furthermore, the user acceptance and usability of new technologies in medicine are examined. To show improvements in healthcare of patients with ARDS, differences in primary and secondary KPIs between control phase and QIS will be tested.Ethics and disseminationEthical approval was obtained from the independent Ethics Committee (EC) at the RWTH Aachen Faculty of Medicine (local EC reference number: EK 102/19) and the respective data protection officer in March 2019. The results of the ASIC QIS will be presented at conferences and published in peer-reviewed journals.Trial registration numberDRKS00014330.

Computational simulation of virtual patients reduces dataset bias and improves machine learning-based detection of ARDS from noisy heterogeneous ICU datasets. Mechanistic virtual patient modeling is used to infer model-derived parameters on individual patients, significantly reducing biases introduced by learning from heterogeneous datasets and allowing improved discovery of patient cohorts driven exclusively by medical conditions.

Machine learning (ML) models are developed on a learning dataset covering only a small part of the data of interest. If model predictions are accurate for the learning dataset but fail for unseen data then generalization error is considered high. This problem manifests itself within all major sub-fields of ML but is especially relevant in medical applications. Clinical data structures, patient cohorts, and clinical protocols may be highly biased among hospitals such that sampling of representative learning datasets to learn ML models remains a challenge. As ML models exhibit poor predictive performance over data ranges sparsely or not covered by the learning dataset, in this study, we propose a novel method to assess their generalization capability among different hospitals based on the convex hull (CH) overlap between multivariate datasets. To reduce dimensionality effects, we used a two-step approach. First, CH analysis was applied to find mean CH coverage between each of the two datasets, resulting in an upper bound of the prediction range. Second, 4 types of ML models were trained to classify the origin of a dataset (i.e., from which hospital) and to estimate differences in datasets with respect to underlying distributions. To demonstrate the applicability of our method, we used 4 critical-care patient datasets from different hospitals in Germany and USA. We estimated the similarity of these populations and investigated whether ML models developed on one dataset can be reliably applied to another one. We show that the strongest drop in performance was associated with the poor intersection of convex hulls in the corresponding hospitals' datasets and with a high performance of ML methods for dataset discrimination. Hence, we suggest the application of our pipeline as a first tool to assess the transferability of trained models. We emphasize that datasets from different hospitals represent heterogeneous data sources, and the transfer from one database to another should be performed with utmost care to avoid implications during real-world applications of the developed models. Further research is needed to develop methods for the adaptation of ML models to new hospitals. In addition, more work should be aimed at the creation of gold-standard datasets that are large and diverse with data from varied application sites.