Gram-negative infection has been linked to hospital water although few studies have examined whether water systems are reservoirs of nosocomial pathogens. This study investigated longitudinal recovery of the opportunistic pathogens Pseudomonas aeruginosa, Stenotrophomonas maltophilia and Acinetobacter baumannii from water outlets of a haematology unit and evaluated Point-Of-Use Filtration (POU-F) as a control measure. In a two-year double cross-over trial, water samples and swabs were taken weekly from 39 showers/taps on the unit. Four study phases alternated between non-filtered (Phases 1 & 3), and filtered outlets (Phases 2 & 4) using Pall AquasafeTM 14-day filters. In Phases 1 & 3; 99% of 1396 samples yielded bacterial growth, with colonies generally too numerous to count. Target species were isolated from 22% of water samples (P. aeruginosa 14%; S. maltophilia 10%) and 10% of swabs. P. aeruginosa was particularly associated with handwash stations and S. maltophilia with showers. A. baumannii was not isolated. With POU-F; 22% of 1242 samples yielded bacterial growth (mean CFU/100ml ,4.6). S. maltophilia was isolated only once from water but never from outlet swabs. PCR typing identified clusters of isolates colonizing different outlets over time but no clear association between water and patient isolates was identified. The incidence of Gram negative infections remained low throughout the study. Without POU-F, water from taps/showers represented a source of bacteria including the target species. POU-F substantially reduced the frequency and number of target species from every outlet, and merits further investigation as an intervention to protect immunocompromised patients from opportunistic pathogens. / School of Life Sciences, University of Bradford and Pall Medical (Pall Europe Ltd)

The multi-million-dollar medical uniform industry has not utilized advancements in garment and textile technology that could positively impact the protection of healthcare professionals and patients. In most cases the uniforms meet basic requirements – they clothe the professional in a recognizable way. Little innovation in design, function and performance, has been applied to these garments. This is particularly evident in the case of the stereotypical white lab coat worn by many physicians, despite evidence indicating that these lab coats may carry contamination and play a role in the spread of deadly bacteria. Healthcare Associated Infections (HAIs) are among the most serious problems facing modern medical care, costing millions of lives and dollars annually worldwide. This research investigates the design and use of the physician’s lab coat – an immediately recognizable symbol of Western medicine. The research identifies the medical, functional, cultural and symbolic roles of the lab coat within the hospital environment and beyond, to the larger the global society. This thesis examines the extent to which the design of medical wear can impact the effect of hospital-acquired infections, support doctor/patient relationships and enhance the performance and behavior of the healthcare professional by envisioning a future lab coat which offers increased protection for physician and patient, aids in communication and enhances the performance of the doctor by utilizing digital technologies incorporated into the lab coat whereby the lab coat becomes the only tool necessary for the physician.

The spread of healthcare-associated infections (HAIs) and multi-drug resistance in healthcare networks is a major public health issue. Evaluating the role of inter-facility patient transfers that form the structure of these networks may provide insights on novel infection control measures. Identifying novel infection control strategies is especially important for multi-drug resistant pathogens such as Carbapenemase-producing Enterobacteriaceae (CPE) due to limited treatment options. The increasing use of inter-individual contact and inter-facility transfer network data in mathematical modelling of HAI spread has helped these models become more realistic; however, they remain limited to a few settings and pathogens. The main objectives of this thesis were two-fold: 1) to better understand the structure of the healthcare networks of France and their impact on HAI spread dynamics; and 2) to assess the role of transfers on the spread of CPE in France during the 2012 to 2015 period. The French healthcare networks are characterized by centralized patient flows towards hubs hospitals and a two-tier community clustering structure. We also found that networks of patients with HAIs form the same underlying structure as that of the general patient population. The number of CPE episodes have increased over time in France and projections estimate that the number of monthly episodes could continue to increase with seasonal peaks in October. The general patient network was used to show that, since 2012, patient transfers have played an increasingly important role over time in the spread of CPE in France. Multiple spreading events of CPE linked to patient transfers were also observed. Despite subtle differences in the flows of patients with an HAI and the general patient population, the general patient network may best inform novel infection control measures for pathogen spread. The structure of healthcare networks may help serve as a basis for novel infection control strategies to tackle HAIs in general but also CPE in particular. Key healthcare hubs in large metropoles and key patient flows connecting hospital communities at the local and regional level should be considered in the development of coordinated regional strategies to control pathogen spread in healthcare systems. / La propagation des infections nosocomiales (IN), notamment liées aux bactéries multi-résistantes, au sein du réseau des hôpitaux, est un grand enjeu de santé publique. L’évaluation du rôle joué par les transferts inter-établissements des patients sur cette propagation pourrait permettre l’élaboration de nouvelles mesures de contrôle. L’identification de nouvelles mesures de contrôle est particulièrement importante pour les bactéries résistantes aux antibiotiques comme les entérobactéries productrices de carbapenemase (EPC) pour lesquelles les possibilités de traitement sont très limitées. L’utilisation des données de réseaux de contact inter-individus et de transferts inter-établissement dans la modélisation mathématique ont rendu ces modèles plus proches de la réalité. Toutefois, ces derniers restent limités à quelques milieux hospitaliers et quelques pathogènes. La thèse a eu pour objectifs de 1) mieux comprendre la structure des réseaux hospitaliers français et leur impact sur la propagation des IN ; et 2) évaluer le rôle des transferts sur la propagation des EPC.Les réseaux hospitaliers français sont caractérisés par des flux de patients vers des hubs et par deux niveaux de communautés des hôpitaux. La structure du réseau de transfert des patients présentant une IN n’est pas différente de celle du réseau général de transfert des patients. Au cours des dernières années, le nombre d’épisode d’EPC a augmenté en France et les prédictions prévoient une poursuite de cette augmentation, avec des pics de saisonnalité en octobre. Ce travail a également montré que, depuis 2012, les transferts de patients jouent avec les années un rôle de plus en plus important sur la diffusion des EPC en France. Des évènements de propagation multiple liée aux transferts sont également de plus en plus souvent observés.En conséquence, la structure du réseau des hôpitaux pourrait servir de base pour la proposition des nouvelles stratégies de contrôles des IN en général, et des EPC en particulier. Les hôpitaux très connectés des grandes métropoles et les flux des patients entre les communautés locale et régionale doivent être considérés pour le développement de mesures de contrôle coordonnées entre établissements de santé.

Introduction. La détection précoce des infections par un système de surveillance efficace permet de mettre en œuvre des mesures de prévention et de contrôle adaptées. L’objectif de cette thèse était d’évaluer les performances d’un système de détection automatique, type syndromique, des patients à risque épidémique à partir des données du dossier médical informatisé des urgences. Population d’étude. 101001 patients ayant consulté aux urgences du groupement Nord des Hospices Civils de Lyon, entre le 01/06/2007 et le 31/03/2011, dont 10895 patients hospitalisés dans l’établissement à l’issue de la consultation. Méthode. Trois étapes ont été nécessaires. 1) Évaluation de la faisabilité d’utiliser les données structurées et textuelles, à l’aide d’une application de traitement automatisé des données textuelles. 2) Construction et évaluation d’algorithmes de repérage, pour les syndromes respiratoire, cutané et gastro-intestinal, de patients avec une infection à risque épidémique à partir des données du dossier médical informatisé des urgences. 3) Évaluation des données du dossier médical des urgences pour la détection d’épidémies communautaires de grippe, comparées aux données régionales de surveillance de la grippe. Résultats et Discussion. Cette thèse a montré que qu’il est possible de repérer des patients à risque épidémique avec une balance raisonnable entre la sensibilité et la spécificité pour des syndromes respiratoires et cutanés. Les algorithmes pour des syndromes gastro-intestinaux n'étaient pas assez spécifiques pour une utilisation de routine. Les données d’urgences ont permis aussi de détecter les épidémies communautaires dès le début de l’épidémie locale / Introduction. The early detection of the infections by an effective surveillance system allows implementing adapted measures of prevention and control. The objective of this thesis was to estimate the performances of an automatic system syndromic-like to detect the patients with potentially transmissible infectious diseases from the emergency department computerized medical record data. Study population. 101,001 adults, who were admitted to the emergency department and hospitalised of the North Hospital In University Hospital of Lyon, between 01/06/2007 and 30/03/2011. Method. Three steps were necessary. 1) Evaluation of the feasibility to use the structured and textual data with an application which automatically extracts and encodes information found in narrative reports. 2) Different algorithms were built for the detection of patients with infectious respiratory, cutaneous or gastrointestinal syndromes, and assessed. 3) Evaluation of the data of the electronic medical record of emergency department for the detection of flu community epidemics, compared with regional surveillance networks for flu. Results and discussion. This thesis showed that it is possible to detect patients with potentially transmissible infectious diseases with reasonable balance between sensitivity and specificity for respiratory and cutaneous syndromes. The algorithms for gastrointestinal syndromes were not specific enough for their routine use. Emergency department data enabled the detection of community outbreaks for flu