Imbalanced prediction of emergency department admission using natural language processing and deep neural network.

The emergency department (ED) plays a very significant role in the hospital. Owing to the rising number of ED visits, medical service points, and ED market, overcrowding of EDs has become serious worldwide. Overcrowding has long been recognized as a vital issue that increases the risk to patients and negative emotions of medical personnel and impacts hospital cost management. For the past years, many researchers have been applying artificial intelligence to reduce crowding situations in the ED. Nevertheless, the datasets in ED hospital admission are naturally inherent with the high-class imbalance in the real world. Previous studies have not considered the imbalance of the datasets, particularly addressing the imbalance. This study purposes to develop a natural language processing model of a deep neural network with an attention mechanism to solve the imbalanced problem in ED admission. The proposed framework is used for predicting hospital admission so that the hospitals can arrange beds early and solve the problem of congestion in the ED. Furthermore, the study compares a variety of methods and obtains the best composition that has the best performance for forecasting hospitalization in ED. The study used the data from a specific hospital in Taiwan as an empirical study. The experimental result demonstrates that almost all imbalanced methods can improve the model's performance. In addition, the natural language processing model of Bi-directional Long Short-Term Memory with attention mechanism has the best results in all-natural language processing methods.

Utilizing online stochastic optimization on scheduling of intensity-modulate radiotherapy therapy (IMRT).

According to Ministry of Health and Welfare of Taiwan, cancer has been one of the major causes of death in Taiwan since 1982. The Intensive-Modulated Radiation Therapy (IMRT) is one of the most important radiotherapies of cancers, especially for Nasopharyngeal cancers, Digestive system cancers and Cervical cancers. For patients, if they can receive the treatment at the earliest possibility while diagnosed with cancers, their survival rate increases. However, the discussion of effective patient scheduling models of IMRT to reduce patients' waiting time is still limited in literature. This study proposed a mathematical model to improve the efficiency of patient scheduling. The research was composed of two stages. In the first stage, the online stochastic algorithm was proposed to improve the performance of present scheduling system. In the second stage the impact of future treatment to reduce patients' waiting time was considered. A genetic algorithm (GA) was then proposed to solve the online stochastic scheduling problem. This research collected data from a practical medical institute and the proposed model was validated with real data. It contributes to both theory and practice by proposing a practical model to assist the medical institute in implementing patient scheduling in a more efficient manner.

Stochastic resource allocation in emergency departments with a multi-objective simulation optimization algorithm.

The number of emergency cases or emergency room visits rapidly increases annually, thus leading to an imbalance in supply and demand and to the long-term overcrowding of hospital emergency departments (EDs). However, current solutions to increase medical resources and improve the handling of patient needs are either impractical or infeasible in the Taiwanese environment. Therefore, EDs must optimize resource allocation given limited medical resources to minimize the average length of stay of patients and medical resource waste costs. This study constructs a multi-objective mathematical model for medical resource allocation in EDs in accordance with emergency flow or procedure. The proposed mathematical model is complex and difficult to solve because its performance value is stochastic; furthermore, the model considers both objectives simultaneously. Thus, this study develops a multi-objective simulation optimization algorithm by integrating a non-dominated sorting genetic algorithm II (NSGA II) with multi-objective computing budget allocation (MOCBA) to address the challenges of multi-objective medical resource allocation. NSGA II is used to investigate plausible solutions for medical resource allocation, and MOCBA identifies effective sets of feasible Pareto (non-dominated) medical resource allocation solutions in addition to effectively allocating simulation or computation budgets. The discrete event simulation model of ED flow is inspired by a Taiwan hospital case and is constructed to estimate the expected performance values of each medical allocation solution as obtained through NSGA II. Finally, computational experiments are performed to verify the effectiveness and performance of the integrated NSGA II and MOCBA method, as well as to derive non-dominated medical resource allocation solutions from the algorithms.

Intelligent Healthcare System Using Patients Confidential Data Communication in Electrocardiogram Signals.

With the advent of the aging era, healthcare and elderly care have become the focus of medical care, especially the care of the elderly with dementia. Patients' confidential data hiding is a useful technology for healthcare and patient information privacy. In this study, we implement an intelligent healthcare system using the multiple-coefficient quantization technology in transform domain to hide patients' confidential data into electrocardiogram (ECG) signals obtained by ECG sensor module. In embedding patients' confidential data, we first consider a non-linear model for optimizing the quality of the embedded ECG signals. Next, we apply simulated annealing (SA) to solve the non-linear model so as to have good signal-to-noise ratio (SNR), root mean square error (RMSE), and relative RMSE (rRMSE). Accordingly, the distortion of the PQRST complexes and the ECG amplitude is very small so that the embedded confidential data can satisfy the requirements of physiological diagnostics. In end devices, one can receive the ECG signals with the embedded confidential data and without the original ECG signals. Experimental results confirm the effectiveness of our method, which remains high quality for each ECG signal with the embedded confidential data no matter how the quantization size Q is increased.

Effects of cow's and goat's milk as fermentation media on the microbial ecology of sugary kefir grains.

In the present study, we have investigated the importance of fermentation media on grain formation and the microbial characteristics of sugary kefir. The sugary kefir grains were fermented in brown sugar, cow's milk or goat's milk. Using culture-dependent and culture-independent methods, we identified the microorganisms present in both the grains and filtrate and then evaluated their distribution. The structure of the grains was also observed by scanning electronic microscopy (SEM). The identification results indicated that there were remarkable changes in microbial ecological profiles of the sugary kefir grains and their filtrates when brown sugar and milk were compared as fermentation media. Three lactic acid bacteria (LAB) species (Leuconostoc mesenteroides, Lactobacillus mali and Lactobacillus hordei) were found in the grains fermented using brown sugar. However, four species, named Leu. mesenteroides, Lactococcus lactis, Bifidobacterium psychraerophilum and Enterococcus faecalis, were identified in the grains fermented using either cow's or goat's milk. The size and structure of the kefir grains were also significantly influenced by the culture medium. We hypothesize that the grains originally may contain many different microorganisms and the identified changes are an adaption to each specific medium during grain formation and growth. The distribution of strains thus may vary depending on the carbon and energy sources available for grain fermentation and these microbial changes will further affect the granulation and growth of the grains. This study is important to our understanding of the mechanism of kefir grain formation and growth because it explores the relationship between fermentation media and kefir microorganisms.