Simultaneous bilateral video-assisted thoracic surgery is safe and feasible for multiple primary lung cancers.

BACKGROUND: The treatment for bilateral synchronous multiple primary lung cancers (MPLC) remains challenging. Simultaneous bilateral video-assisted thoracic surgery (VATS) may be an optimal treatment with curative intent, but its safety and feasibility are controversial. METHODS: One hundred and fifty-eight patients who underwent simultaneous bilateral VATS (simultaneous group) and 79 who underwent two-staged bilateral VATS (two-staged group) were included in this study. Their medical records were retrospectively reviewed and analyzed. RESULTS: The majority of patients were female and non-smokers. The most common surgical plan was lobectomy and contralateral wedge resection in both groups. There was no significant difference in the postoperative complication rate between the simultaneous groups and two-staged group (13.3% vs. 11.4%, p = 0.73). Patients who underwent simultaneous bilateral resection had shorter hospital stays, shorter anesthesia time and less chest drainage compared with those who underwent two-staged resection. Advanced TNM stage, complicated surgical plan and aggressive lymph node resection were risk factors for postoperative complications in simultaneous bilateral VATS. Patients in two groups had similar overall survival and disease free survival (p = 0.2). CONCLUSIONS: Simultaneous bilateral VATS for bilateral lung nodule resection is as safe and feasible as two-staged bilateral VATS. Patients who underwent simultaneous bilateral resection had similar or even better outcomes compared to that of the two-staged group. Simultaneous bilateral VATS is potentially an optimal treatment option for patients with erarly cTNM stage and good physical condition.

Positive fluid balance and poor outcomes after initial intensive care unit admission in sepsis resuscitation: a retrospective study.

Introduction: Fluid resuscitation of patients with sepsis is crucial. This study explored the role of fluid balance in the early resuscitation of sepsis patients in the intensive care unit (ICU). Material and methods: A retrospective study of patients with sepsis using the Peking Union Medical College Hospital Intensive Care Medical Information System and Database from January 2014 to June 2020 was performed. Based on the survival status on day 28, the training cohort was divided into an alive group (n = 1,803) and a deceased group (n = 429). Univariate and multivariate analyses were used to identify risk factors, and the integrated learning XGBoost algorithm was used to construct a model for predicting outcomes. ROC and Kaplan-Meier survival curves were used to evaluate the effectiveness of the model. A verification cohort (n = 433) was used to verify the model. Results: Univariate analysis showed that fluid balance is an important covariate. Based on the scatterplot distribution, a significant difference in mortality was determined between groups stratified with a balance of 1000 ml. There were associations in the multivariate analysis between poor outcomes and sex, PO2/FiO2, serum creatinine, FiO2, platelets, respiratory rate, SPO2, temperature, and total fluid volume (1000 ml). Among these variables, total fluid balance (1000 ml) had an OR of 1.98 (CI: 1.41-2.77, p < 0.001). Therefore, the model was built with these nine factors using XGBoost. Cross validation was used to verify generalizability. This model performed better than the SOFA and APACHE II models. The result was well verified in the verification cohort. A causal forest model suggested that patients with hypoxemia may suffer from positive fluid balance. Conclusions: Sepsis fluid resuscitation in the ICU should be a targeted and goal-oriented treatment. A new prognostic prediction model was constructed and indicated that a 6-hour positive fluid balance after ICU initial admission is a risk factor for poor outcomes in sepsis patients. A 6-hour fluid balance above 1000 ml should be performed with caution.

A BIGRU-Based Stacked Attention Network for Biomedical Named Entity Recognition with Chinese EMRs.

Biomedical named entity recognition (BNER) is an effective method to structure the medical text data. It is an important basic task for building the medical application services such as the medical knowledge graphs and the intelligent auxiliary diagnosis systems. Existing medical named entity recognition methods generally leverage the word embedding model to construct text representation, and then integrate multiple semantic understanding models to enhance the semantic understanding ability of the model to achieve high-performance entity recognition. However, in the medical field, there are many professional terms that rarely appear in the general field, which cannot be represented well by the general domain word embedding model. Second, existing approaches typically only focus on the extraction of global semantic features, which generate a loss of local semantic features between characters. Moreover, as the word embedding dimension becomes much higher, the standard single-layer structure fails to fully and deeply extract the global semantic features. We put forward the BIGRU-based Stacked Attention Network (BSAN) model for biomedical named entity recognition. Firstly, we use the large-scale real-world medical electronic medical record (EMR) data to fine-tune BERT to build the proprietary embedding representations of the medical terms. Second, we use the Convolutional Neural Network model to extract semantic features. Finally, a stacked BIGRU is constructed using a multi-layer structure and a novel stacking method. It not only enables comprehensive and in-depth extraction of global semantic features, but also requires less time. Experimentally validated on the real-world datasets in Chinese EMRs, the proposed BSAN model achieves 90.9% performance on F1-values, which is stronger than the BNER performance of other state-of-the-art models.

Multidimensional dynamic prediction model for hospitalized patients with the omicron variant in China.

Purpose: To establish dynamic prediction models by machine learning using daily multidimensional data for coronavirus disease 2019 (COVID-19) patients. Methods: Hospitalized COVID-19 patients at Peking Union Medical College Hospital from Nov 2nd, 2022, to Jan 13th, 2023, were enrolled in this study. The outcome was defined as deterioration or recovery of the patient's condition. Demographics, comorbidities, laboratory test results, vital signs, and treatments were used to train the model. To predict the following days, a separate XGBoost model was trained and validated. The Shapley additive explanations method was used to analyze feature importance. Results: A total of 995 patients were enrolled, generating 7228 and 3170 observations for each prediction model. In the deterioration prediction model, the minimum area under the receiver operating characteristic curve (AUROC) for the following 7 days was 0.786 (95% CI 0.721-0.851), while the AUROC on the next day was 0.872 (0.831-0.913). In the recovery prediction model, the minimum AUROC for the following 3 days was 0.675 (0.583-0.767), while the AUROC on the next day was 0.823 (0.770-0.876). The top 5 features for deterioration prediction on the 7th day were disease course, length of hospital stay, hypertension, and diastolic blood pressure. Those for recovery prediction on the 3rd day were age, D-dimer levels, disease course, creatinine levels and corticosteroid therapy. Conclusion: The models could accurately predict the dynamics of Omicron patients' conditions using daily multidimensional variables, revealing important features including comorbidities (e.g., hyperlipidemia), age, disease course, vital signs, D-dimer levels, corticosteroid therapy and oxygen therapy.

Measuring the worldwide spread of COVID-19 using a comprehensive modeling method.

BACKGROUND: With the global spread of COVID-19, detecting high-risk countries/regions timely and dynamically is essential; therefore, we sought to develop automatic, quantitative and scalable analysis methods to observe and estimate COVID-19 spread worldwide and further generate reliable and timely decision-making support for public health management using a comprehensive modeling method based on multiple mathematical models. METHODS: We collected global COVID-19 epidemic data reported from January 23 to September 30, 2020, to observe and estimate its possible spread trends. Countries were divided into three outbreak levels: high, middle, and low. Trends analysis was performed by calculating the growth rate, and then country grouping was implemented using group-based trajectory modeling on the three levels. Individual countries from each group were also chosen to further disclose the outbreak situations using two predicting models: the logistic growth model and the SEIR model. RESULTS: All 187 observed countries' trajectory subgroups were identified using two grouping strategies: with and without population consideration. By measuring epidemic trends and predicting the epidemic size and peak of individual countries, our study found that the logistic growth model generally estimated a smaller epidemic size than the SEIR model. According to SEIR modeling, confirmed cases in each country would take an average of 9-12 months to reach the outbreak peak from the day the first case occurred. Additionally, the average number of cases at the peak time will reach approximately 10-20% of the countries' populations, and the countries with high trends and a high predicted size must pay special attention and implement public health interventions in a timely manner. CONCLUSIONS: We demonstrated comprehensive observations and predictions of the COVID-19 outbreak in 187 countries using a comprehensive modeling method. The methods proposed in this study can measure COVID-19 development from multiple perspectives and are generalizable to other epidemic diseases. Furthermore, the methods also provide reliable and timely decision-making support for public health management.

Evaluation of ICUs and weight of quality control indicators: an exploratory study based on Chinese ICU quality data from 2015 to 2020.

This study aimed to explore key quality control factors that affected the prognosis of intensive care unit (ICU) patients in Chinese mainland over six years (2015-2020). The data for this study were from 31 provincial and municipal hospitals (3425 hospital ICUs) and included 2 110 685 ICU patients, for a total of 27 607 376 ICU hospitalization days. We found that 15 initially established quality control indicators were good predictors of patient prognosis, including percentage of ICU patients out of all inpatients (%), percentage of ICU bed occupancy of total inpatient bed occupancy (%), percentage of all ICU inpatients with an APACHE II score ⩾15 (%), three-hour (surviving sepsis campaign) SSC bundle compliance (%), six-hour SSC bundle compliance (%), rate of microbe detection before antibiotics (%), percentage of drug deep venous thrombosis (DVT) prophylaxis (%), percentage of unplanned endotracheal extubations (%), percentage of patients reintubated within 48 hours (%), unplanned transfers to the ICU (%), 48-h ICU readmission rate (%), ventilator associated pneumonia (VAP) (per 1000 ventilator days), catheter related blood stream infection (CRBSI) (per 1000 catheter days), catheter-associated urinary tract infections (CAUTI) (per 1000 catheter days), in-hospital mortality (%). When exploratory factor analysis was applied, the 15 indicators were divided into 6 core elements that varied in weight regarding quality evaluation: nosocomial infection management (21.35%), compliance with the Surviving Sepsis Campaign guidelines (17.97%), ICU resources (17.46%), airway management (15.53%), prevention of deep-vein thrombosis (14.07%), and severity of patient condition (13.61%). Based on the different weights of the core elements associated with the 15 indicators, we developed an integrated quality scoring system defined as F score=21.35%xnosocomial infection management + 17.97%xcompliance with SSC guidelines + 17.46%×ICU resources + 15.53%×airway management + 14.07%×DVT prevention + 13.61%×severity of patient condition. This evidence-based quality scoring system will help in assessing the key elements of quality management and establish a foundation for further optimization of the quality control indicator system.

Chinese experts' consensus on the application of intensive care big data.

The development of intensive care medicine is inseparable from the diversified monitoring data. Intensive care medicine has been closely integrated with data since its birth. Critical care research requires an integrative approach that embraces the complexity of critical illness and the computational technology and algorithms that can make it possible. Considering the need of standardization of application of big data in intensive care, Intensive Care Medicine Branch of China Health Information and Health Care Big Data Society, Standard Committee has convened expert group, secretary group and the external audit expert group to formulate Chinese Experts' Consensus on the Application of Intensive Care Big Data (2022). This consensus makes 29 recommendations on the following five parts: Concept of intensive care big data, Important scientific issues, Standards and principles of database, Methodology in solving big data problems, Clinical application and safety consideration of intensive care big data. The consensus group believes this consensus is the starting step of application big data in the field of intensive care. More explorations and big data based retrospective research should be carried out in order to enhance safety and reliability of big data based models of critical care field.

[Effects of different drying methods on ginsenosides based on UHPLC-Q-Exactive Orbitrap MS technique].

The present study quickly identified the ginsenosides in fresh Panax ginseng and specified the effects of different drying methods(50 ℃-drying, 80 ℃-drying, and-70 ℃ freeze-drying) on ginsenosides.Three P.ginseng products by different drying methods were prepared, and the UHPLC-Q-Exactive Orbitrap high-resolution liquid mass spectrometry(MS) technique was applied to perform gradient elution using water-acetonitrile as the mobile phase, and the data collected in the negative ion mode were analyzed using X Calibur 2.2.The results showed that 57 saponins were identified from fresh P.ginseng.As revealed by the comparison with the fresh P.ginseng, in terms of the loss of ginsenosides, the dried products were ranked as the dried product at 50 ℃, freeze-dried products at-70 ℃, and the dried product at 80 ℃ in the ascending order.This study elucidated the effects of different drying methods on the types and relative content of ginsenosides, which can provide references for the processing of P.ginseng in the producing areas.

Effects of Individualized Nursing Based on Zero-Defect Theory on Perioperative Patients Undergoing Laparoscopic Cholecystectomy.

Objective: This study is aimed at analyzing the effects of individualized nursing based on the zero-defect theory on perioperative patients undergoing laparoscopic cholecystectomy. Methods: 174 patients who underwent laparoscopic cholecystectomy from 1st November 2019 to 30th November 2020 were enrolled as the research subjects and randomly divided into control and observation groups. The patients in the control group received conventional perioperative nursing care, and the patients in the observation group were treated with individualized nursing based on the zero-defect theory. Results: The heart rate, diastolic blood pressure, and systolic blood pressure level of patients in two groups after nursing decreased significantly, and the reduction in the observation group was more significant than that in the control group. The depression and anxiety scores of the two groups after nursing were decreased, and the decrease in the observation group was significantly greater than that in the control group. The time to first postoperative exhaust, return to normal intake, out-of-bed activity, and hospital stay in the observation group was less than that in the control group. The incidence of postoperative complications in the observation group was substantially lower than that in the control group. The satisfaction degree of nursing care in the observation group was significantly higher than that in the control group. Conclusion: Individualized nursing care based on zero-defect theory can effectively reduce the perioperative psychological stress response of patients with laparoscopic cholecystectomy. It helps to improve the negative emotions of depression and anxiety, promotes the recovery of disease, reduces postoperative complications, and improves nursing satisfaction, which is worthy of clinical promotion.

Diagnostic accuracy of red blood cell distribution width for neonatal sepsis.

INTRODUCTION: Red blood cell distribution width (RDW) is a biomarker for the diagnosis and prognosis of many diseases. However, the relevance between RDW and neonatal sepsis (NS) have not reached a consensus yet; the perform of RDW in the diagnosis of neonatal sepsis is still not clear. The aim of this meta-analysis was to estimate the significance of RDW in neonatal sepsis and the perform of RDW in diagnosis of neonatal sepsis. EVIDENCE ACQUISITION: We used Pubmed, Embase, Web of science, CNKI and Google academic database to find all articles that met the inclusion criteria until July 1, 2020. EVIDENCE SYNTHESIS: Fifteen eligible studies involving 1362 newborns were included in the meta-analysis after two independent investigators read the title, abstract and full text in detail. The pooled result of this meta-analysis showed that RDW was significantly higher in the NS group than in the control group (WMD=3.224; 95%CI: 2.359-4.090, P<0.001). In addition, the overall pooled sensitivity, specificity, PLR, NLR and DOR were 0.88 (95%CI:0.66-0.96), 0.90 (95%CI:0.65-0.98), 9.2 (95%CI:2.1-40.3), 0.14(95%CI:0.04-0.43) and 66.9 (95%CI:8.73-513.26), respectively. The area under the SROC curve (AUC) was 0.95 (95%CI:0.93-0.96). CONCLUSIONS: The meta-analysis demonstrated that newborns with sepsis had an elevated RDW level than healthy controls. RDW levels have significant correlated with neonatal sepsis; and RDW can be used as a cheap and satisfactory diagnostic biomarker for neonatal sepsis with a relatively high performance.

Early Prediction of Mortality, Severity, and Length of Stay in the Intensive Care Unit of Sepsis Patients Based on Sepsis 3.0 by Machine Learning Models.

Background: Early prediction of the clinical outcome of patients with sepsis is of great significance and can guide treatment and reduce the mortality of patients. However, it is clinically difficult for clinicians. Methods: A total of 2,224 patients with sepsis were involved over a 3-year period (2016-2018) in the intensive care unit (ICU) of Peking Union Medical College Hospital. With all the key medical data from the first 6 h in the ICU, three machine learning models, logistic regression, random forest, and XGBoost, were used to predict mortality, severity (sepsis/septic shock), and length of ICU stay (LOS) (>6 days, ≤ 6 days). Missing data imputation and oversampling were completed on the dataset before introduction into the models. Results: Compared to the mortality and LOS predictions, the severity prediction achieved the best classification results, based on the area under the operating receiver characteristics (AUC), with the random forest classifier (sensitivity = 0.65, specificity = 0.73, F1 score = 0.72, AUC = 0.79). The random forest model also showed the best overall performance (mortality prediction: sensitivity = 0.50, specificity = 0.84, F1 score = 0.66, AUC = 0.74; LOS prediction: sensitivity = 0.79, specificity = 0.66, F1 score = 0.69, AUC = 0.76) among the three models. The predictive ability of the SOFA score itself was inferior to that of the above three models. Conclusions: Using the random forest classifier in the first 6 h of ICU admission can provide a comprehensive early warning of sepsis, which will contribute to the formulation and management of clinical decisions and the allocation and management of resources.

Early warning of citric acid overdose and timely adjustment of regional citrate anticoagulation based on machine learning methods.

BACKGROUND: Regional citrate anticoagulation (RCA) is an important local anticoagulation method during bedside continuous renal replacement therapy. To improve patient safety and achieve computer assisted dose monitoring and control, we took intensive care units patients into cohort and aiming at developing a data-driven machine learning model to give early warning of citric acid overdose and provide adjustment suggestions on citrate pumping rate and 10% calcium gluconate input rate for RCA treatment. METHODS: Patient age, gender, pumped citric acid dose value, 5% NaHCO3 solvent, replacement fluid solvent, body temperature value, and replacement fluid PH value as clinical features, models attempted to classify patients who received regional citrate anticoagulation into correct outcome category. Four models, Adaboost, XGBoost, support vector machine (SVM) and shallow neural network, were compared on the performance of predicting outcomes. Prediction results were evaluated using accuracy, precision, recall and F1-score. RESULTS: For classifying patients at the early stages of citric acid treatment, the accuracy of neutral networks model is higher than Adaboost, XGBoost and SVM, the F1-score of shallow neutral networks (90.77%) is overall outperformed than other models (88.40%, 82.17% and 88.96% for Adaboost, XGBoost and SVM). Extended experiment and validation were further conducted using the MIMIC-III database, the F1-scores for shallow neutral networks, Adaboost, XGBoost and SVM are 80.00%, 80.46%, 80.37% and 78.90%, the AUCs are 0.8638, 0.8086, 0.8466 and 0.7919 respectively. CONCLUSION: The results of this study demonstrated the feasibility and performance of machine learning methods for monitoring and adjusting local regional citrate anticoagulation, and further provide decision-making recommendations to clinicians point-of-care.

Selection strategy for sedation depth in critically ill patients on mechanical ventilation.

BACKGROUND: Analgesia and sedation therapy are commonly used for critically ill patients, especially mechanically ventilated patients. From the initial nonsedation programs to deep sedation and then to on-demand sedation, the understanding of sedation therapy continues to deepen. However, according to different patient's condition, understanding the individual patient's depth of sedation needs remains unclear. METHODS: The public open source critical illness database Medical Information Mart for Intensive Care III was used in this study. Latent profile analysis was used as a clustering method to classify mechanically ventilated patients based on 36 variables. Principal component analysis dimensionality reduction was used to select the most influential variables. The ROC curve was used to evaluate the classification accuracy of the model. RESULTS: Based on 36 characteristic variables, we divided patients undergoing mechanical ventilation and sedation and analgesia into two categories with different mortality rates, then further reduced the dimensionality of the data and obtained the 9 variables that had the greatest impact on classification, most of which were ventilator parameters. According to the Richmond-ASS scores, the two phenotypes of patients had different degrees of sedation and analgesia, and the corresponding ventilator parameters were also significantly different. We divided the validation cohort into three different levels of sedation, revealing that patients with high ventilator conditions needed a deeper level of sedation, while patients with low ventilator conditions required reduction in the depth of sedation as soon as possible to promote recovery and avoid reinjury. CONCLUSION: Through latent profile analysis and dimensionality reduction, we divided patients treated with mechanical ventilation and sedation and analgesia into two categories with different mortalities and obtained 9 variables that had the greatest impact on classification, which revealed that the depth of sedation was limited by the condition of the respiratory system.

A Clinical Prediction Model to Predict Heparin Treatment Outcomes and Provide Dosage Recommendations: Development and Validation Study.

BACKGROUND: Unfractionated heparin is widely used in the intensive care unit as an anticoagulant. However, weight-based heparin dosing has been shown to be suboptimal and may place patients at unnecessary risk during their intensive care unit stay. OBJECTIVE: In this study, we intended to develop and validate a machine learning-based model to predict heparin treatment outcomes and to provide dosage recommendations to clinicians. METHODS: A shallow neural network model was adopted in a retrospective cohort of patients from the Multiparameter Intelligent Monitoring in Intensive Care III (MIMIC III) database and patients admitted to the Peking Union Medical College Hospital (PUMCH). We modeled the subtherapeutic, normal, and supratherapeutic activated partial thromboplastin time (aPTT) as the outcomes of heparin treatment and used a group of clinical features for modeling. Our model classifies patients into 3 different therapeutic states. We tested the prediction ability of our model and evaluated its performance by using accuracy, the kappa coefficient, precision, recall, and the F1 score. Furthermore, a dosage recommendation module was designed and evaluated for clinical decision support. RESULTS: A total of 3607 patients selected from MIMIC III and 1549 patients admitted to the PUMCH who met our criteria were included in this study. The shallow neural network model showed results of F1 scores 0.887 (MIMIC III) and 0.925 (PUMCH). When compared with the actual dosage prescribed, our model recommended increasing the dosage for 72.2% (MIMIC III, 1240/1718) and 64.7% (PUMCH, 281/434) of the subtherapeutic patients and decreasing the dosage for 80.9% (MIMIC III, 504/623) and 76.7% (PUMCH, 277/361) of the supratherapeutic patients, suggesting that the recommendations can contribute to clinical improvements and that they may effectively reduce the time to optimal dosage in the clinical setting. CONCLUSIONS: The evaluation of our model for predicting heparin treatment outcomes demonstrated that the developed model is potentially applicable for reducing the misdosage of heparin and for providing appropriate decision recommendations to clinicians.

Noninvasive Real-Time Mortality Prediction in Intensive Care Units Based on Gradient Boosting Method: Model Development and Validation Study.

BACKGROUND: Monitoring critically ill patients in intensive care units (ICUs) in real time is vitally important. Although scoring systems are most often used in risk prediction of mortality, they are usually not highly precise, and the clinical data are often simply weighted. This method is inefficient and time-consuming in the clinical setting. OBJECTIVE: The objective of this study was to integrate all medical data and noninvasively predict the real-time mortality of ICU patients using a gradient boosting method. Specifically, our goal was to predict mortality using a noninvasive method to minimize the discomfort to patients. METHODS: In this study, we established five models to predict mortality in real time based on different features. According to the monitoring, laboratory, and scoring data, we constructed the feature engineering. The five real-time mortality prediction models were RMM (based on monitoring features), RMA (based on monitoring features and the Acute Physiology and Chronic Health Evaluation [APACHE]), RMS (based on monitoring features and Sequential Organ Failure Assessment [SOFA]), RMML (based on monitoring and laboratory features), and RM (based on all monitoring, laboratory, and scoring features). All models were built using LightGBM and tested with XGBoost. We then compared the performance of all models, with particular focus on the noninvasive method, the RMM model. RESULTS: After extensive experiments, the area under the curve of the RMM model was 0.8264, which was superior to that of the RMA and RMS models. Therefore, predicting mortality using the noninvasive method was both efficient and practical, as it eliminated the need for extra physical interventions on patients, such as the drawing of blood. In addition, we explored the top nine features relevant to real-time mortality prediction: invasive mean blood pressure, heart rate, invasive systolic blood pressure, oxygen concentration, oxygen saturation, balance of input and output, total input, invasive diastolic blood pressure, and noninvasive mean blood pressure. These nine features should be given more focus in routine clinical practice. CONCLUSIONS: The results of this study may be helpful in real-time mortality prediction in patients in the ICU, especially the noninvasive method. It is efficient and favorable to patients, which offers a strong practical significance.

Exploring Fever of Unknown Origin Intelligent Diagnosis Based on Clinical Data: Model Development and Validation.

BACKGROUND: Fever of unknown origin (FUO) is a group of diseases with heterogeneous complex causes that are misdiagnosed or have delayed diagnoses. Previous studies have focused mainly on the statistical analysis and research of the cases. The treatments are very different for the different categories of FUO. Therefore, how to intelligently diagnose FUO into one category is worth studying. OBJECTIVE: We aimed to fuse all of the medical data together to automatically predict the categories of the causes of FUO among patients using a machine learning method, which could help doctors diagnose FUO more accurately. METHODS: In this paper, we innovatively and manually built the FUO intelligent diagnosis (FID) model to help clinicians predict the category of the cause and improve the manual diagnostic precision. First, we classified FUO cases into four categories (infections, immune diseases, tumors, and others) according to the large numbers of different causes and treatment methods. Then, we cleaned the basic information data and clinical laboratory results and structured the electronic medical record (EMR) data using the bidirectional encoder representations from transformers (BERT) model. Next, we extracted the features based on the structured sample data and trained the FID model using LightGBM. RESULTS: Experiments were based on data from 2299 desensitized cases from Peking Union Medical College Hospital. From the extensive experiments, the precision of the FID model was 81.68% for top 1 classification diagnosis and 96.17% for top 2 classification diagnosis, which were superior to the precision of the comparative method. CONCLUSIONS: The FID model showed excellent performance in FUO diagnosis and thus would be a potentially useful tool for clinicians to enhance the precision of FUO diagnosis and reduce the rate of misdiagnosis.

Evaluation of the Secondary Transmission Pattern and Epidemic Prediction of COVID-19 in the Four Metropolitan Areas of China.

Understanding the transmission dynamics of COVID-19 is crucial for evaluating its spread pattern, especially in metropolitan areas of China, as its spread could lead to secondary outbreaks. In addition, the experiences gained and lessons learned from China have the potential to provide evidence to support other metropolitan areas and large cities outside China with their emerging cases. We used data reported from January 24, 2020, to February 23, 2020, to fit a model of infection, estimate the likely number of infections in four high-risk metropolitan areas based on the number of cases reported, and increase the understanding of the COVID-19 spread pattern. Considering the effect of the official quarantine regulations and travel restrictions for China, which began January 23~24, 2020, we used the daily travel intensity index from the Baidu Maps app to roughly simulate the level of restrictions and estimate the proportion of the quarantined population. A group of SEIR model statistical parameters were estimated using Markov chain Monte Carlo (MCMC) methods and fitting on the basis of reported data. As a result, we estimated that the basic reproductive number, R 0, was 2.91 in Beijing, 2.78 in Shanghai, 2.02 in Guangzhou, and 1.75 in Shenzhen based on the data from January 24, 2020, to February 23, 2020. In addition, we inferred the prediction results and compared the results of different levels of parameters. For example, in Beijing, the predicted peak number of cases was 467 with a peak time of March 01, 2020; however, if the city were to implement different levels (strict, moderate, or weak) of travel restrictions or regulation measures, the estimation results showed that the transmission dynamics would change and that the peak number of cases would differ by between 54% and 209%. We concluded that public health interventions would reduce the risk of the spread of COVID-19 and that more rigorous control and prevention measures would effectively contain its further spread, and awareness of prevention should be enhanced when businesses and social activities return to normal before the end of the epidemic. Further, the experiences gained and lessons learned from China offer the potential to provide evidence supporting other metropolitan areas and big cities with their emerging cases outside China.

Toward Optimal Heparin Dosing by Comparing Multiple Machine Learning Methods: Retrospective Study.

BACKGROUND: Heparin is one of the most commonly used medications in intensive care units. In clinical practice, the use of a weight-based heparin dosing nomogram is standard practice for the treatment of thrombosis. Recently, machine learning techniques have dramatically improved the ability of computers to provide clinical decision support and have allowed for the possibility of computer generated, algorithm-based heparin dosing recommendations. OBJECTIVE: The objective of this study was to predict the effects of heparin treatment using machine learning methods to optimize heparin dosing in intensive care units based on the predictions. Patient state predictions were based upon activated partial thromboplastin time in 3 different ranges: subtherapeutic, normal therapeutic, and supratherapeutic, respectively. METHODS: Retrospective data from 2 intensive care unit research databases (Multiparameter Intelligent Monitoring in Intensive Care III, MIMIC-III; e-Intensive Care Unit Collaborative Research Database, eICU) were used for the analysis. Candidate machine learning models (random forest, support vector machine, adaptive boosting, extreme gradient boosting, and shallow neural network) were compared in 3 patient groups to evaluate the classification performance for predicting the subtherapeutic, normal therapeutic, and supratherapeutic patient states. The model results were evaluated using precision, recall, F1 score, and accuracy. RESULTS: Data from the MIMIC-III database (n=2789 patients) and from the eICU database (n=575 patients) were used. In 3-class classification, the shallow neural network algorithm performed the best (F1 scores of 87.26%, 85.98%, and 87.55% for data set 1, 2, and 3, respectively). The shallow neural network algorithm achieved the highest F1 scores within the patient therapeutic state groups: subtherapeutic (data set 1: 79.35%; data set 2: 83.67%; data set 3: 83.33%), normal therapeutic (data set 1: 93.15%; data set 2: 87.76%; data set 3: 84.62%), and supratherapeutic (data set 1: 88.00%; data set 2: 86.54%; data set 3: 95.45%) therapeutic ranges, respectively. CONCLUSIONS: The most appropriate model for predicting the effects of heparin treatment was found by comparing multiple machine learning models and can be used to further guide optimal heparin dosing. Using multicenter intensive care unit data, our study demonstrates the feasibility of predicting the outcomes of heparin treatment using data-driven methods, and thus, how machine learning-based models can be used to optimize and personalize heparin dosing to improve patient safety. Manual analysis and validation suggested that the model outperformed standard practice heparin treatment dosing.

Exploiting Publication Contents and Collaboration Networks for Collaborator Recommendation.

Thanks to the proliferation of online social networks, it has become conventional for researchers to communicate and collaborate with each other. Meanwhile, one critical challenge arises, that is, how to find the most relevant and potential collaborators for each researcher? In this work, we propose a novel collaborator recommendation model called CCRec, which combines the information on researchers' publications and collaboration network to generate better recommendation. In order to effectively identify the most potential collaborators for researchers, we adopt a topic clustering model to identify the academic domains, as well as a random walk model to compute researchers' feature vectors. Using DBLP datasets, we conduct benchmarking experiments to examine the performance of CCRec. The experimental results show that CCRec outperforms other state-of-the-art methods in terms of precision, recall and F1 score.