Search

Abstract Background Hyperglycemic crises are associated with high morbidity and mortality. Previous studies have proposed methods to predict adverse outcomes of patients in hyperglycemic crises; however, artificial intelligence (AI) has never been used to predict adverse outcomes. We implemented an AI model integrated with the hospital information system (HIS) to clarify whether AI could predict adverse outcomes. Methods We included 2,666 patients with hyperglycemic crises from emergency departments (ED) between 2009 and 2018. The patients were randomized into a 70%/30% split for AI model training and testing. Twenty-two feature variables from the electronic medical records were collected. The performance of the multilayer perceptron (MLP), logistic regression, random forest, Light Gradient Boosting Machine (LightGBM), support vector machine (SVM), and K-nearest neighbor (KNN) algorithms was compared. We selected the best algorithm to construct an AI model to predict sepsis or septic shock, intensive care unit (ICU) admission, and all-cause mortality within 1 month. The outcomes between the non-AI and AI groups were compared after implementing the HIS and predicting the hyperglycemic crisis death (PHD) score. Results The MLP had the best performance in predicting the three adverse outcomes, compared with the random forest, logistic regression, SVM, KNN, and LightGBM models. The areas under the curves (AUCs) using the MLP model were 0.852 for sepsis or septic shock, 0.743 for ICU admission, and 0.796 for all-cause mortality. Furthermore, we integrated the AI predictive model with the HIS to assist decision making in real time. No significant differences in ICU admission or all-cause mortality were detected between the non-AI and AI groups. The AI model performed better than the PHD score for predicting all-cause mortality (AUC 0.796 vs. 0.693). Conclusions A real-time AI predictive model is a promising method for predicting adverse outcomes in ED patients with hyperglycemic crises. Further studies recruiting more patients are warranted.

Hypoglycemia is a common metabolic disorder that occurs in the neonatal period. Early identification of neonates at risk of developing hypoglycemia can optimize therapeutic strategies in neonatal care. This study aims to develop a machine learning model and implement a predictive application to assist clinicians in accurately predicting the risk of neonatal hypoglycemia within four hours after birth. Our retrospective study analyzed data from neonates born ≥35 weeks gestational age and admitted to the well-baby nursery between 1 January 2011 and 31 August 2021. We collected electronic medical records of 2687 neonates from a tertiary medical center in Southern Taiwan. Using 12 clinically relevant features, we evaluated nine machine learning approaches to build the predictive models. We selected the models with the highest area under the receiver operating characteristic curve (AUC) for integration into our hospital information system (HIS). The top three AUC values for the early neonatal hypoglycemia prediction models were 0.739 for Stacking, 0.732 for Random Forest and 0.732 for Voting. Random Forest is considered the best model because it has a relatively high AUC and shows no significant overfitting (accuracy of 0.658, sensitivity of 0.682, specificity of 0.649, F1 score of 0.517 and precision of 0.417). The best model was incorporated in the web-based application integrated into the hospital information system. Shapley Additive Explanation (SHAP) values indicated mode of delivery, gestational age, multiparity, respiratory distress, and birth weight < 2500 gm as the top five predictors of neonatal hypoglycemia. The implementation of our machine learning model provides an effective tool that assists clinicians in accurately identifying at-risk neonates for early neonatal hypoglycemia, thereby allowing timely interventions and treatments.

Evidence-based practice (EBP) is an essential component of healthcare practice that ensures the delivery of high-quality care by integrating the best available evidence. This study aimed to explore factors influencing EBP among nursing professionals in Taiwan. A cross-sectional survey study was conducted with 752 registered nurses and nurse practitioners recruited from a regional teaching hospital in southern Taiwan. EBP competency was evaluated using the Taipei Evidence-Based Practice Questionnaire (TEBPQ). The results showed that participation in evidence-based courses or training within the past year had the strongest association with EBP competencies (Std. B = 0.157, p < 0.001). Holding a graduate degree (Std. B = 0.151, p < 0.001), working in gynecology or pediatrics (Std. B = 0.126, p < 0.001), searching the literature in electronic databases (Std. B = 0.072, p = 0.039), and able to read academic articles in English (Std. B = 0.088, p = 0.005) were significantly associated with higher TEBPQ scores. Younger age (Std. B = −0.105, p = 0.005) and male gender (Std. B = 0.089, p = 0.010) were also identified as factors contributing to higher EBP competencies. The study highlights the importance of ongoing professional development, including EBP training and language proficiency, in enhancing EBP competencies among nursing professionals in Taiwan.

BackgroundNegation and speculation unrelated to abnormal findings can lead to false-positive alarms for automatic radiology report highlighting or flagging by laboratory information systems. ObjectiveThis internal validation study evaluated the performance of natural language processing methods (NegEx, NegBio, NegBERT, and transformers). MethodsWe annotated all negative and speculative statements unrelated to abnormal findings in reports. In experiment 1, we fine-tuned several transformer models (ALBERT [A Lite Bidirectional Encoder Representations from Transformers], BERT [Bidirectional Encoder Representations from Transformers], DeBERTa [Decoding-Enhanced BERT With Disentangled Attention], DistilBERT [Distilled version of BERT], ELECTRA [Efficiently Learning an Encoder That Classifies Token Replacements Accurately], ERNIE [Enhanced Representation through Knowledge Integration], RoBERTa [Robustly Optimized BERT Pretraining Approach], SpanBERT, and XLNet) and compared their performance using precision, recall, accuracy, and F1-scores. In experiment 2, we compared the best model from experiment 1 with 3 established negation and speculation-detection algorithms (NegEx, NegBio, and NegBERT). ResultsOur study collected 6000 radiology reports from 3 branches of the Chi Mei Hospital, covering multiple imaging modalities and body parts. A total of 15.01% (105,755/704,512) of words and 39.45% (4529/11,480) of important diagnostic keywords occurred in negative or speculative statements unrelated to abnormal findings. In experiment 1, all models achieved an accuracy of >0.98 and F1-score of >0.90 on the test data set. ALBERT exhibited the best performance (accuracy=0.991; F1-score=0.958). In experiment 2, ALBERT outperformed the optimized NegEx, NegBio, and NegBERT methods in terms of overall performance (accuracy=0.996; F1-score=0.991), in the prediction of whether diagnostic keywords occur in speculative statements unrelated to abnormal findings, and in the improvement of the performance of keyword extraction (accuracy=0.996; F1-score=0.997). ConclusionsThe ALBERT deep learning method showed the best performance. Our results represent a significant advancement in the clinical applications of computer-aided notification systems.

BackgroundFor the intensivists, accurate assessment of the ideal timing for successful weaning from the mechanical ventilation (MV) in the intensive care unit (ICU) is very challenging.PurposeUsing artificial intelligence (AI) approach to build two-stage predictive models, namely, the try-weaning stage and weaning MV stage to determine the optimal timing of weaning from MV for ICU intubated patients, and implement into practice for assisting clinical decision making.MethodsAI and machine learning (ML) technologies were used to establish the predictive models in the stages. Each stage comprised 11 prediction time points with 11 prediction models. Twenty-five features were used for the first-stage models while 20 features were used for the second-stage models. The optimal models for each time point were selected for further practical implementation in a digital dashboard style. Seven machine learning algorithms including Logistic Regression (LR), Random Forest (RF), Support Vector Machines (SVM), K Nearest Neighbor (KNN), lightGBM, XGBoost, and Multilayer Perception (MLP) were used. The electronic medical records of the intubated ICU patients of Chi Mei Medical Center (CMMC) from 2016 to 2019 were included for modeling. Models with the highest area under the receiver operating characteristic curve (AUC) were regarded as optimal models and used to develop the prediction system accordingly.ResultsA total of 5,873 cases were included in machine learning modeling for Stage 1 with the AUCs of optimal models ranging from 0.843 to 0.953. Further, 4,172 cases were included for Stage 2 with the AUCs of optimal models ranging from 0.889 to 0.944. A prediction system (dashboard) with the optimal models of the two stages was developed and deployed in the ICU setting. Respiratory care members expressed high recognition of the AI dashboard assisting ventilator weaning decisions. Also, the impact analysis of with- and without-AI assistance revealed that our AI models could shorten the patients’ intubation time by 21 hours, besides gaining the benefit of substantial consistency between these two decision-making strategies.ConclusionWe noticed that the two-stage AI prediction models could effectively and precisely predict the optimal timing to wean intubated patients in the ICU from ventilator use. This could reduce patient discomfort, improve medical quality, and lower medical costs. This AI-assisted prediction system is beneficial for clinicians to cope with a high demand for ventilators during the COVID-19 pandemic.

Abstract Background Predicting outcomes in older patients with influenza in the emergency department (ED) by machine learning (ML) has never been implemented. Therefore, we conducted this study to clarify the clinical utility of implementing ML. Methods We recruited 5508 older ED patients (≥65 years old) in three hospitals between 2009 and 2018. Patients were randomized into a 70%/30% split for model training and testing. Using 10 clinical variables from their electronic health records, a prediction model using the synthetic minority oversampling technique preprocessing algorithm was constructed to predict five outcomes. Results The best areas under the curves of predicting outcomes were: random forest model for hospitalization (0.840), pneumonia (0.765), and sepsis or septic shock (0.857), XGBoost for intensive care unit admission (0.902), and logistic regression for in-hospital mortality (0.889) in the testing data. The predictive model was further applied in the hospital information system to assist physicians’ decisions in real time. Conclusions ML is a promising way to assist physicians in predicting outcomes in older ED patients with influenza in real time. Evaluations of the effectiveness and impact are needed in the future.

Background: Tuberculosis (TB) is one of the leading causes of death worldwide and a major cause of ill health. Without treatment, the mortality rate of TB is approximately 50%; with treatment, most patients with TB can be cured. However, anti-TB drug treatments may result in many adverse effects. Therefore, it is important to detect and predict these adverse effects early. Our study aimed to build models using an artificial intelligence/machine learning approach to predict acute hepatitis, acute respiratory failure, and mortality after TB treatment. Materials and Methods: Adult patients (age ≥ 20 years) who had a TB diagnosis and received treatment from January 2004 to December 2021 were enrolled in the present study. Thirty-six feature variables were used to develop the predictive models with AI. The data were randomly stratified into a training dataset for model building (70%) and a testing dataset for model validation (30%). These algorithms included XGBoost, random forest, MLP, light GBM, logistic regression, and SVM. Results: A total of 2248 TB patients in Chi Mei Medical Center were included in the study; 71.7% were males, and the other 28.3% were females. The mean age was 67.7 ± 16.4 years. The results showed that our models using the six AI algorithms all had a high area under the receiver operating characteristic curve (AUC) in predicting acute hepatitis, respiratory failure, and mortality, and the AUCs ranged from 0.920 to 0.766, 0.884 to 0.797, and 0.834 to 0.737, respectively. Conclusions: Our AI models were good predictors and can provide clinicians with a valuable tool to detect the adverse prognosis in TB patients early.

Objective: Environmental factors, eating habits, and different ages might affect the profiles of allergy sensitization. The purpose of this study was to survey the different profiles of allergen sensitization in different ages in eastern Taiwan. Materials and Methods: We analyzed the allergic patients who had allergen sensitization examinations by the Phadiatop (atopy screen; IBT Laboratories, Lenexa, KS, USA) and the Pharmacia CAP System method at Haulien Tzu Chi Hospital from January 2010 to December 2015. Results were compared in different ages. Results: A total of 15,455 patients were analyzed. The food and aeroallergen screen sensitization rate of children was significantly higher than that of adults (44.0% vs. 9.9% and 61.9% vs. 52.2% P < 0.05). Children had statistically significantly higher cow milk allergen-specific sensitization rate than that of adults (32.9% vs. 5.8% P < 0.05). The higher sensitization of shrimp occurred in adults than children. (33.6% vs. 24.8% P < 0.05). Conclusion: Our data demonstrate that children have higher cow milk allergy sensitization and adults have higher sensitization of shrimp.

Abstract Background A big-data-driven and artificial intelligence (AI) with machine learning (ML) approach has never been integrated with the hospital information system (HIS) for predicting major adverse cardiac events (MACE) in patients with chest pain in the emergency department (ED). Therefore, we conducted the present study to clarify it. Methods In total, 85,254 ED patients with chest pain in three hospitals between 2009 and 2018 were identified. We randomized the patients into a 70%/30% split for ML model training and testing. We used 14 clinical variables from their electronic health records to construct a random forest model with the synthetic minority oversampling technique preprocessing algorithm to predict acute myocardial infarction (AMI)

Growing evidence suggests that the gut microbiota and their metabolites are associated with bone homeostasis and fragility. However, this association is limited to microbial taxonomic differences. This study aimed to explore whether gut bacterial community associations, composition, and functions are associated with osteopenia and osteoporosis. We compared the gut bacterial community composition and interactions of healthy postmenopausal women with normal bone density (n = 8) with those of postmenopausal women with osteopenia (n = 18) and osteoporosis (n = 21) through 16S rRNA sequencing coupled with network biology and statistical analyses. The results of this study showed reduced alpha diversity in patients with osteoporosis, followed by that in patients with osteopenia, then in healthy controls. Taxonomic analysis revealed that significantly enriched bacterial genera with higher abundance was observed in patients with osteoporosis and osteopenia than in healthy subjects. Additionally, a co-occurrence network revealed that, compared to healthy controls, bacterial interactions were higher in patients with osteoporosis, followed by those with osteopenia. Further, NetShift analysis showed that a higher number of bacteria drove changes in the microbial community structure of patients with osteoporosis than osteopenia. Correlation analysis revealed that most of these driver bacteria had a significant positive relationship with several significant metabolic pathways. Further, ordination analysis revealed that height and T-score were the primary variables influencing the gut microbial community structure. Taken together, this study evaluated that microbial community interaction is more important than the taxonomic differences in knowing the critical role of gut microbiota in postmenopausal women associated with osteopenia and osteoporosis. Additionally, the significantly enriched bacteria and functional pathways might be potential biomarkers for the prognosis and treatment of postmenopausal women with osteopenia and osteoporosis.

Ascophyllum nodosum and Fucus vesiculosus both contain unique polyphenols called phlorotannins. Phlorotannins reportedly possess various pharmacological activities. A previous study reported that the activity of phlorotannin is strongly correlated with the normalization of metabolic function, and phlorotannins are extremely promising nutrients for use in the treatment of metabolic syndrome. To date, no study has explored the antihyperlipidemic effects of phlorotannins from A. nodosum and F. vesiculosus in animal models. Therefore, in the present study, we investigated the effects of phlorotannins using a rat model of high-energy diet (HED)-induced hyperlipidemia. The results showed that the rats that were fed an HED and treated with phlorotannin-rich extract from A. nodosum and F. vesiculosus had significantly lower serum fasting blood sugar (FBS), aspartate aminotransferase (AST), alanine aminotransferase (ALT), total cholesterol (TC), triacylglyceride (TG) and free fatty acids (FFAs) levels and hepatic TG level and had higher serum insulin, high-density lipoprotein cholesterol (HDL-C) levels and lipase activity in their fat tissues than in the case with the rats that were fed the HED alone. A histopathological analysis revealed that phlorotannin-rich extract could significantly reduce the size of adipocytes around the epididymis. In addition, the rats treated with phlorotannin-rich extract had significantly lowered interleukin 6 (IL-6) and tumor necrosis factor alpha (TNF-α) levels and increased superoxide dismutase (SOD) and glutathione peroxidase (GPX) activities than did those in the HED group. These results suggested that the phlorotannin-rich extract stimulated lipid metabolism and may have promoted lipase activity in rats with HED-induced hyperlipidemia. Our results indicated that A. nodosum and F. vesiculosus, marine algae typically used as health foods, have strong antihyperlipidemic effects and may, therefore, be useful for preventing atherosclerosis. These algae may be incorporated into antihyperlipidemia pharmaceuticals and functional foods.

The emergency department (ED) is at the forefront of medical care, and the medical team needs to make outright judgments and treatment decisions under time constraints. Thus, knowing how to make personalized and precise predictions is a very challenging task. With the advancement of artificial intelligence (AI) technology, Chi Mei Medical Center (CMMC) adopted AI, the Internet of Things (IoT), and interaction technologies to establish diverse prognosis prediction models for eight diseases based on the ED electronic medical records of three branch hospitals. CMMC integrated these predictive models to form a digital AI dashboard, showing the risk status of all ED patients diagnosed with any of these eight diseases. This study first explored the methodology of CMMC’s AI development and proposed a four-tier AI dashboard architecture for ED implementation. The AI dashboard’s ease of use, usefulness, and acceptance was also strongly affirmed by the ED medical staff. The ED AI dashboard is an effective tool in the implementation of real-time risk monitoring of patients in the ED and could improve the quality of care as a part of best practice. Based on the results of this study, it is suggested that healthcare institutions thoughtfully consider tailoring their ED dashboard designs to adapt to their unique workflows and environments.

Traumatic brain injury (TBI) remains a critical public health challenge. Although studies have found several prognostic factors for TBI, a useful early predictive tool for mortality has yet to be developed in the triage of the emergency room. This study aimed to use machine learning algorithms of artificial intelligence (AI) to develop predictive models for TBI patients in the emergency room triage. We retrospectively enrolled 18,249 adult TBI patients in the electronic medical records of three hospitals of Chi Mei Medical Group from January 2010 to December 2019, and undertook the 12 potentially predictive feature variables for predicting mortality during hospitalization. Six machine learning algorithms including logistical regression (LR) random forest (RF), support vector machines (SVM), LightGBM, XGBoost, and multilayer perceptron (MLP) were used to build the predictive model. The results showed that all six predictive models had high AUC from 0.851 to 0.925. Among these models, the LR-based model was the best model for mortality risk prediction with the highest AUC of 0.925; thus, we integrated the best model into the existed hospital information system for assisting clinical decision-making. These results revealed that the LR-based model was the best model to predict the mortality risk in patients with TBI in the emergency room. Since the developed prediction system can easily obtain the 12 feature variables during the initial triage, it can provide quick and early mortality prediction to clinicians for guiding deciding further treatment as well as helping explain the patient’s condition to family members.

Successful weaning from prolonged mechanical ventilation (MV) is an important issue in respiratory care centers (RCCs). Delayed or premature extubation increases both the risk of adverse outcomes and healthcare costs. However, the accurate evaluation of the timing of successful weaning from MV is very challenging in RCCs. This study aims to utilize artificial intelligence algorithms to build predictive models for the successful timing of the weaning of patients from MV in RCCs and to implement a dashboard with the best model in RCC settings. A total of 670 intubated patients in the RCC in Chi Mei Medical Center were included in the study. Twenty-six feature variables were selected to build the predictive models with artificial intelligence (AI)/machine-learning (ML) algorithms. An interactive dashboard with the best model was developed and deployed. A preliminary impact analysis was then conducted. Our results showed that all seven predictive models had a high area under the receiver operating characteristic curve (AUC), which ranged from 0.792 to 0.868. The preliminary impact analysis revealed that the mean number of ventilator days required for the successful weaning of the patients was reduced by 0.5 after AI intervention. The development of an AI prediction dashboard is a promising method to assist in the prediction of the optimal timing of weaning from MV in RCC settings. However, a systematic prospective study of AI intervention is still needed.

Chronic obstructive pulmonary disease (COPD) is one of the leading causes of mortality and contributes to high morbidity worldwide. Patients with COPD have a higher risk for acute respiratory failure, ventilator dependence, and mortality after hospitalization compared with the general population. Accurate and early risk detection will provide more information for early management and better decision making. This study aimed to build prediction models using patients’ characteristics, laboratory data, and comorbidities for early detection of acute respiratory failure, ventilator dependence, and mortality in patients with COPD after hospitalization. We retrospectively collected the electronic medical records of 5061 patients with COPD in three hospitals of the Chi Mei Medical Group, Taiwan. After data cleaning, we built three prediction models for acute respiratory failure, ventilator dependence, and mortality using seven machine learning algorithms. Based on the AUC value, the best model for mortality was built by the XGBoost algorithm (AUC = 0.817), the best model for acute respiratory failure was built by random forest algorithm (AUC = 0.804), while the best model for ventilator dependence was built by LightGBM algorithm (AUC = 0.809). A web service application was implemented with the best models and integrated into the existing hospital information system for physician’s trials and evaluations. Our machine learning models exhibit excellent predictive quality and can therefore provide physicians with a useful decision-making reference for the adverse prognosis of COPD patients.

Healthcare Artificial Intelligence (AI) has the greatest opportunity for development. Since healthcare and technology are two of Taiwan’s most competitive industries, the development of healthcare AI is an excellent chance for Taiwan to improve its health-related services. From the perspective of economic development, promoting healthcare AI must be a top priority. However, despite having many breakthroughs in research and pilot projects, healthcare AI is still considered rare and is broadly used in the healthcare setting. Based on a medical center in Taiwan that has introduced a variety of healthcare AI into practice, this study discussed and analyzed the issues and concerns in the development and scaling of medical AIs from the perspective of various stakeholders in the healthcare setting, including the government, healthcare institutions, users (healthcare workers), and AI providers. The present study also identified critical influential factors for the deployment and scaling of healthcare AI. It is hoped that this paper can serve as an important reference for the advancement of healthcare AI not only in Taiwan but also in other countries.

(1) Background: Effective healthcare collaboration not only improves the outcomes of patients, but also provides benefits to healthcare providers. A patient-centered communication platform, a so-called “one-stop platform”, is necessary to promote interprofessional collaboration (IPC) for optimal patient care. (2) Methods: Chi Mei Medical Center developed a patient-centered computerized platform to fulfill interprofessional collaboration needs. The platform features a spiral-shaped integrated care area and a communication area that allows the medical team to access patients’ information including the medical care they received within seven days, and veritably shows whether the team members have read communication messages. After pilot adoption, an online survey was conducted. (3) Results: A one-stop IPC platform was implemented and promoted for patient care. The online survey revealed that medical team members have high positive appraisal of the platform. It also pointed out that resistance to change among the medical team still has a significant impact on behavioral intention. (4) Conclusions: The interprofessional collaboration platform was recognized by the medical teams of Chi Mei Medical Center as an effective and convenient tool for assisting clinical decision making. However, actions to reduce user resistance to change and encourage collaboration among team members still need to be continued. Shared decision making within physicians and patients will be valuable to develop in the platform in the future.

A shape memory alloy (SMA) is an alloy which can eliminate deformation at lower temperatures and restore its original shape upon heating. SMAs have been receiving considerable attention in the research field of materials science, and their applications include robotics, automotive, aerospace, and biomedical industries. Observing the SMA’s shaping and restoration processes is important for understanding its working principles and applications. However, the transformation of its crystal structure with temperature can only be seen using special equipment, such as a transmission electron microscope (TEM), which is an expensive apparatus and the operation requires professional skills. In this study, a teaching module is designed using virtual reality (VR) technology and research results of an SMA to show its shape memory properties, shaping and restoration processes, as well as the real-life applications in an immersive and interactive way. A teaching experiment has been conducted to analyze students’ learning effectiveness using the teaching module (the experimental group) compared with that of using real SMA materials as the teaching aids (the control group). Two classes of students in the Department of Materials Science (one as the experimental group and the other as the control group) were selected as the samples by convenience sampling from a university in North Taiwan. The experimental group contained 52 students and the control group contained 70 students. A nonequivalent pretest-posttest design was adopted to explore whether the two groups had a significant difference in learning effectiveness. The experimental results reveal that the teaching module can improve the learning effectiveness significantly (p = 0.001), and the questionnaire results also show that a majority of the students had positive attitudes about the teaching module. They believed that it could increase their learning motivation and help them understand the properties and applications of the SMA.

Henoch-Schönlein purpura (HSP) is a disorder whose cause and pathogenesis is unknown; some familial cases of this disease have been reported. The clinical heterogeneity in HSP may be conferred by a number of genetic loci, including the major histocompatibility complex. The racial and genetic factors responsible for the occurrence of the familial cases of HSP in Taiwan are unclear. The purpose of this study was to examine the racial and genetic factors in familial HSP cases in Taiwan. We retrospectively collected the HSP cases in our hospital during 2006 through 2010 and observed that familial HSP cases were only in Taroko Aborigines. Six cases of HSP in 3 Taroko families were found, and their human leukocyte antigens (HLA) were studied in the tissue typing laboratory of our hospital, to determine the possible association with familial HSP cases in Taiwanese Aborigines. Our results suggest an increased frequency of familial HSP cases with HLA-A24 in Taiwanese Taroko Aborigines. We concluded that racial and genetic predisposition was the possible cause for the familial occurrence of and renal involvement in HSP in Taiwanese Aborigines.