BMC Medical Informatics and Decision Making

Internet social networking tools and the emerging web 2.0 technologies are providing a new way for web users and health workers in information sharing and knowledge dissemination. Based on the characters of immediate, two-way and large scale of impact, the internet social networking tools have been utilized as a solution in emergency response during disasters. This paper highlights the use of internet social networking in disaster emergency response and public health management of disasters by focusing on a case study of the typhoon Morakot disaster in Taiwan. In the case of typhoon disaster in Taiwan, internet social networking and mobile technology were found to be helpful for community residents, professional emergency rescuers, and government agencies in gathering and disseminating real-time information, regarding volunteer recruitment and relief supplies allocation. We noted that if internet tools are to be integrated in the development of emergency response system, the accessibility, accuracy, validity, feasibility, privacy and the scalability of itself should be carefully considered especially in the effort of applying it in resource poor settings. This paper seeks to promote an internet-based emergency response system by integrating internet social networking and information communication technology into central government disaster management system. Web-based networking provides two-way communication which establishes a reliable and accessible tunnel for proximal and distal users in disaster preparedness and management.

There has been indisputable growth in adoption of electronic medical record (EMR) systems in the recent years. However, physicians’ progress in using these systems has stagnated when measured with maturity scales. While this so-called ceiling effect has been observed and its consequences described in previous studies, there is a paucity of research on the elements that could explain such an outcome. We first suggest that in the context of EMR systems we are in presence of a “tiered ceiling effect” and then we show why such phenomenon occurs. We conducted in-depth case studies in three primary care medical practices in Canada where physicians had been using EMR systems for 3 years or more. A total of 37 semi-structured interviews were conducted with key informants: family physicians (about half of the interviews), nurses, secretaries, and administrative managers. Additional information was obtained through notes taken during observations of users interacting with their EMR systems and consultation of relevant documents at each site. We used abductive reasoning to infer explanations of the observed phenomenon by going back and forth between the case data and conceptual insights. Our analysis shows that a ceiling effect has taken place in the three clinics. We identified a set of conditions preventing the users from overcoming the ceiling. In adopting an EMR system, all three clinics essentially sought improved operational efficiency. This had an influence on the criteria used to assess the systems available on the market and eventually led to the adoption of a system that met the specified criteria without being optimal. Later, training sessions focussed on basic functionalities that minimally disturbed physicians’ habits while helping their medical practices become more efficient. Satisfied with the outcome of their system use, physicians were likely to ignore more advanced EMR system functionalities. This was because their knowledge about EMR systems came almost exclusively from a single source of information: their EMR system vendors. This knowledge took the form of interpretations of what the innovation was (know-what), with little consideration of the rationales for innovation adoption (know-why) or hands-on strategies for adopting, implementing and assimilating the innovation in the organization (know-how). This paper provides a holistic view of the technological innovation process in primary care and contends that limited learning, satisficing behaviours and organizational inertia are important factors leading to the ceiling effect frequently experienced in the EMR system assimilation phase.

Institutions, government departments, and healthcare professionals engage in social media because it facilitates reaching a large number of people simultaneously. YouTube provides a platform whereby anyone can upload videos and gain feedback on their content from other users. Many YouTube videos are related to health and science, and many people search YouTube for health-related information. YouTube has been acknowledged as a key public information source in recent crises caused by Zika, H1N1, swine flu, and most recently, COVID-19. YouTube videos were collected from the YouTube Application Programming Interface (API) using the search terms COVID-19, coronavirus, COVID19, and corona. The search was conducted on April 4 and 5, 2020. The initial investigation found a total of 1084 videos. The second step involved identifying and verifying the videos for their relationship to COVID-19 information and excluding videos that did not relate to COVID-19 or were in a language other than English and Hindi. An analysis of YouTube videos covering COVID-19, uploaded in early 2020, in English and Hindi. The sample comprised 349 videos (n = 334 English). Videos were characterized by contributor, duration, content, and reception (views/likes/dislikes/comments). The majority contained general information, with only 4.01% focusing on symptoms and 11.17% on treatment and outcomes. Further, the majority (n = 229) were short videos of under 10 min duration. Videos provided by government and health care professionals comprised 6.87% and 5.74% % of the sample, respectively. News channels uploaded 71.63% of videos. YouTube may provide a significant resource for disseminating of information on public health issues like outbreaks of viral infections and should be utilized by healthcare agencies for this purpose. However, there is currently no way to determine whether a video has been produced or verified by authorized healthcare professionals. This limitation needs to be addressed so that the vital distribution services offered by platforms like YouTube can be fully utilized for increasing public understanding of healthcare science, particularly during a crisis such as a pandemic.

Machine learning has been used extensively in clinical text classification tasks. Deep learning approaches using word embeddings have been recently gaining momentum in biomedical applications. In an effort to automate the identification of altered mental status (AMS) in emergency department provider notes for the purpose of decision support, we compare the performance of classic bag-of-words-based machine learning classifiers and novel deep learning approaches. We used a case-control study design to extract an adequate number of clinical notes with AMS and non-AMS based on ICD codes. The notes were parsed to extract the history of present illness, which was used as the clinical text for the classifiers. The notes were manually labeled by clinicians. As a baseline for comparison, we tested several traditional bag-of-words based classifiers. We then tested several deep learning models using a convolutional neural network architecture with three different types of word embeddings, a pre-trained word2vec model and two models without pre-training but with different word embedding dimensions. We evaluated the models on 1130 labeled notes from the emergency department. The deep learning models had the best overall performance with an area under the ROC curve of 98.5% and an accuracy of 94.5%. Pre-training word embeddings on the unlabeled corpus reduced training iterations and had performance that was statistically no different than the other deep learning models. This supervised deep learning approach performs exceedingly well for the detection of AMS symptoms in clinical text in our environment. Further work is needed for the generalizability of these findings, including evaluation of these models in other types of clinical notes and other environments. The results seem promising for the ultimate use of these types of classifiers in combination with other information derived from the electronic health records as input for clinical decision support.

Videoconferencing has been proposed as a way of improving access to healthcare for older adults in care homes. Despite this, effective uptake of videoconferencing remains varied. This study evaluates a videoconferencing service for care home staff seeking support from healthcare professionals for the care of residents. The aim was to explore factors affecting the uptake and sustainability of videoconferencing in care homes, to establish what works for whom, in which circumstances and respects. The findings informed recommendations for commissioners and strategic managers on how best to implement videoconferencing for remote healthcare provision in care homes for older adults. Realist evaluation was used to develop, refine and test theories around the uptake and maintenance of videoconferencing in three care homes across Yorkshire and the Humber, England. The care homes were selected using maximum variation sampling regarding the extent to which they used videoconferencing. A developmental inquiry framework and realist interviews were used to identify Context, Mechanism and Outcome Configurations (CMOCs) regarding uptake and sustainability of the service. Participants included care home residents (aged > 65) and staff, relatives and strategic managers of care home chains. The interviews were an iterative process conducted alongside data analysis. Transcripts of audio recordings were entered into NVIVO 12, initially coded into themes, then hypotheses developed, refined and tested. Outcomes were generated in relation to two main contextual factors, these were: (1) communication culture in the home and (2) the prior knowledge and experience that staff have of videoconferencing. The key facilitators identified were aspects of leadership, social links within the home and psychological safety which promoted shared learning and confidence in using the technology. Videoconferencing is a valuable tool, but successful implementation and sustainability are dependent on care home culture and staff training to promote confidence through positive and supported experiences.

Predicting progression from Mild Cognitive Impairment (MCI) to Alzheimer’s Disease (AD) is an utmost open issue in AD-related research. Neuropsychological assessment has proven to be useful in identifying MCI patients who are likely to convert to dementia. However, the large battery of neuropsychological tests (NPTs) performed in clinical practice and the limited number of training examples are challenge to machine learning when learning prognostic models. In this context, it is paramount to pursue approaches that effectively seek for reduced sets of relevant features. Subsets of NPTs from which prognostic models can be learnt should not only be good predictors, but also stable, promoting generalizable and explainable models. We propose a feature selection (FS) ensemble combining stability and predictability to choose the most relevant NPTs for prognostic prediction in AD. First, we combine the outcome of multiple (filter and embedded) FS methods. Then, we use a wrapper-based approach optimizing both stability and predictability to compute the number of selected features. We use two large prospective studies (ADNI and the Portuguese Cognitive Complaints Cohort, CCC) to evaluate the approach and assess the predictive value of a large number of NPTs. The best subsets of features include approximately 30 and 20 (from the original 79 and 40) features, for ADNI and CCC data, respectively, yielding stability above 0.89 and 0.95, and AUC above 0.87 and 0.82. Most NPTs learnt using the proposed feature selection ensemble have been identified in the literature as strong predictors of conversion from MCI to AD. The FS ensemble approach was able to 1) identify subsets of stable and relevant predictors from a consensus of multiple FS methods using baseline NPTs and 2) learn reliable prognostic models of conversion from MCI to AD using these subsets of features. The machine learning models learnt from these features outperformed the models trained without FS and achieved competitive results when compared to commonly used FS algorithms. Furthermore, the selected features are derived from a consensus of methods thus being more robust, while releasing users from choosing the most appropriate FS method to be used in their classification task.