The Flatland competition aimed at finding novel approaches to solve the vehicle re-scheduling problem (VRSP). The VRSP is concerned with scheduling trips in traffic networks and the re-scheduling of vehicles when disruptions occur, for example the breakdown of a vehicle. While solving the VRSP in various settings has been an active area in operations research (OR) for decades, the ever-growing complexity of modern railway networks makes dynamic real-time scheduling of traffic virtually impossible. Recently, multi-agent reinforcement learning (MARL) has successfully tackled challenging tasks where many agents need to be coordinated, such as multiplayer video games. However, the coordination of hundreds of agents in a real-life setting like a railway network remains challenging and the Flatland environment used for the competition models these real-world properties in a simplified manner. Submissions had to bring as many trains (agents) to their target stations in as little time as possible. While the best submissions were in the OR category, participants found many promising MARL approaches. Using both centralized and decentralized learning based approaches, top submissions used graph representations of the environment to construct tree-based observations. Further, different coordination mechanisms were implemented, such as communication and prioritization between agents. This paper presents the competition setup, four outstanding solutions to the competition, and a cross-comparison between them.
Abstract At our tertiary children’s hospital, infections with newly detected methicillin-resistant Staphylococcus aureus (MRSA) among children attending primary (age 6–12 years) and secondary school (age 13–16 years) nearly doubled in 2018 compared to previous years. This observation initiated an epidemiological outbreak investigation including phenotypic (susceptibility testing) and genotypic (whole genome sequencing) characterization of the isolates. In addition, a cross-sectional study was conducted to determine source of the outbreak, colonization frequency and to identify risk factors for transmission using a questionnaire. As a result, 49 individuals were detected with 57 corresponding isolates. Based on the case definition combined with whole genome sequencing, a core cluster was identified that shared common genetic features and a similar antimicrobial susceptibility pattern (efflux-mediated macrolide resistance, tetracycline susceptibility along with presence of Panton-Valentine leukocidin). Epidemiologic evaluation identified a distinct school as a common risk factor. However, the source of the clustered infections within that school could not be further specified. No further cases could be detected after decolonization of infected and colonized children.
False-positive results in the diagnostic of meningitis and encephalitis pose important challenges. This study aimed to determine false-positive rates for Haemophilus influenzae in cerebrospinal fluids evaluated by the BioFire FilmArray ® Meningitis/Encephalitis Panel. We conducted a retrospective study of all H. influenzae -positive FilmArray ® . Meningitis/Encephalitis Panel results from June 2016 to October 2019 in two Swiss university hospitals. Cases were classified as true positive, likely true-positive, and likely false-positive results according to cerebrospinal fluid culture, H. influenzae -specific quantitative real-time PCR (qPCR), and Gram staining, as well as culture of other materials. We performed 3,082 panels corresponding to 2,895 patients: 0.6% of the samples (18/3,082) were positive for H. influenzae . Culture and H. influenzae -specific qPCR were performed on 17/18 (94.4%) and 3/18 (16.7%) cerebrospinal fluid samples, respectively; qPCR was negative in all cases. Among 17 samples sent for culture, 10 concerned patients were not treated with antibiotics prior to lumbar puncture. Only 1/17 revealed growth of H. influenzae and was classified as a true positive. We further classified 3/18 (16.7%) cases with the identification of Gram-negative rods in the cerebrospinal fluid or positive blood cultures for H. influenzae as likely true-positive and 14/18 (77.8%) cases as likely false-positive. Diagnostic results should always be interpreted together with the clinical presentation, cerebrospinal fluid analysis, and other available microbiological results. All H. influenzae -positive results should be viewed with special caution and a H. influenzae- specific qPCR should be systematically considered.
Background An accurate and timely identification of bacterial species is critical in clinical diagnostics. Species identification allows a potential first adaptation of empiric antibiotic treatments before the resistance profile is available. Matrix assisted Laser Desorption Ionization Time of Flight mass spectrometry (MALDI-TOF MS) is a widely used method for bacterial species identification. However, important challenges in species identification remain. These arise from (i) incomplete databases, (ii) close relatedness of species of interest, and (iii) spectral quality, which is currently vaguely defined. Methods We selected 47 clinically relevant bacterial isolates from 39 species, which can be challenging to identify by MALDI-TOF MS. We measured these isolates under various analytical conditions on two MALDI-TOF MS systems. First, we identified spectral features, which were associated with correct species identification in three different databases. Considering these features, we then systematically compared spectra produced with three different sample preparation protocols. In addition, we varied quantities of bacterial colony material applied and bacterial colony age. Results We identified (i) the number of ribosomal marker peaks detected, (ii) the median relative intensity of ribosomal marker peaks, (iii) the sum of the intensity of all detected peaks, (iv) a high measurement precision, and (v) reproducibility of peaks to act as good proxies of spectral quality. We found that using formic acid, measuring bacterial colonies at a young age, and frequently calibrating the MALDI-TOF MS device increase mass spectral quality. We further observed significant differences in spectral quality between different bacterial taxa and optimal measurement conditions vary per taxon. Conclusion We identified and applied quality measures for MALDI-TOF MS and optimized spectral quality in routine settings. Phylogenetic marker peaks can be reproducibly detected and provide an increased resolution and the ability to distinguish between challenging species such as those within the Enterobacter cloacae complex, Burkholderia cepacia complex, or viridans streptococci.
. Intra-articular fractures are commonly scanned by computer tomography. But even complex fractures located in the bone shaft are usually only diagnosed on 2D radiographs. The surgeon must therefore himself interpret the 2D pictures to determine fracture’s 3D configuration. Minimally invasive procedures that do not expose the fracture during surgery are increasing in popularity. Consequently, young surgeons will hardly ever see the true 3D configuration of the shaft fractures they are treating. This lack of opportunity to learn the correspondence between 2D radiogram and actual 3D fracture morphology is a cause of concern in surgical education. In this study we suggest a method to improve comprehension of 3D fracture morphology. We propose creating a concatenated 3D bone model by merging a short scan of the pertinent fracture with a generic intact bone model. Thus a full-length concatenated model is obtained that depicts both fracture and surrounding intact bone.
From 1974 to 1989, 37,392 patients were admitted to the divisions of general internal medicine of the CHDM hospitals. 19,082 of them were treated with a minor analgesic or an NSAID. In 95 of the exposed patients, an allergic or a pseudoallergic reaction to one or two of these drugs was observed. From 1981 to 1990, general practitioners, hospitals and the pharmaceutical industry reported to SANZ 158 individual cases with comparable reactions to 175 exposures of the same kind. Of the 15 different syndromes and symptoms registered in both institutions, most were reactions of the skin, mainly the usual maculopapular exanthemas (rash), urticaria and angioedema. In the CHDM, allergic or pseudoallergic reactions were observed in 0.23% of patients exposed to minor analgesics (including ASA preparations on a daily dose up to 1.0 g and pyrazolones, mainly metamizole, propyphenazone) and in 0.81% of patients exposed to NSAIDs (including the pyrazolone oxyphenbutazone). In the experience of the Comprehensive Hospital Drug Monitoring in Berne and St. Gallen (CHDM) and the Spontaneous Adverse Drug Reactions Center of Switzerland (SANZ).
