Data Masking for Chinese Electronic Medical Records with Named Entity Recognition

With the rapid development of information technology,the electronifi-cation of medical records has gradually become a trend.In China,the population base is huge and the supporting medical institutions are numerous,so this reality drives the conversion of paper medical records to electronic medical records.Electronic medical records are the basis for establishing a smart hospital and an important guarantee for achieving medical intelligence,and the massive amount of electronic medical record data is also an important data set for conducting research in the medical field.However,electronic medical records contain a large amount of private patient information,which must be desensitized before they are used as open resources.Therefore,to solve the above problems,data masking for Chinese electronic medical records with named entity recognition is proposed in this paper.Firstly,the text is vectorized to satisfy the required format of the model input.Secondly,since the input sentences may have a long or short length and the relationship between sentences in context is not negligible.To this end,a neural network model for named entity recognition based on bidirectional long short-term memory(BiLSTM)with conditional random fields(CRF)is constructed.Finally,the data masking operation is performed based on the named entity recog-nition results,mainly using regular expression filtering encryption and principal component analysis(PCA)word vector compression and replacement.In addi-tion,comparison experiments with the hidden markov model(HMM)model,LSTM-CRF model,and BiLSTM model are conducted in this paper.The experi-mental results show that the method used in this paper achieves 92.72%Accuracy,92.30%Recall,and 92.51%F1_score,which has higher accuracy compared with other models.

Maturity Assessment of Hospital Information Systems Based on Electronic Medical Record Adoption Model (EMRAM)— Private Hospital Cases in Iran

Introduction: Today, information technology is considered as an important national development principle in each country which is applied in different fields. Health care as a whole and the hospitals could be regarded as a field and organizations with most remarkable IT applications respectively. Although different benchmarks and frameworks have been developed to assess different aspects of Hospital Information Systems (HISs) by various researchers, there is not any suitable reference model yet to benchmark HIS in the world. Electronic Medical Record Adoption Model (EMRAM) has been currently presented and is globally well-known to benchmark the rate of HIS utilization in the hospitals. Notwithstanding, this model has not been introduced in Iran so far. Methods: This research was carried out based on an applied descriptive method in three private hospitals of Isfahan—one of the most important provinces of Iran—in the year 2015. The purpose of this study was to investigate IT utilization stage in three selected private hospitals. Conclusion: The findings revealed that HIS is not at the center of concern in studied hospitals and is in the first maturity stage in accordance with EMRAM. However, hospital managers are enforced and under the pressure of different beneficiaries including insurance companies to improve their HIS. Therefore, it could be concluded that these types of hospitals are still far away from desirable conditions and need to enhance their IT utilization stage significantly.

Development of Medical Informatization in the Era of Big Data

The purpose of this paper is to discuss the development of medical informatization in the era of big data.Through literature review and theoretical analysis,the development of medical informatization in the era of big data is deeply discussed.The results show that medical informatization has developed rapidly in the era of big data,and its role in clinical decision-making,scientific research,teaching,and management has become increasingly prominent.The development of medical informatization in the era of big data has important purposes and methods,which can produce important results and conclusions and provide strong support for the development of the medical field.

A Method for Extracting Electronic Medical Record Entities by Fusing Multichannel Self-Attention Mechanism with Location Relationship Features

With the implementation of the“Internet+”strategy,electronic medi-cal records are generally applied in the medicalfield.Deep mining of electronic medical record content data is an effective means to obtain medical knowledge and analyse patients’states,but the existing methods for extracting entities from electronic medical records have problems of redundant information,overlapping entities,and low accuracy rates.Therefore,this paper proposes an entity extrac-tion method for electronic medical records based on the network framework of BERT-BiLSTM,which incorporates a multichannel self-attention mechanism and location relationship features.First,the text input sequence was encoded using the BERT-BiLSTM network framework,and the global semantic information of the sentence was mined more deeply using the multichannel self-attention mech-anism.Then,the position relation characteristic was used to extract the local semantic message of the text,and the position relation characteristic of the word and the position embedding matrix of the whole sentence were obtained.Next,the extracted global semantic information was stitched with the positional embedding matrix of the sentence to obtain the current entity classification matrix.Finally,the proposed method was validated on the dataset of Chinese medical text entity relationship extraction and the 2010i2b2/VA relationship corpus,and the exper-imental results indicate that the proposed method surpasses existing methods in terms of precision,recall,F1 value and training time.

Unidirectional Identity-Based Proxy Re-Signature with Key Insulation in EHR Sharing System

The introduction of the electronic medical record(EHR)sharing system has made a great contribution to the management and sharing of healthcare data.Considering referral treatment for patients,the original signature needs to be converted into a re-signature that can be verified by the new organization.Proxy re-signature(PRS)can be applied to this scenario so that authenticity and nonrepudiation can still be insured for data.Unfortunately,the existing PRS schemes cannot realize forward and backward security.Therefore,this paper proposes the first PRS scheme that can provide key-insulated property,which can guarantee both the forward and backward security of the key.Although the leakage of the private key occurs at a certain moment,the forward and backward key will not be attacked.Thus,the purpose of key insulation is implemented.What’s more,it can update different corresponding private keys in infinite time periods without changing the identity information of the user as the public key.Besides,the unforgeability of our scheme is proved based on the extended Computational Diffie-Hellman assumption in the random oracle model.Finally,the experimental simulation demonstrates that our scheme is feasible and in possession of promising properties.

Mobile EMR Use for Epidemiological Surveillance on a Medical Service Trip in Honduras: A Pilot Study

Rationale: Medical treatment on short-term primary care medical service trips (MSTs) is generally symptom-based and supplemented by point-of-care testing. This pilot study contributes to the effective planning for such austere settings based on predicted symptomology. Objective: We aimed to prospectively document the epidemiology of patients seen during two low-resource clinics on a MST in Honduras and apply predefined case definitions adapted from guidelines used by international healthcare organizations (e.g. World Health Organization). Methods: An observational design was used to track the epidemiology during two clinics on an MST in Limon, Honduras in March 2015. The QuickChart mobile electronic medical record (EMR) application was piloted to document diagnoses according to predefined case definitions. Results: The most commonly diagnosed syndromes were upper respiratory complaints (20.19%), nonspecific abdominal complaints (20.19%), general pain (15.38%), hypertension (9.62%), pruritus (6.73%), and asthma/ COPD (4.81%). The case definitions accounted for 94% of all complaints and diagnoses on the brigade. Discussion: The distribution of common patient diagnoses on this MST was similar to that which had been reported elsewhere. The use of broader symptom-based case definitions for epidemiologic surveillance could also facilitate the syndromic management of patients seen on MSTs, and improve the consistency of treatment offered. Conclusion: Case definitions for common syndromes on primary care MSTs may be a feasible method of standardizing patient management. Preliminary use of the QuickChart EMR was acceptable for documentation of epidemiology in the field. Further study is necessary to investigate the reliability of syndromic diagnostic criteria between different clinicians and in a variety of MST settings.

E_Breast: A Computerized Database Management System for Breast Diseases Patients in a Low Income Country

Objectives: To report our experience in using an electronic database for management of breast diseases in a developing country. Materials and methods: E-Breast is a database developed on FileMaker Pro Advanced to serve as patient file and breast diseases registry. The development of the platform, its usage and advantages on a manual filing system are described. Results: For 6 years, we use this database, which accounts more than 2000 patients and includes data from more than 10 years. An overview of the activity is easily generated by E-Breast. The generated reports are used to the routine care of patients, statistics and clinical research. Data entered are immediately useful in addition to simultaneously implement the database for clinical research. Many custom features are integrated. For research purposes, the system has the ability to perform detailed analyses on subsets defined by the user as breast cancer, breast benign diseases, etc. Conclusion: E-Breast has proven to be a useful way of documentation that has become an integral and essential part of the daily activity and also a valuable research tool.

Classification of metabolic-associated fatty liver disease subtypes based on TCM clinical phenotype

Objective:To classify the subtypes of metabolic-associated fatty liver disease(MAFLD)and provide new insights into the heterogeneity of MAFLD.Methods:Electronic medical records(EMR)of MAFLD diagnosed in accordance with the diagnostic criteria of Hubei Provincial Hospital of Traditional Chinese Medicine from 2016-2020 were included in the study.for physical annotation,and the data on each clinical phenotype was normalized according to corresponding aspirational standards.The MAFLD heterogeneous medical record network(HEMnet)was constructed using sex,age,disease diagnosis,symptoms,and Western medicine prescriptions as nodes and the co-occurrence times between phenotypes as edges.K-means clustering was used for disease classification.Relative risk(RR)was used to assess the specificity of each phenotype.Statistical methods were used to compare differences in laboratory indicators among subtypes.Results:A total of patients(12,626)with a mean age of 55.02(±14.21)years were included in the study.MAFLD can be divided into five subtypes:digestive diseases(C0),mental disorders and gynecological diseases(C1),chronic liver diseases and decompensated complications(C2),diabetes mellitus and its complications(C3),and immune joint system diseases(C4).Conclusions:Patients with MAFLD experience various symptoms and complications.The classification of MAFLD based on the HEMnet method is highly reliable.

Serum Sodium Fluctuation Prediction among ICU Patients Using Neural Network Algorithm:Analysis of the MIMIC-IV Database

Sodium homeostasis disorder is one of the most common abnormal symptoms of elderly patients in intensive care unit(ICU),which may lead to physiological disorders of many organs.The current prediction of serum sodium in ICU is mainly based on the subjective judgment of doctors’experience.This study aims at this problem by studying the clinical retrospective electronic medical record data of ICU to establish a machine learning model to predict the short-term serum sodium value of ICU patients.The data set used in this study is the open-source intensive care medical information set Medical Information Mart for Intensive Care(MIMIC)-IV.The time point of serum sodium detection was selected from the ICU clinical records,and the ICU records of 25risk factors related to serum sodium were extracted from the patients within the first 12 h for statistical analysis.A prediction model of serum sodium value within 48 h was established using a feedforward neural network,and compared with previous methods.Our research results show that the neural network learning model can predict the development of serum sodium in patients using physiological indicators recorded in clinical electronic medical records within 12 h,and has better prediction effect than the serum sodium formula and other machine learning models.

Overview of CCKS 2020 Task 3: Named Entity Recognition and Event Extraction in Chinese Electronic Medical Records

The China Conference on Knowledge Graph and Semantic Computing(CCKS)2020 Evaluation Task 3 presented clinical named entity recognition and event extraction for the Chinese electronic medical records.Two annotated data sets and some other additional resources for these two subtasks were provided for participators.This evaluation competition attracted 354 teams and 46 of them successfully submitted the valid results.The pre-trained language models are widely applied in this evaluation task.Data argumentation and external resources are also helpful.

Word Embedding Bootstrapped Deep Active Learning Method to Information Extraction on Chinese Electronic Medical Record

Electronic medical record (EMR) containing rich biomedical information has a great potential in disease diagnosis and biomedical research. However, the EMR information is usually in the form of unstructured text, which increases the use cost and hinders its applications. In this work, an effective named entity recognition (NER) method is presented for information extraction on Chinese EMR, which is achieved by word embedding bootstrapped deep active learning to promote the acquisition of medical information from Chinese EMR and to release its value. In this work, deep active learning of bi-directional long short-term memory followed by conditional random field (Bi-LSTM+CRF) is used to capture the characteristics of different information from labeled corpus, and the word embedding models of contiguous bag of words and skip-gram are combined in the above model to respectively capture the text feature of Chinese EMR from unlabeled corpus. To evaluate the performance of above method, the tasks of NER on Chinese EMR with “medical history” content were used. Experimental results show that the word embedding bootstrapped deep active learning method using unlabeled medical corpus can achieve a better performance compared with other models.

Secure and Efficient Data Storage and Sharing Scheme Based on Double Blockchain

In the digital era,electronic medical record(EMR)has been a major way for hospitals to store patients’medical data.The traditional centralized medical system and semi-trusted cloud storage are difficult to achieve dynamic balance between privacy protection and data sharing.The storage capacity of blockchain is limited and single blockchain schemes have poor scalability and low throughput.To address these issues,we propose a secure and efficient medical data storage and sharing scheme based on double blockchain.In our scheme,we encrypt the original EMR and store it in the cloud.The storage blockchain stores the index of the complete EMR,and the shared blockchain stores the index of the shared part of the EMR.Users with different attributes can make requests to different blockchains to share different parts according to their own permissions.Through experiments,it was found that cloud storage combined with blockchain not only solved the problem of limited storage capacity of blockchain,but also greatly reduced the risk of leakage of the original EMR.Content Extraction Signature(CES)combined with the double blockchain technology realized the separation of the privacy part and the shared part of the original EMR.The symmetric encryption technology combined with Ciphertext-Policy Attribute-Based Encryption(CP–ABE)not only ensures the safe storage of data in the cloud,but also achieves the consistency and convenience of data update,avoiding redundant backup of data.Safety analysis and performance analysis verified the feasibility and effectiveness of our scheme.

Comparative visual analytics for assessing medical records with sequence embedding

Machine learning for data-driven diagnosis has been actively studied in medicine to provide better healthcare.Supporting analysis of a patient cohort similar to a patient under treatment is a key task for clinicians to make decisions with high confidence.However,such analysis is not straightforward due to the characteristics of medical records:high dimensionality,irregularity in time,and sparsity.To address this challenge,we introduce a method for similarity calculation of medical records.Our method employs event and sequence embeddings.While we use an autoencoder for the event embedding,we apply its variant with the self-attention mechanism for the sequence embedding.Moreover,in order to better handle the irregularity of data,we enhance the self-attention mechanism with consideration of different time intervals.We have developed a visual analytics system to support comparative studies of patient records.To make a comparison of sequences with different lengths easier,our system incorporates a sequence alignment method.Through its interactive interface,the user can quickly identify patients of interest and conveniently review both the temporal and multivariate aspects of the patient records.We demonstrate the effectiveness of our design and system with case studies using a real-world dataset from the neonatal intensive care unit of UC Davis.

Optimization of the Classic Transfer-Stacking Model Migration Algorithm: A Way to Solve Time-Varying Performance Degradation of Acute Kidney Injury Clinical Prediction Model

Acute Kidney Injury (AKI) is one of the most common acute and critical illnesses in general wards and intensive care units. Its high morbidity and high fatality rate have become a major global public health problem. There are often serious lags in clinical diagnosis of AKI. Early diagnosis and timely intervention and effective care become critical. The use of electronic medical record data to build an AKI risk prediction model has been proven to help prevent the occurrence of AKI. However, in actual clinical applications, the distribution of historical data and new data will continue to vary over time, resulting in a significant decrease in the performance of the model. How to solve the problem of model performance degradation over time will be a core challenge for the long-term use of predictive models in clinical applications. Aiming at the above problems, this paper studies the classic Transfer-Stacking model migration algorithm. Aiming at the lack of this algorithm, such as the loss of a large amount of feature information of the target domain and poor fit when integrating the model of the target domain, the Accumulate-Transfer-Stacking algorithm is proposed to improve it. Improvements include: 1) Optimize the input vector and model integration algorithm of Transfer-Stacking’s target domain model. 2) Optimize Transfer-Stacking from a single-source domain model to a multi-source domain model. The experimental results show that for the improved algorithm proposed in this paper when the data is sufficient and insufficient, the average AUC value of the model on the data of subsequent years is 0.89 and 0.87, and the average F1 Score value is 0.45 and 0.36. Moreover, this method is significantly better than the unimproved Transfer-Stacking algorithm and baseline method, and can effectively overcome the problem of data distribution heterogeneity caused by time factors.

Translation in Data Mining to Advance Personalized Medicine for Health Equity

Personalized medicine is the development of “tailored” therapies that reflect traditional medical approaches with the incorporation of the patient’s unique genetic profile and the environmental basis of the disease. These individualized strategies encompass disease prevention and diagnosis, as well as treatment strategies. Today’s healthcare workforce is faced with the availability of massive amounts of patient- and disease-related data. When mined effectively, these data will help produce more efficient and effective diagnoses and treatment, leading to better prognoses for patients at both the individual and population level. Designing preventive and therapeutic interventions for those patients who will benefit most while minimizing side effects and controlling healthcare costs requires bringing diverse data sources together in an analytic paradigm. A resource to clinicians in the development and application of personalized medicine is largely facilitated, perhaps even driven, by the analysis of “big data”. For example, the availability of clinical data warehouses is a significant resource for clinicians in practicing personalized medicine. These “big data” repositories can be queried by clinicians, using specific questions, with data used to gain an understanding of challenges in patient care and treatment. Health informaticians are critical partners to data analytics including the use of technological infrastructures and predictive data mining strategies to access data from multiple sources, assisting clinicians’ interpretation of data and development of personalized, targeted therapy recommendations. In this paper, we look at the concept of personalized medicine, offering perspectives in four important, influencing topics: 1) the availability of “big data” and the role of biomedical informatics in personalized medicine, 2) the need for interdisciplinary teams in the development and evaluation of personalized therapeutic approaches, and 3) the impact of electronic medical record systems and clinical data warehouses on the field of personalized medicine. In closing, we present our fourth perspective, an overview to some of the ethical concerns related to personalized medicine and health equity.

Classification of Medical Image Notes for Image Labeling by Using MinBERT

The lack of labeled image data poses a serious challenge to the application of artificial intelligence(AI)in medical image diagnosis.Medical image notes contain valuable patient information that could be used to label images for machine learning tasks.However,most image note texts are unstructured with heterogeneity and short-paragraph characters,which fail traditional keyword-based techniques.We utilized a deep learning approach to recover missing labels for medical image notes automatically by using a combination of deep word embedding and deep neural network classifiers.Bidirectional encoder representations from transformers trained on medical image notes corpus(MinBERT)were proposed.We applied the proposed techniques to two typical classification tasks:Medical image type identification and clinical diagnosis identification.The two methods significantly outperformed baseline methods and presented high accuracies of 99.56%and 99.72%in image type identification and of 94.56%and 92.45%in clinical diagnosis identification.Visualization analysis further indicated that word embedding could efficiently capture semantic similarities and regularities across diverse expressions.Results indicated that our proposed framework could accurately recover the missing label information of medical images through the automatic extraction of electronic medical record information.Hence,it could serve as a powerful tool for exploring useful training data in various medical AI applications.

Integrated medical resource consumption stratification in hospitalized patients:an Auto Triage Management model based on accurate risk,cost and length of stay prediction

Triage management plays important roles in hospitalized patients for disease severity stratification and medical burden analysis.Although progression risks have been extensively researched for numbers of diseases,other crucial indicators that reflect patients’economic and time costs have not been systematically studied.To address the problems,we developed an automatic deep learning based Auto Triage Management(ATM)Framework capable of accurately modelling patients’disease progression risk and health economic evaluation.Based on them,we can first discover the relationship between disease progression and medical system cost,find potential features that can more precisely aid patient triage in resource allocation,and allow treatment plan searching that has cured patients.Applying ATM in COVID-19,we built a joint model to predict patients’risk,the total length of stay(Lo S)and cost when at-admission,and remaining Lo S and cost at a given hospitalized time point,with C-index0.930 and 0.869 for risk prediction,mean absolute error(MAE)of 5.61 and 5.90 days for total Lo S prediction in internal and external validation data.

Time to rebuild the doctor-patient relationship in China

In China,the doctor-patient relationship(DPR)has been tense in recent years and continues to deteriorate.From 2009 to 2018,295 severe medical violence events were reported on social media,in which 362 doctors were injured and 24 lost their lives(1).According to a survey conducted in 2018 by the national Chinese Medical Doctor Association(CMDA),62% of doctors had experienced varying degrees of medical disputes and 66% had experienced varying degrees of doctor-patient conflict,dominated by verbal violence(accounting for 51%of cases).

Symptom network topological features predict the effectiveness of herbal treatment for pediatric cough

Pediatric cough is a heterogeneous condition in terms of symptoms and the underlying disease mechanisms.Symptom phenotypes hold complicated interactions between each other to form an intricate network structure.This study aims to investigate whether the network structure of pediatric cough symptoms is associated with the prognosis and outcome of patients.A total of 384 cases were derived from the electronic medical records of a highly experienced traditional Chinese medicine (TCM) physician.The data were divided into two groups according to the therapeutic effect,namely,an invalid group (group A with 40 cases of poor efficacy) and a valid group (group B with 344 cases of good efficacy).Several well-established analysis methods,namely,statistical test,correlation analysis,and complex network analysis,were used to analyze the data.This study reports that symptom networks of patients with pediatric cough are related to the effectiveness of treatment: a dense network of symptoms is associated with great difficulty in treatment.Interventions with the most different symptoms in the symptom network may have improved therapeutic effects.

Semi-Supervised Noisy Label Learning for Chinese Clinical Named Entity Recognition

This paper describes our approach for the Chinese clinical named entity recognition(CNER) task organized by the 2020 China Conference on Knowledge Graph and Semantic Computing(CCKS) competition. In this task, we need to identify the entity boundary and category labels of six entities from Chinese electronic medical record(EMR). We constructed a hybrid system composed of a semi-supervised noisy label learning model based on adversarial training and a rule post-processing module. The core idea of the hybrid system is to reduce the impact of data noise by optimizing the model results. Besides, we used post-processing rules to correct three cases of redundant labeling, missing labeling, and wrong labeling in the model prediction results. Our method proposed in this paper achieved strict criteria of 0.9156 and relax criteria of 0.9660 on the final test set, ranking first.