Search

BackgroundDelirium is an acute neurocognitive disorder that affects up to half of older hospitalized medical patients and can lead to dementia, longer hospital stays, increased health costs, and death. Although delirium can be prevented and treated, it is difficult to identify and predict. ObjectiveThis study aimed to improve machine learning models that retrospectively identify the presence of delirium during hospital stays (eg, to measure the effectiveness of delirium prevention interventions) by using the natural language processing (NLP) technique of sentiment analysis (in this case a feature that identifies sentiment toward, or away from, a delirium diagnosis). MethodsUsing data from the General Medicine Inpatient Initiative, a Canadian hospital data and analytics network, a detailed manual review of medical records was conducted from nearly 4000 admissions at 6 Toronto area hospitals. Furthermore, 25.74% (994/3862) of the eligible hospital admissions were labeled as having delirium. Using the data set collected from this study, we developed machine learning models with, and without, the benefit of NLP methods applied to diagnostic imaging reports, and we asked the question “can NLP improve machine learning identification of delirium?” ResultsAmong the eligible 3862 hospital admissions, 994 (25.74%) admissions were labeled as having delirium. Identification and calibration of the models were satisfactory. The accuracy and area under the receiver operating characteristic curve of the main model with NLP in the independent testing data set were 0.807 and 0.930, respectively. The accuracy and area under the receiver operating characteristic curve of the main model without NLP in the independent testing data set were 0.811 and 0.869, respectively. Model performance was also found to be stable over the 5-year period used in the experiment, with identification for a likely future holdout test set being no worse than identification for retrospective holdout test sets. ConclusionsOur machine learning model that included NLP (ie, sentiment analysis in medical image description text mining) produced valid identification of delirium with the sentiment analysis, providing significant additional benefit over the model without NLP.

The picture-word interference (PWI) paradigm has been used to investigate the time course of processes involved in word retrieval, but is challenging to implement online due to dependence on measurements of vocal reaction time. We performed a series of four experiments to examine picture-word interference and facilitation effects in a form of covert picture naming, with and without gamification. A target picture was accompanied by an audio word distractor that was either unrelated, phonologically-related, associatively-related, or categorically-related to the picture. Participants were instructed to judge whether the name of the target picture ended in the phoneme assigned to the block by pressing corresponding keys as quickly and accurately as possible. Experiments 1 and 2 successfully replicated categorical interference and phonological facilitation effects at different optimal stimulus-onset-asynchronies (SOAs) between words and pictures. Experiment 3 demonstrated that a key gamification feature (collecting coins) motivated faster speed at the expense of accuracy in the gamified vs. experimental format of the task. Experiment 4 adopted the optimal SOAs and verified that the gamification reveals expected interference and facilitation effects despite the speed-accuracy tradeoff. These studies confirmed that categorical interference occurs earlier than phonological facilitation, while both processes are independent from articulation and inherent to word retrieval itself. The covert PWI paradigm and its gamification have methodological value for neuroimaging studies in which articulatory artifacts obscure word retrieval processes, and may be developed into potential online word-finding assessments that can reveal word retrieval difficulties with greater sensitivity.

Clustering prior to regression analysis improves the accuracy of prediction in clinical decision making. However, most previously described methods focused on numerical data only. This paper investigated how well textual features can improve the accuracy of regression predictions. Preliminary diagnosis, diagnosis summary, and drug names used in prescriptions as provided in the MIMIC II dataset were used to derive textual features. We proposed the bag-of-entities indexing method, which relies on named entity recognition, a machine learning technique used for locating and identifying words into predefined classes. The proposed technique captured meaningful phrases from texts in health records and represented them in numerical vector format. Dimensionality of the data space was reduced using principal component analysis. The additional well-tuned textual features were then combined with existing numerical features in using cluster-boosted regression to predict patient mortality in ICU. The experimental results showed prediction improvement obtained from textual features over the use of numerical features only. We found that using the proposed indexing method outperformed traditional word-vector representation approaches (bag-of-words and bag-of-bigrams) as well as a state-of-the-art approach (Doc2vec) in terms of resulting accuracy in predicting death status. Moreover, instead of directly interpreting, the identifiable individual features were grouped into types and summarized. The summarized de-identified data of textual features handled by the proposed framework can support predictive classification while also reducing privacy concerns. Grouping of similar patients based on their electronic health records also benefits physicians through the improved differential diagnosis and effective treatment planning.

Introduction The global COVID-19 pandemic continues to have wide-ranging implications for health, including psychological well-being. A growing corpus of research reviews has emerged on the topic of psychological resilience in the context of the pandemic. However, this body of work has not been systematically reviewed for its quality, nor with respect to findings on the effectiveness of tools and strategies for psychological resilience. To this end, a meta-review protocol is proposed with the following objectives: (1) identify review work on the topic of psychological resilience during COVID-19; (2) assess the quality of this review work using A MeaSurement Tool to Assess systematic Reviews; (3) assess the risk of bias in this work; (4) generate a narrative summary of the key points, strengths and weaknesses; (5) identify the psychological resilience strategies that have been reviewed; (6) identify how these strategies have been evaluated for their effectiveness; (7) identify what outcomes were measured and (8) summarise the findings on strategies for psychological resilience so far, providing recommendations, if possible.Methods and analysis A systematic meta-review will be conducted in accordance with the Preferred Reporting Items for Systematic Reviews for Protocols and Joanna Briggs Institute umbrella review guidelines. Electronic searches of general databases, especially Web of Science, Scopus and PubMed, will be conducted. Only results from January 2020 onwards will be considered, coinciding with the COVID-19 pandemic. Only results in English will be included. Descriptive statistics, thematic analysis and narrative summaries describing the nature of the reviewed work and evaluation of psychological resilience strategies will be carried out.Ethics and dissemination Ethical approval is not needed for systematic review protocols. The results of the meta-review will be published in an international peer-reviewed journal. The raw and summarised data will be shared in the journal or other open venues.PROSPERO registration number CRD42021235288.

Abstract Background Systematic reviews are infrequently used by health care managers (HCMs) and policy-makers (PMs) in decision-making. HCMs and PMs co-developed and tested novel systematic review of effects formats to increase their use. Methods A three-phased approach was used to evaluate the determinants to uptake of systematic reviews of effects and the usability of an innovative and a traditional systematic review of effects format. In phase 1, survey and interviews were conducted with HCMs and PMs in four Canadian provinces to determine perceptions of a traditional systematic review format. In phase 2, systematic review format prototypes were created by HCMs and PMs via Conceptboard©. In phase 3, prototypes underwent usability testing by HCMs and PMs. Results Two hundred two participants (80 HCMs, 122 PMs) completed the phase 1 survey. Respondents reported that inadequate format (Mdn = 4; IQR = 4; range = 1–7) and content (Mdn = 4; IQR = 3; range = 1–7) influenced their use of systematic reviews. Most respondents (76%; n = 136/180) reported they would be more likely to use systematic reviews if the format was modified. Findings from 11 interviews (5 HCMs, 6 PMs) revealed that participants preferred systematic reviews of effects that were easy to access and read and provided more information on intervention effectiveness and less information on review methodology. The mean System Usability Scale (SUS) score was 55.7 (standard deviation [SD] 17.2) for the traditional format; a SUS score

BACKGROUND: Asthma is a common chronic condition. Work-related asthma (WRA) has a large socioeconomic impact and is increasing in prevalence but remains under-recognized. Although international guidelines recommend patient education, no widely available educational tool exists.

Introduction: Cognitive screening in settings such as emergency departments (ED) is frequently carried out using paper-and-pencil tests that require administration by trained staff. These assessments often compete with other clinical duties and thus may not be routinely administered in these busy settings. Literature has shown that the presence of cognitive impairments such as dementia and delirium are often missed in older ED patients. Failure to recognize delirium can have devastating consequences including increased mortality (Kakuma et al., 2003). Given the demands on emergency staff, an automated cognitive test to screen for delirium onset could be a valuable tool to support delirium prevention and management.In earlier research we examined the concurrent validity of a serious game, and carried out an initial assessment of its potential as a delirium screening tool (Tong et al., 2016). In this paper, we examine the test-retest reliability of the game, as it is an important criterion in a cognitive test for detecting risk of delirium onset.Objective: To demonstrate the test-retest reliability of the screening tool over time in a clinical sample of older emergency patients. A secondary objective is to assess whether there are practice effects that might make game performance unstable over repeated presentations.Materials and methods: Adults over the age of 70 were recruited from a hospital ED. Each patient played our serious game in an initial session soon after they arrived in the ED, and in follow up sessions conducted at eight-hour intervals (for each participant there were up to five follow up sessions, depending on how long the person stayed in the ED).Results: A total of 114 adults (61 females, 53 males) between the ages of 70 and 104 years (M=81 years, SD=7) participated in our study after screening out delirious patients. We observed a test-retest reliability of the serious game (as assessed by correlation r-values) between 0.5 and 0.8 across adjacent sessions.Conclusions: The game-based assessment for cognitive screening has relatively strong test-retest reliability and little evidence of practice effects among elderly emergency patients, and may be a useful supplement to existing cognitive assessment methods.

Human activity recognition is a key to a lot of applications such as healthcare and smart home. In this study, we provide a comprehensive survey on recent advances and challenges in human activity recognition (HAR) with deep learning. Although there are many surveys on HAR, they focused mainly on the taxonomy of HAR and reviewed the state-of-the-art HAR systems implemented with conventional machine learning methods. Recently, several works have also been done on reviewing studies that use deep models for HAR, whereas these works cover few deep models and their variants. There is still a need for a comprehensive and in-depth survey on HAR with recently developed deep learning methods. [ABSTRACT FROM AUTHOR]