The rapid digitalization of the healthcare domain in recent years highlighted the need for advanced predictive methods particularly based upon deep learning methods. Deep learning methods which are capable of dealing with time- series data have recently emerged in various fields such as natural language processing, machine translation, and the Intensive Care Unit (ICU). The recent applications of deep learning in ICU have increasingly received attention, and it has shown promising results for different clinical tasks; however, there is still a need for the benchmark models as far as a handful of public datasets are available in ICU. In this thesis, a novel benchmark model of four clinical tasks on a multi-center publicly available dataset is presented; we employed deep learning models to predict clinical studies. We believe this benchmark model can facilitate and accelerate the research in ICU by allowing other researchers to build on top of it. Moreover, we investigated the effectiveness of the proposed method to predict the risk of delirium in the varying observation and prediction windows, the variable ranking is provided to ease the implementation of a screening tool for helping caregivers at the bedside. Ultimately, an attention-based interpretable neural network is proposed to predict the outcome and rank the most influential variables in the model predictions’ outcome. Our experimental findings show the effectiveness of the proposed approaches in improving the application of deep learning models in daily ICU practice.

Objective: The diagnosis of delirium in intensive care unit patients is frequently missed. Key symptoms to identify delirium are motoric alterations, changes in activity level, and delirium-specific movements. This study aimed to explore features collected by a single wrist-worn accelerometer as indicators of delirium. Methods: The study included twenty-two patients in the intensive care unit. The data was collected with the GENEActiv accelerometer device and the activity level was calculated. Differences between the delirious and nondelirious patients were tested. Results: Differences in activity level and rest-activity patterns were noticed between the delirious and non-delirious patients. However, the differences were not found to be significant. Conclusion: Activity patterns revealed differences between delirious and non‐delirious patients. Further study is required to confirm the potential of actigraphy in the early detection of delirium in the intensive care unit. / Mål: Diagnosen delirium hos intensivvårdspatienter missas ofta. Nyckelsymptom för att identifiera delirium är motoriska förändringar, förändringar i aktivitetsnivå och deliriumspecifika rörelser. Denna studie syftade till att utforska funktioner som samlats in av en enskild handledsburen accelerometer som indikatorer på delirium. Metod: Studien omfattade tjugotvå patienter på intensivvårdsavdelningen. Data samlades in med GENEActiv accelerometerenheten och aktivitetsnivån beräknades. Skillnader mellan de delirious och icke-delirious patienterna testades. Resultat: Skillnader i aktivitetsnivå och viloaktivitetsmönster noterades mellan de deliriösa och icke-deliriösa patienterna. Skillnaderna visade sig dock inte vara signifikanta. Slutsats: Aktivitetsmönster avslöjade skillnader mellan deliriösa och ickedelirösa patienter. Ytterligare studier krävs för att bekräfta potentialen för aktigrafi vid tidig upptäckt av delirium på intensivvårdsavdelningen.

Introduction: Oral hygiene is a significant aspect of nursing care. Endocarditis, stroke, lung cancer, and hypertension have been associated with poor oral hygiene. Research exploring oral care practices for mechanically ventilated patients is well documented. In contrast, oral hygiene for the non-mechanically ventilated acute care population remains underestimated. The purpose of this study was to establish a baseline of the type, frequency, and consistency of oral hygiene being performed on non-mechanically ventilated ICU patients and explore how the oral care provided was documented. Methodology: A literature search was conducted and reported as a literature review. The databases CINAHL Plus with Full Text, MEDLINE, PsychINFO, Academic Search Premier, and Cochrane Database of Systematic Reviews were searched. Key terms used were "oral hygiene," "oral care," "oral intensity," "mouth rinse," "mouth care," chlorhexidine rinse and ICU, "intensive care unit," "critical care" and infection*, pneumonia*, NV, non-ventilat*, and nonventilat*. The articles' selection addressed type, frequency, consistency, and/or documentation of oral hygiene in ICU patients, particularly non-mechanically ventilated patients, if available. Inclusion criteria consisted of English language, and academic journal articles. No specified publication date was placed as a restriction. The results were limited to English language, academic journal articles, peer reviewed research articles, evidence-based articles or practices, and articles published within the last ten years (2006 to 2016). All articles on oral hygiene practices in the ICU or critical care population were included. Articles that did not relate to oral hygiene practices in acute care, ICU patients, or critically ill hospitalized patients were excluded. Articles focused solely on the mechanically ventilated or intubated population were also excluded. Results: The review yielded very few articles focusing solely on non-mechanically ventilated ICU patients. Nevertheless, resulting data showed four areas common to oral hygiene practices in non-mechanically ventilated patients in the ICU: type of documentation, type of products, frequency of care, and personnel providing care. Documentation was found to be lacking compared to personnel's self-reported frequency of oral care. Oral hygiene products were found to be consistent in non-mechanically ventilated patients, while there was no consistency of products used in the general acute care population. Oral hygiene was self-reported by staff members to have been performed an average of two to three times per day for non-mechanically ventilated patients. Oral hygiene self-reported frequency was found to be inconsistent among the general acute care population. Lastly, registered nurses (RNs) were the primary providers of oral hygiene to patients. Conclusions: Findings support the existing gap in the literature on oral hygiene practices in non-mechanically ventilated patients in the ICU. Despite evidence documenting the impact of oral hygiene on health, further research is guaranteed.

No description available.

No description available.

No description available.

By taking advantage of the increasing amount of available electronic health data, applications of machine learning in the intensive care unit have the potential to improve medical diagnostics and risk stratification. This thesis proposes an approach for early onset prediction of pneumothorax with such technique, using time series data extracted from a clinical database. The prevalence of pneumothorax among patients is identified through ICD-9 codes, and labels for the onset are obtained by relying on proxies closely related to the condition. Both simple algorithms and deep learning networks are used in a sliding window-based prediction framework, and the importance of each feature is measured with weighted Shapley values. The results proved the feasibility of this approach using Long Short-Term Memory models, although the number of false positives is noticeably high. Mechanical ventilation was the most contributing feature for the outcome. In future work, the focus should be on addressing the large class imbalance that prevails, along with considering more well-founded methods of target labeling.

In the intensive care unit, patients with crucial, life-threatening conditions are admitted and need constant monitoring. Here, the need for a quick and efficient decision support tool is the greatest. The use of machine learning has shown promising results in identifying patients at risk of different severe conditions in the intensive care unit and detection at an early stage is crucial in order to take preventive measures. This especially applies to conditions that can be hard to manage once developed, such as circulatory shock. In this master’s thesis, a machine learning modeling approach is suggested to detect and differentiate the onset of three types of circulatory shock – cardiogenic, hypovolemic and septic shock. Data was used from the open-source database MIMIC which represents thousands of patients from intensive care. The data was preprocessed and labels for the three shock types were created using ICD-9 codes combined with a proxy that is closely related to the condition – vasopressor. Different machine learning algorithms were then used for a static onset prediction as a base. The best performing models were also trained for a dynamic onset prediction in order to make predictions up to four hours ahead of onset. All models were evaluated using different evaluation metrics and at last, an interpretation method was used to enable a simpler interpretation of the results. The final results show that it is possible to detect and distinguish between the three types of shock, up to four hours ahead of onset. For future developments, further development and validation using more data should be the main focus before testing it in a clinical setting.

Background: NICU experiences pose significant challenges for parents, especially immigrants, necessitating comprehensive support. South Asian immigrants, comprising 25% of Canada's visible minority population, face unique adversities related to gender roles, hindering access to essential health services and integration. Coupled with unfavourable social determinants of health (SDoH), these challenges worsen issues like inadequate prenatal care, education, and nutrition, predictors of adverse maternal and neonatal health outcomes. Existing studies lack insight into the specific experiences of South Asian immigrant mothers in the NICU. This study investigates these experiences. Methods: Using a qualitative descriptive approach, we recruited four participants for semi-structured interviews, supplemented by a demographic questionnaire and participant observation. Qualitative content analysis was employed for data analysis. Findings: Four key themes were identified from the interviews: (1) Seeking Understanding, (2) Cultural Influence on NICU Experience, (3) Motherhood Journey, and (4) Circle of Care. Implications: This study fills a gap in NICU research for South Asian immigrant women, providing a foundation for future nursing research and practice. It underscores the importance of communication and preparation for discharge delays to ease parental concerns. Additionally, it emphasizes culturally sensitive care practices and encourages further exploration of cultural influences on hospital experiences. Insights from this study can benefit other ethno-racial immigrant groups. / Thesis / Master of Science in Nursing (MSN) / Existing research offers insights into the general challenges and distress often associated with mothers' experiences in the NICU. However, there is little evidence to understand the specific experiences of South Asian immigrant mothers within this context. The objective of this thesis is to describe and understand the experiences in the NICU reported by this population. Employing a qualitative description methodology, this study engaged four eligible participants. Data collection entailed semi-structured interviews alongside a demographic questionnaire. Employing qualitative content analysis, four overarching themes were identified: (1) Seeking to Understand, (2) The Impact of South Asian Culture on the NICU Experience, (3) Becoming a Mother One Step at a Time, and (4) Circle of Care.

京都大学 / 新制・課程博士 / 博士(医学) / 甲第25157号 / 医博第5043号 / 新制||医||1070(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 石見 拓, 教授 西浦 博, 教授 江木 盛時 / 学位規則第4条第1項該当 / Doctor of Agricultural Science / Kyoto University / DFAM