En central del i patientsäkerhetsarbetet inom hälso- och sjukvård är att personalen rapporterar risker och avvikelser som upptäcks i verksamheten. En avvikelse är ett faktum när verksamheten inte når upp till satta mål och krav utifrån lagar, föreskrifter och beslut och kan vara allt från småstrul som tar tid från patientarbete till en allvarlig skada som inträffar (SOSFS 2011:9) Att ständigt arbeta med förbättringar är ett måste för att anpassa verksamheten utifrån nya uppkomna behov krav och förutsättningar. Att ständigt identifierade brister och ta tillvara på medarbetares förslag till förbättring, samt sträva efter en samsyn är förutsättningar som krävs för att verksamheten ständigt ska förbättras. Studiens syfte var att undersöka hur man använder rapporterade avvikelser som underlag till att förbättra sin verksamhet samt vilka förutsättningar som krävs för att arbeta med avvikelser på ett systematiskt sätt. Studien utgörs av flervalsmetodforskning där data samlats in via semistrukturerade intervjuer med tre vårdenhetschefer inom en verksamhet för att få en fördjupad förståelse av förutsättningar och utmaningar med att använda avvikelser för att förbättra sin verksamhet. För att få en förstärkt och bredare bild av studiens frågeställningar kompletterades datainsamlingen med en enkät där 60 medarbetare inom fyra vårdenheter svarat. Empirisk data analyserades genom tematisk analys. I resultatet framkommer fem övergripande tema som både chefer och medarbetare identifierat som viktiga förutsättningar för att förbättra sin verksamhet med grund i avvikelser; Samsyn, Struktur, Chefens stöd, Lärande och Medarbetares delaktighet som även inkluderar fler underliggande identifierade subteman. Utmaningar identifierades som tre övergripande tema; Fokus på rätt sak, Prioritering av tid samt Integrera förbättringsarbetet i det dagliga arbetet. En av frågeställningarna handlade också om hur medarbetares förslag till förbättring fångas upp och resulterade i intressanta reflektioner både hos medarbetare och chefer. / A central aspect of patient safety work within healthcare is for the staff to report risks and deviations. A deviation occurs when healthcare fails to meet set goals and requirements according to laws, regulations, and decisions. It may include minor issues that take time away from patient, to serious incidents causing harm (SOSFS 2011:9). Continuous improvements are necessary to adapt the care unit based on new emerging needs and conditions. Constantly identifying shortcomings, make use of employees' improvement suggestions, and striving for consensus are prerequisites for continuous improvement in care units. This study aimed to investigate how reported deviations are utilized as a basis for improving operations, as well as the prerequisites for working with deviations in a systematic manner. The study was designed as a mixed-methods research design, with data collected through semi-structured interviews with three unit managers within a healthcare organization to gain an deepened understanding of the prerequisites and challenges associated with using deviations to improve care units. Data collection was complemented with a survey in which 60 employees from four units responded to enhance and broaden the understanding of the research questions. Empirical data were analyzed through thematic analysis. The results showed five overarching themes that both managers and employees identified as important prerequisites for improving operations based on deviations: Consensus, Structure, Managerial Support, Learning, and Employee involvement, including several underlying subthemes. Challenges were identified in three overarching themes: Focus on the right things, Time prioritization, and Integrating improvement work into daily operations. One of the research questions also focused on how employees' improvement suggestions are captured, resulting in interesting reflections among both employees and managers

京都大学 / 新制・課程博士 / 博士(理学) / 甲第24386号 / 理博第4885号 / 新制||理||1699(附属図書館) / 京都大学大学院理学研究科数学・数理解析専攻 / (主査)教授 平岡 裕章, 教授 COLLINS Benoit Vincent Pierre, 教授 坂上 貴之 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

The presented thesis describes the historical evolution of growth and laboratory reference values and the methods for their creation – leading finally to a family of methods applied in the WHO Multicentre Growth Reference Study (MGRS). The following part describes these methods, their assumptions, and model diagnostics. The original article at the beginning of part III combines these methods with resampling to be able to use LMS -type methods on data containing different dependencies like follow-up measures and family relationships. This method has been applied in the estimation of reference values of several laboratory values in the context of the LIFE child study. Three papers are already published and are also presented. In part III. The next section concentrates on the accompanying R package childsds. A short summary and outlook conclude the work.:I Introduction 1 Introduction 1.1 Standardization and Reference Values 1.2 The Scope of this Thesis 2 Historical Aspects 2.1 Development of Basic Concepts 2.2 Growth References and Growth Charts 2.3 Laboratory Reference Values II Methods 3 Method Selection 3.1 Nomenclature 3.2 The Choice of Statistical Method 3.3 GAMLSS 3.4 LMS-type Methods 4 Model Diagnostics 4.1 Normalized Quantile Residuals 4.2 Diagnostic Plots 4.3 Numerical Statistics III Original Articles & R Pakackage 5 Original Article 6 Articles Using the Proposed Method 7 R Package childsds 7.1 The Example Data 7.2 The Fitting Functions 7.3 The Classes 7.4 The Back-transformation Function 7.5 Convenience Functions 7.6 Collection of Reference Values 8 Summary and Outlook 8.1 Summary 8.2 SDS Values and Quality Control 8.3 The R Package / Die vorliegende Arbeit soll die Entwicklung und Bedeutung von Referenzwerten im pädiatrischen Kontext beschreiben. Der zugehörige Artikel beschreibt die Kombination der von der WHO empfohlene Methode und Resampling, um die Herleitung von Referenzwerten auch im Fall von Messwiederholungen und anderen Abhängigkeiten zu ermöglichen. Drei weitere Artikel, in denen die Methode angewendet wurde und eine kurze Vorstellung des zugehörigen, begleitend entwickelten R Pakets (statistische Software) bilden Teil III der Arbeit.:I Introduction 1 Introduction 1.1 Standardization and Reference Values 1.2 The Scope of this Thesis 2 Historical Aspects 2.1 Development of Basic Concepts 2.2 Growth References and Growth Charts 2.3 Laboratory Reference Values II Methods 3 Method Selection 3.1 Nomenclature 3.2 The Choice of Statistical Method 3.3 GAMLSS 3.4 LMS-type Methods 4 Model Diagnostics 4.1 Normalized Quantile Residuals 4.2 Diagnostic Plots 4.3 Numerical Statistics III Original Articles & R Pakackage 5 Original Article 6 Articles Using the Proposed Method 7 R Package childsds 7.1 The Example Data 7.2 The Fitting Functions 7.3 The Classes 7.4 The Back-transformation Function 7.5 Convenience Functions 7.6 Collection of Reference Values 8 Summary and Outlook 8.1 Summary 8.2 SDS Values and Quality Control 8.3 The R Package

Machine tools need to be tested to check they behave correctly while machining. A bigger production (capacity), improving the accuracy in final products (quality) or reducing costs are some of the main goals. Reducing the maintenance of the machines, their nonproductive time and a higher quality of the final parts, have a strong influence in the costs. Even machines built to the same specification present different properties and behaviours, which leads into a reduction of the flexibility when it comes to move operations among them. This study is focused on the measurement and modelling of four machine tools (M1, M2, M3 and M4), with same specifications, from the static point of view. Methods used to measure all these properties are the circular tests under loaded conditions, with the device Loaded Double Ball Bar (LDBB), which measures positional accuracy and static stiffness. Different pressures (0.5, 1, 3, 5, 6, 7 bar) and locations for the LDBB have been used.   After the tests were performed, all the machines shown good properties with the detail of M4, which had a lower stiffness than the other three ones due to its table attachments. / Maskinverktyg måste testas för att kontrollera att de beter sig korrekt vid bearbetning. En större produktion (kapacitet), förbättra noggrannheten i slutprodukterna (kvalitet) eller sänka kostnaderna är några av huvudmålen. Att minska underhållet av maskinerna, deras icke-produktiva tid och en högre kvalitet på de slutliga delarna har ett starkt inflytande i kostnaderna. Även maskiner konstruerade för samma specifikation presenterar olika egenskaper och beteenden, vilket leder till en minskning av flexibiliteten när det gäller att flytta verksamheten bland dem. Denna studie är inriktad på mätning och modellering av fyra verktygsmaskiner (M1, M2, M3 och M4), med samma specifikationer, ur statisk synvinkel. Metoder som används för att mäta alla dessa egenskaper är de cirkulära testerna under laddade förhållanden, med enheten Loaded Double Ball Bar (LDBB), som mäter positionsnoggrannhet och statisk styvhet. Olika tryck (0,5, 1, 3, 5, 6, 7 bar) och platser för LDBB har använts.   Efter testerna visade alla maskiner goda egenskaper med detaljerna i M4, som hade en lägre styvhet än de andra tre på grund av dess bordsfästanordningar.

In this thesis, we study how sovereign credit default swaps are able to measure systemic risk as well as how they can be used to construct optimal portfolios to minimize risk. We define the clustering coefficient as a proxy for systemic risk and design an optimization problem with the goal of minimizing the mean absolute deviation of the clustering coefficient on a group of nine European countries. Additionally, we define a metric we call the diversity score that measures the diversification of any given portfolio. We solve this problem for a baseline set of parameters, then spend the remainder of the thesis modifying these parameters to investigate how the optimal solution and diversity score are impacted.

We study the fluctuation effects in the seminal cyclic predator-prey model in population dynamics due to Robert May and Warren Leonard both in the zero-dimensional and two-dimensional spatial version. We compute the mean time to extinction of a stable set of coexisting populations driven by large fluctuations. We see that the contribution of large fluctuations to extinction can be captured by a quasi-stationary approximation and the Wentzel–Kramers–Brillouin (WKB) eikonal ansatz. We see that near the Hopf bifurcation, extinctions are fast owing to the flat non-Gaussian distribution whereas away from the bifurcation, extinctions are dominated by large fluctuations of the fat tails of the distribution. We compare our results to Gillespie simulations and a single-species theoretical calculation. In addition, we study the spatio-temporal pattern formation of the stochastic May--Leonard model through the Doi-Peliti coherent state path integral formalism to obtain a coarse-grained Langevin description, i.e. the Complex Ginzburg Landau equation with stochastic noise in one complex field. We see that when one restricts the internal reaction noise to small amplitudes, one can obtain a simple form for the stochastic noise correlations that modify the Complex Ginzburg Landau equation. Finally, we study the effect of coupling a spatially extended May--Leonard model in two dimensions with symmetric predation rates to one with asymmetric rates that is prone to reach extinction. We show that the symmetric region induces otherwise unstable coexistence spiral patterns in the asymmetric May--Leonard lattice. We obtain the stability criterion for this pattern induction as we vary the strength of the extinction inducing asymmetry. This research was sponsored by the Army Research Office and was accomplished under Grant Number W911NF-17-1-0156. / Doctor of Philosophy / In the field of ecology, the cyclic predator-prey patterns in a food web are relevant yet independent to the hierarchical archetype. We study the paradigmatic cyclic May--Leonard model of three species, both analytically and numerically. First, we employ well--established techniques in large-deviation theory to study the extinction of populations induced by large but rare fluctuations. In the zero--dimensional version of the model, we compare the mean time to extinction computed from the theory to numerical simulations. Secondly, we study the stochastic spatial version of the May--Leonard model and show that for values close to the Hopf bifurcation, in the limit of small fluctuations, we can map the coarse-grained description of the model to the Complex Ginsburg Landau Equation, with stochastic noise corrections. Finally, we explore the induction of ecodiversity through spatio-temporal spirals in the asymmetric version of the May--Leonard model, which is otherwise inclined to reach an extinction state. This is accomplished by coupling to a symmetric May-Leonard counterpart on a two-dimensional lattice. The coupled system creates conditions for spiral formation in the asymmetric subsystem, thus precluding extinction.

No description available.

På Karolinska Universitetssjukhuset har det återkommande rapporterats om flödesavvikelser hos infusionspumpar. En hypotes är att pumpanvändare drar i infusionsaggregates silikondel när en infusion förbereds och att detta ger upphov till flödesavvikelser. Detta projekt fokuserar på att utreda om töjning av slangsetets silikondel kan orsaka flödesavvikelser. Projektet utvärderade tre grupper av slangset där två av grupperna bestod av infusionsaggregat vars silikondel töjdes till olika längder och en grupp fungerade som referensgrupp. Resultatet visar att silikondelen förlängs vid töjning samt att flödesavvikelser uppstår om töjningen är tillräckligt stor. Trots att flödesavvikelserna påvisats så bedöms töjningen som utförts vara mycket större än den töjning som utförts av vårdpersonalen. Därför bedömmer vi att flödesavvikelserna som rapporterats på Karolinska Universitetssjukhuset inte beror på denna typ av felhantering av slangset. / At Karolinska University Hospital, recurrent reports of flow deviations in infusion pumps have been observed. One hypothesis is that pump users pull the silicone component of the administration set when preparing an infusion, leading to flow deviations. This project focuses on investigating whether elongation of the silicone component of the administration set can cause flow deviations. The project evaluated three groups of administration sets, where two groups consisted of administration sets which were elongated to different lenghts, while one group served as reference group. Results show that elongation of the silicone component occurs and that flow deviations occur if the elongation is large enough. However, despite the observed flow deviations, the elongation performed in the study is considered much larger than what healthcare personnel typically apply. Therefore, we find it unlikely that the reported flow deviations at Karolinska University Hospital are due to mishandling of the administration set.

The goal of this project is to find a suitable Few-Shot Learning (FSL) model that can be used in a fault detection system for use in an industrial setting. A dataset of Printed Circuit Board (PCB) images has been created to train different FSL models. This dataset is meant for evaluating FSL models in the specialized setting of fault detection in PCB manufacturing. FSL is a part of deep learning that has seen a large amount of development recently. Few-shot learning allows neural networks to learn on small datasets. In this thesis, various state-of-the-art FSL algorithms are implemented and tested on the custom PCB dataset. Different backbones are used to establish a benchmark for the tested FSL algorithms on three different datasets. Those datasets are ImageNet, PCB Defects, and the created PCB dataset. Our results show that ProtoNets combined with ResNet12 backbone achieved the highest accuracy in two test scenarios. In those tests, the model combination achieved 87.20%and 92.27% in 1-shot and 5-shot test scenarios, respectively. This thesis presents a Few-Shot Anomaly Detection (FSAD) model based on Vision Transformers (ViT). The model is compared to the state-of-the-art FSAD model DevNet on the MVTec-AD dataset. DevNet and ViT are chosen for comparison because they both approach the problem by dividing images into patches. How the models handle the image patches is however very different. The results indicate that ViT Deviation does not obtain as high AUC-ROC and AUC-PR scores as DevNet. This is because of the use of the very deep ViT architecture in the ViT Deviation model. A shallower transformer-based model is believed to be better suited for FSAD. Improvements for ViT Deviation are suggested for future work. The most notable suggested improvement is the use of the FS-CT architecture as a FSAD model because of the high accuracy it achieves in classification. / Målet med detta projekt är att hitta en lämplig Few-Shot Learning(FSL) modell som kan användas i ett feldetekteringssystem för användning i en industriell miljö. Ett dataset av Printed Circuit Board(PCB) bilder har skapats för att träna olika FSL-modeller. Detta datasetär avsedd för att utvärdera FSL-modeller i det specialiserade områdetfeldetektering vid PCB-tillverkning. FSL är en del av djupinlärningsom har utvecklats mycket den senaste tiden. FSL tillåter neuralanätverk att lära sig på små datamängder.I detta examensarbete implementeras och testas olika state-of-theart FSL algoritmer på det anpassade PCB-datasetet. Olika ryggradsmodeller används för att upprätta ett riktmärke för de testade FSL-algoritmernapå tre olika dataset. Dessa dataset är ImageNet[6], PCB Defects[14]och det skapade PCB-datasetet. Våra resultat visar att ProtoNets ikombination med ResNet12-ryggraden uppnådde den högsta noggrannheten i två testscenarier. I dessa tester uppnådde modellkombinationen 87,20% och 92,27% i testscenarier med 1-shot respektive5-shot.Detta examensarbete presenterar en Few-Shot Anomaly Detectionmodell (FSAD) baserad på Vision Transformers (ViT). Modellen jämförs med FSAD-modellen DevNet på MVTec-AD-datasetet. DevNetoch ViT väljs för jämförelse eftersom de båda angriper problemetgenom att dela upp bilder i mindre lappar. Hur modellerna hanterarlapparna är dock väldigt olika. Resultaten indikerar att ViT-Deviationinte får lika hög AUC-ROC och AUC-PR som DevNet. Detta beror påanvändningen av den mycket djupa ViT-arkitekturen i ViT Deviationmodellen. En grundare ViT-baserad modell tros vara bättre lämpadför FSAD. Förbättringar för ViT-Deviation föreslås för framtida arbete.Den mest anmärkningsvärda föreslagna förbättringen är användningen av FS-CT-arkitekturen som en FSAD-modell på grund av de lovande resultaten den uppnår i klassificering.

Knowledge of battery health is very important. It provides insight into the capacity of a given system and allows the operators to plan ahead more efficiently. But measuring state of health (SoH) of a battery is difficult, and takes time. More importantly, the battery needs to be taken out of operation to be analysed correctly. This paper aims to evaluate a proposed linear regression method for predicting battery health, based on easily acquired operational data. The main predictor being voltage deviation, a characteristic of battery voltages during charge/discharge cycles. Using this method, the only time a battery would need to be extracted is to gather training data. Then, the model could be used for similar batteries to predict their SoH. Meaning those systems would never need to be halted, increasing productivity. The results of this paper is that the data used was not suitable for linear regression. There were problems with heteroskedasticity and non-normality of the residuals, but mainly the estimated parameter for the relationship between voltage deviation and SoH ran contrary to established theory. Which could not be overlooked. Therefore, the estimated models should not be used to predict SoH. To accomplish the goal of accurate SoH prediction, more research should be conducted and a better sample used.