Chinese Electronic Health Records Research Articles

Automatic quantitative stroke severity assessment based on Chinese clinical named entity recognition with domain-adaptive pre-trained large language model

Background:Stroke is a prevalent disease with a significant global impact. Effective assessment of stroke severity is vital for an accurate diagnosis, appropriate treatment, and optimal clinical outcomes. The National Institutes of Health Stroke Scale (NIHSS) is a widely used scale for quantitatively assessing stroke severity. However, the current manual scoring of NIHSS is labor-intensive, time-consuming, and sometimes unreliable. Applying artificial intelligence (AI) techniques to automate the quantitative assessment of stroke on vast amounts of electronic health records (EHRs) has attracted much interest. Objective:This study aims to develop an automatic, quantitative stroke severity assessment framework through automating the entire NIHSS scoring process on Chinese clinical EHRs. Methods:Our approach consists of two major parts: Chinese clinical named entity recognition (CNER) with a domain-adaptive pre-trained large language model (LLM) and automated NIHSS scoring. To build a high-performing CNER model, we first construct a stroke-specific, densely annotated dataset “Chinese Stroke Clinical Records” (CSCR) from EHRs provided by our partner hospital, based on a stroke ontology that defines semantically related entities for stroke assessment. We then pre-train a Chinese clinical LLM coined “CliRoberta” through domain-adaptive transfer learning and construct a deep learning-based CNER model that can accurately extract entities directly from Chinese EHRs. Finally, an automated, end-to-end NIHSS scoring pipeline is proposed by mapping the extracted entities to relevant NIHSS items and values, to quantitatively assess the stroke severity. Results:Results obtained on a benchmark dataset CCKS2019 and our newly created CSCR dataset demonstrate the superior performance of our domain-adaptive pre-trained LLM and the CNER model, compared with the existing benchmark LLMs and CNER models. The high F1 score of 0.990 ensures the reliability of our model in accurately extracting the entities for the subsequent automatic NIHSS scoring. Subsequently, our automated, end-to-end NIHSS scoring approach achieved excellent inter-rater agreement (0.823) and intraclass consistency (0.986) with the ground truth and significantly reduced the processing time from minutes to a few seconds. Conclusion:Our proposed automatic and quantitative framework for assessing stroke severity demonstrates exceptional performance and reliability through directly scoring the NIHSS from diagnostic notes in Chinese clinical EHRs. Moreover, this study also contributes a new clinical dataset, a pre-trained clinical LLM, and an effective deep learning-based CNER model. The deployment of these advanced algorithms can improve the accuracy and efficiency of clinical assessment, and help improve the quality, affordability and productivity of healthcare services.

The Impact of Prostate-Specific Antigen Screening on Prostate Cancer Incidence and Mortality in China: 13-Year Prospective Population-Based Cohort Study.

The status of prostate-specific antigen (PSA) screening is unclear in China. Evidence regarding the optimal frequency and interval of serial screening for prostate cancer (PCa) is disputable. This study aimed to depict the status of PSA screening and to explore the optimal screening frequency for PCa in China. A 13-year prospective cohort study was conducted using the Chinese Electronic Health Records Research in Yinzhou study's data set. A total of 420,941 male participants aged ≥45 years were included between January 2009 and June 2022. Diagnosis of PCa, cancer-specific death, and all-cause death were obtained from the electronic health records and vital statistic system. Hazard ratios (HRs) with 95% CIs were estimated using Cox regression analysis. The cumulative rate of ever PSA testing was 17.9% with an average annual percent change (AAPC) of 8.7% (95% CI 3.6%-14.0%) in the past decade in China. People with an older age, a higher BMI, higher waist circumference, tobacco smoking and alcohol drinking behaviors, higher level of physical activity, medication use, and comorbidities were more likely to receive PSA screening, whereas those with a lower education level and a widowed status were less likely to receive the test. People receiving serial screening ≥3 times were at a 67% higher risk of PCa detection (HR 1.67; 95% CI 1.48-1.88) but a 64% lower risk of PCa-specific mortality (HR 0.36; 95% CI 0.18-0.70) and a 28% lower risk of overall mortality (HR 0.72; 95% CI 0.67-0.77). People following a serial screening strategy at least once every 4 years were at a 25% higher risk of PCa detection (HR 1.25; 95% CI 1.13-1.36) but 70% (HR 0.30; 95% CI 0.16-0.57) and 23% (HR 0.77; 95% CI 0.73-0.82) lower risks of PCa-specific and all-cause mortality, respectively. This study reveals a low coverage of PSA screening in China and provides the first evidence of its benefits in the general Chinese population. The findings of this study indicate that receiving serial screening at least once every 4 years is beneficial for overall and PCa-specific survival. Further studies based on a nationwide population and with long-term follow-up are warranted to identify the optimal screening interval in China.

GWBNER: A named entity recognition method based on character glyph and word boundary features for Chinese EHRs

Electronic Health Records (EHRs) contain unprecedented volumes of data related to health such as diagnosis and treatment information. Based on the technologies of named entity recognition (NER), EHRs mining has become a research focus in health domain. Nevertheless, the complex medical entities are challenging to be recognized, especially in Chinese EHRs. In this paper, a Glyph and Word Boundary-based Named Entity Recognition (GWBNER) method is presented, which takes into account both the Chinese character glyph and word boundary features in Chinese EHRs. Specifically, the character glyphs are utilized to capture the character-level global structural features, and the word boundaries are adopted to extract the word-level local structural features in Chinese. Therefore, the medical entities are fully recognized with rich semantic features from diverse perspectives in Chinese EHR texts. Finally, we conduct extensive experiments to evaluate the novel approach, and the F1 scores of 0.879, 0.998 and 0.962 are achieved on three EHR datasets, respectively. The experimental results demonstrate the optimal performance of the GWBNER method compared with the state-of-the-art models in NER community.

MP44-19 EFFECTS OF PROSTATE-SPECIFIC ANTIGEN SCREENING ON PROSTATE CANCER INCIDENCE AND MORTALITY: A POPULATION-BASED COHORT STUDY IN CHINA

You have accessJournal of UrologyCME1 Apr 2023MP44-19 EFFECTS OF PROSTATE-SPECIFIC ANTIGEN SCREENING ON PROSTATE CANCER INCIDENCE AND MORTALITY: A POPULATION-BASED COHORT STUDY IN CHINA Yongle Zhan, Xiaohao Ruan, Da Huang, Jingyi Huang, Jinlun Huang, Tsun Tsun Chun, Brian Ho, Ada Ng, James Tsu, and Rong Na Yongle ZhanYongle Zhan More articles by this author , Xiaohao RuanXiaohao Ruan More articles by this author , Da HuangDa Huang More articles by this author , Jingyi HuangJingyi Huang More articles by this author , Jinlun HuangJinlun Huang More articles by this author , Tsun Tsun ChunTsun Tsun Chun More articles by this author , Brian HoBrian Ho More articles by this author , Ada NgAda Ng More articles by this author , James TsuJames Tsu More articles by this author , and Rong NaRong Na More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003290.19AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Evidence regarding the benefit of prostate-specific antigen (PSA) testing for prostate cancer (PCa) is limited in China. Scarce research so far explores the optimal frequency and interval of repeated PSA testing. This study aims to evaluate the effects of serial screening in different frequencies and intervals against PCa incidence and mortality among Chinese population. METHODS: Data of this study came from the Chinese Electronic Health Records Research in Yinzhou (CHERRY) Study. Briefly, the CHERRY study is a population-based cohort study comprising more than one million adults enrolled from 1 January 2009 in Yinzhou, China. The population census, primary care, electronic medical records, health checks, surveillance, and death registry databases have been integrated into the CHERRY study. A total of 420,941 male participants aged over 45 at recruitment were included in the present study. Among them, a subgroup of 75,362 men who have undergone any PSA testing was selected for analysis. The number of receiving PSA screening, screening frequency and interval, PCa diagnosis and death were extracted from the electronic health records system. RESULTS: People receiving more than three times of PSA testing had 67% (95% confidence interval [CI]: 48-88%) higher risks of PCa detection, but 64% (95%CI: 30-82%) lower risks of PCa-specific mortality and 28% (95%CI: 23-33%) lower risks of overall mortality. Receiving at least one screening per four years could increase 25% (95%CI: 13-36%) risks of PCa detection, but decrease 70% (95%CI: 43-84%) and 23% (95%CI: 18-27%) risks of PCa-specific and overall mortality. For those repeated attenders, 2-yr to 3-yr intervals could decrease 50-85% risks of PCa-specific mortality and 9-57% risks of overall mortality (Figure 1). CONCLUSIONS: This study provides the first evidence for the benefits of PSA screening in Chinese general population, and supports that receiving serial screening at least once per four-year is beneficial for overall and prostate cancer-specific survival. Source of Funding: This work was in supported by grants from National Natural Science Foundation of China (Grant No. 81972645), Shanghai Youth Talent Support Program, intramural funding of The University of Hong Kong to Dr. Rong Na, and Shanghai Sailing Program (22YF1440500) to Dr. Da Huang © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e619 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Yongle Zhan More articles by this author Xiaohao Ruan More articles by this author Da Huang More articles by this author Jingyi Huang More articles by this author Jinlun Huang More articles by this author Tsun Tsun Chun More articles by this author Brian Ho More articles by this author Ada Ng More articles by this author James Tsu More articles by this author Rong Na More articles by this author Expand All Advertisement PDF downloadLoading ...

Diagnostic validation and development of an algorithm for identification of intussusception in children using electronic health records of Ningbo city in China

ABSTRACTBackgroundMonitoring the risk of intussusception after the introduction of rotavirus vaccines is recommended by the World Health Organization (WHO). Although the validity of intussusception monitoring using electronic health records (EHRs) has been confirmed previously, no similar studies have been conducted in China. We aimed to verify the diagnosis and determine an algorithm with the best performance for identification of intussusception using Chinese EHR databases.Research design and methodsUsing the Regional Health Information Platform in Ningbo, patients aged 0–72 months from 2015 to 2021 with any related visits for intussusception were included. The algorithms were based on diagnostic codes or keywords in different clinical scenarios, and their performance was evaluated with positive predictive value (PPV) and sensitivity in line with the Brighton guidelines.ResultsBrighton level 1 intussusception was confirmed in 2958 patients with 3246 episodes. Fine-tuned algorithms combining the appearance of the relevant ICD-10 codes or the Chinese keyword ‘Chang Tao’ in any diagnostic reports with the results of enema treatments or related surgeries showed the highest sensitivity, while the highest PPV was obtained by further criteria based on typical radiographic appearances.ConclusionIntussusception could be identified and validated internally using EHRs in Ningbo.

An Efficient Method for Deidentifying Protected Health Information in Chinese Electronic Health Records: Algorithm Development and Validation.

BackgroundWith the popularization of electronic health records in China, the utilization of digitalized data has great potential for the development of real-world medical research. However, the data usually contains a great deal of protected health information and the direct usage of this data may cause privacy issues. The task of deidentifying protected health information in electronic health records can be regarded as a named entity recognition problem. Existing rule-based, machine learning–based, or deep learning–based methods have been proposed to solve this problem. However, these methods still face the difficulties of insufficient Chinese electronic health record data and the complex features of the Chinese language.ObjectiveThis paper proposes a method to overcome the difficulties of overfitting and a lack of training data for deep neural networks to enable Chinese protected health information deidentification.MethodsWe propose a new model that merges TinyBERT (bidirectional encoder representations from transformers) as a text feature extraction module and the conditional random field method as a prediction module for deidentifying protected health information in Chinese medical electronic health records. In addition, a hybrid data augmentation method that integrates a sentence generation strategy and a mention-replacement strategy is proposed for overcoming insufficient Chinese electronic health records.ResultsWe compare our method with 5 baseline methods that utilize different BERT models as their feature extraction modules. Experimental results on the Chinese electronic health records that we collected demonstrate that our method had better performance (microprecision: 98.7%, microrecall: 99.13%, and micro-F1 score: 98.91%) and higher efficiency (40% faster) than all the BERT-based baseline methods.ConclusionsCompared to baseline methods, the efficiency advantage of TinyBERT on our proposed augmented data set was kept while the performance improved for the task of Chinese protected health information deidentification.

PIAT: An Evolutionarily Intelligent System for Deep Phenotyping of Chinese Electronic Health Records

Electronic health record (EHR) resources are valuable but remain underexplored because most clinical information, especially phenotype information, is buried in the free text of EHRs. An intelligent annotation tool plays an important role in unlocking the full potential of EHRs by transforming free-text phenotype information into a computer-readable form. Deep phenotyping has shown its advantage in representing phenotype information in EHRs with high fidelity; however, most existing annotation tools are not suitable for the deep phenotyping task. Here, we developed an intelligent annotation tool named PIAT with a major focus on the deep phenotyping of Chinese EHRs. PIAT can improve the annotation efficiency for EHR-based deep phenotyping with a simple but effective interactive interface, automatic preannotation support, and a learning mechanism. Specifically, experts can proofread automatic annotation results from the annotation algorithm in the web-based interactive interface, and EHRs reviewed by experts can be used for evolving the underlying annotation algorithm. In this way, the annotation process of deep phenotyping EHRs will become easier. In conclusion, we create a powerful intelligent system for the deep phenotyping of Chinese EHRs. It is hoped that our work will inspire further studies in constructing intelligent systems for deep phenotyping English and non-English EHRs.

Applications of the NDR and DIAL models for risk prediction on cardiovascular disease in patients with type 2 diabetes in Ningbo

Objective: To validate the performance of cardiovascular risk prediction models based on the Sweden National Diabetes Register (NDR) and Diabetes Lifetime-perspective prediction (DIAL) model for assessing risks of 5-year and 10-year cardiovascular disease (CVD) among Chinese patients with type 2 diabetes. Methods: Based on the Chinese Electronic Health Records Research in Yinzhou study, 83 503 patients with type 2 diabetes aged 30-75 years without a history of CVD at baseline were included from January 1, 2010 to December 31, 2020. Recalibrated NDR model was used to estimate 5-year risk, while the recalibrated DIAL model was used to predict 5-year and 10-year risks. The competing events adjusted Kaplan-Meier analysis was used to obtain the observed cardiovascular events. Discrimination C statistics evaluated model accuracy, calibration χ2 value, and calibration plots. Results: Through a median follow-up of 7.0 years, 7 326 cardiovascular events, and 2 937 non-vascular deaths were identified among a total of 83 503 subjects. The recalibrated NDR model overestimated 5-year risk by 39.4% in men and 8.6% in women, whereas the overestimation for the recalibrated DIAL model was 14.6% in men and 50.1% in women. The DIAL model had a better discriminative ability (C-statistic=0.681, 95%CI: 0.672-0.690) than NDR model (C-statistic=0.667, 95%CI: 0.657-0.677) in 5-year risk prediction for men, and the models had a similar ability for women (C-statistic=0.699, 95%CI: 0.690-0.708 for NDR and C-statistic=0.698, 95%CI: 0.689-0.706 for DIAL). The prediction accuracy of the DIAL model was improved in the 10-year risk, with the underestimation being 1.6% for men and the overestimation being 12.8% for women. Conclusions: Both recalibrated NDR and DIAL models overestimated 5-year cardiovascular risk in Chinese patients with type 2 diabetes, while the higher overestimation was shown using the DIAL model. However, the improvement was found in predicting 10-year CVD risk using the DIAL model, which suggested the value of lifetime risk prediction and indicated the need for research on the lifetime risk prediction model for cardiovascular risk assessment in Chinese patients with type 2 diabetes.

Deep Phenotyping of Chinese Electronic Health Records by Recognizing Linguistic Patterns of Phenotypic Narratives With a Sequence Motif Discovery Tool: Algorithm Development and Validation.

BackgroundPhenotype information in electronic health records (EHRs) is mainly recorded in unstructured free text, which cannot be directly used for clinical research. EHR-based deep-phenotyping methods can structure phenotype information in EHRs with high fidelity, making it the focus of medical informatics. However, developing a deep-phenotyping method for non-English EHRs (ie, Chinese EHRs) is challenging. Although numerous EHR resources exist in China, fine-grained annotation data that are suitable for developing deep-phenotyping methods are limited. It is challenging to develop a deep-phenotyping method for Chinese EHRs in such a low-resource scenario.ObjectiveIn this study, we aimed to develop a deep-phenotyping method with good generalization ability for Chinese EHRs based on limited fine-grained annotation data.MethodsThe core of the methodology was to identify linguistic patterns of phenotype descriptions in Chinese EHRs with a sequence motif discovery tool and perform deep phenotyping of Chinese EHRs by recognizing linguistic patterns in free text. Specifically, 1000 Chinese EHRs were manually annotated based on a fine-grained information model, PhenoSSU (Semantic Structured Unit of Phenotypes). The annotation data set was randomly divided into a training set (n=700, 70%) and a testing set (n=300, 30%). The process for mining linguistic patterns was divided into three steps. First, free text in the training set was encoded as single-letter sequences (P: phenotype, A: attribute). Second, a biological sequence analysis tool—MEME (Multiple Expectation Maximums for Motif Elicitation)—was used to identify motifs in the single-letter sequences. Finally, the identified motifs were reduced to a series of regular expressions representing linguistic patterns of PhenoSSU instances in Chinese EHRs. Based on the discovered linguistic patterns, we developed a deep-phenotyping method for Chinese EHRs, including a deep learning–based method for named entity recognition and a pattern recognition–based method for attribute prediction.ResultsIn total, 51 sequence motifs with statistical significance were mined from 700 Chinese EHRs in the training set and were combined into six regular expressions. It was found that these six regular expressions could be learned from a mean of 134 (SD 9.7) annotated EHRs in the training set. The deep-phenotyping algorithm for Chinese EHRs could recognize PhenoSSU instances with an overall accuracy of 0.844 on the test set. For the subtask of entity recognition, the algorithm achieved an F1 score of 0.898 with the Bidirectional Encoder Representations from Transformers–bidirectional long short-term memory and conditional random field model; for the subtask of attribute prediction, the algorithm achieved a weighted accuracy of 0.940 with the linguistic pattern–based method.ConclusionsWe developed a simple but effective strategy to perform deep phenotyping of Chinese EHRs with limited fine-grained annotation data. Our work will promote the second use of Chinese EHRs and give inspiration to other non–English-speaking countries.

Prediction of acute kidney injury after cardiac surgery: model development using a Chinese electronic health record dataset

BackgroundAcute kidney injury (AKI) is a major complication following cardiac surgery that substantially increases morbidity and mortality. Current diagnostic guidelines based on elevated serum creatinine and/or the presence of oliguria potentially delay its diagnosis. We presented a series of models for predicting AKI after cardiac surgery based on electronic health record data.MethodsWe enrolled 1457 adult patients who underwent cardiac surgery at Nanjing First Hospital from January 2017 to June 2019. 193 clinical features, including demographic characteristics, comorbidities and hospital evaluation, laboratory test, medication, and surgical information, were available for each patient. The number of important variables was determined using the sliding windows sequential forward feature selection technique (SWSFS). The following model development methods were introduced: extreme gradient boosting (XGBoost), random forest (RF), deep forest (DF), and logistic regression. Model performance was accessed using the area under the receiver operating characteristic curve (AUROC). We additionally applied SHapley Additive exPlanation (SHAP) values to explain the RF model. AKI was defined according to Kidney Disease Improving Global Outcomes guidelines.ResultsIn the discovery set, SWSFS identified 16 important variables. The top 5 variables in the RF importance matrix plot were central venous pressure, intraoperative urine output, hemoglobin, serum potassium, and lactic dehydrogenase. In the validation set, the DF model exhibited the highest AUROC (0.881, 95% confidence interval [CI] 0.831–0.930), followed by RF (0.872, 95% CI 0.820–0.923) and XGBoost (0.857, 95% CI 0.802–0.912). A nomogram model was constructed based on intraoperative longitudinal features, achieving an AUROC of 0.824 (95% CI 0.763–0.885) in the validation set. The SHAP values successfully illustrated the positive or negative contribution of the 16 variables attributed to the output of the RF model and the individual variable’s effect on model prediction.ConclusionsOur study identified 16 important predictors and provided a series of prediction models to enhance risk stratification of AKI after cardiac surgery. These novel predictors might aid in choosing proper preventive and therapeutic strategies in the perioperative management of AKI patients.

Clinical Outcome and Medical Cost of Originator and Generic Antihypertensive Drugs: A Population-Based Study in Yinzhou, China.

Background: The substitution of generic drugs can effectively alleviate the rapid growth of drug costs; however, the clinical effectiveness and medical costs of originator products and generics were barely studied in China. Objectives: To compare the effectiveness of antihypertensive drugs and hypertension-related medical costs between originator and generic initiators in Yinzhou, China. Methods: We conducted a population-based retrospective cohort study using the Chinese Electronic Health Records Research in Yinzhou (CHERRY), from July 1, 2011, to December 31, 2018. Hypertension patients initiating with originator products were compared with patients initiating with generic counterparts. We used 1:1 propensity score matching to pair the two groups based on sociodemographic, clinical, and health service utilization variables. Cox proportional regression was adopted to compare the rate of hospitalization for hypertension-related cardiovascular disease between matched originator and generic initiators. Wilcoxon matched-pairs signed-rank test was used to compare annual hypertension-related medical costs. Results: Matched pairs (10,535) of patients were included in the comparative study of originator products and generics, corresponding to seven antihypertensive drugs including amlodipine, felodipine, nifedipine, irbesartan, losartan, valsartan, and metoprolol. The average age of patients included in the analysis was around 60 years (originator vs. generics initiators: from 59.0 vs. 59.1 years in losartan to 62.9 vs. 63.6 years in nifedipine). Higher hospitalization rates among originator initiators were observed for three calcium channel blockers (hazard ratio[95% CI]: amlodipine, 3.18[1.43, 7.11]; felodipine, 3.60[1.63, 7.98]; and nifedipine, 3.86[1.26, 11.81]; respectively). The remaining four out of seven drugs of the clinical endpoint estimates showed comparable outcomes between originator products and generics (hazard ratio[95% CI]: irbesartan, 1.19[0.50, 2.84]; losartan, 1.84[0.84, 4.07]; valsartan, 2.04[0.72, 5.78]; and metoprolol, 1.25[0.56, 2.80]; respectively). Higher median annual hypertension-related medical costs were observed in originator initiators (all p < 0.001), except for metoprolol (p = 0.646). Conclusion: We observed comparable or even better clinical outcomes and less medical cost associated with the use of antihypertensive generics compared to originator counterparts. This could help increase patient and provider confidence in the efficacy of generic medicines to manage hypertension diseases.

Disease prediction based on multi-type data fusion from Chinese electronic health record.

Disease prediction by using a variety of healthcare data to assist doctors in disease diagnosis is becoming a more and more important research topic recently. This paper proposes a disease prediction model that fuses multiple types of encoded representations of Chinese electronic health records (EHRs). The model framework utilizes a multi-head self-attention mechanism, which combines textual and numerical features to enhance text representations. The BiLSTM-CRF and TextCNN models are used, respectively, to extract entities and then obtain the embedding representations of them. The representations of text and entities in it are combined together for formulating representations of EHRs. The experimental results on EHRs data collected from a Three Grade Class B Hospital General in Gansu Province, China, show that our model achieved an F1 score of 91.92%, which outperforms the previous baseline methods.

Unsupervised multi-granular Chinese word segmentation and term discovery via graph partition.

This study aims at realizing unsupervised term discovery in Chinese electronic health records (EHRs) by using the word segmentation technique. The existing supervised algorithms do not perform satisfactorily in the case of EHRs, as annotated medical data are scarce. We propose an unsupervised segmentation method (GTS) based on the graph partition principle, whose multi-granular segmentation capability can help realize efficient term discovery. A sentence is converted to an undirected graph, with the edge weights based on n-gram statistics, and ratio cut is used to split the sentence into words. The graph partition is solved efficiently via dynamic programming, and multi-granularity is realized by setting different partition numbers. A BERT-based discriminator is trained using generated samples to verify the correctness of the word boundaries. The words that are not recorded in existing dictionaries are retained as potential medical terms. We compared the GTS approach with mature segmentation systems for both word segmentation and term discovery. MD students manually segmented Chinese EHRs at fine and coarse granularity levels and reviewed the term discovery results. The proposed unsupervised method outperformed all the competing algorithms in the word segmentation task. In term discovery, GTS outperformed the best baseline by 17 percentage points (a 47% relative percentage of increment) on F1-score. In the absence of annotated training data, the graph partition technique can effectively use the corpus statistics and even expert knowledge to realize unsupervised word segmentation of EHRs. Multi-granular segmentation can be used to provide potential medical terms of various lengths with high accuracy.

Bayesian Double Feature Allocation for Phenotyping With Electronic Health Records

Electronic health records (EHR) provide opportunities for deeper understanding of human phenotypes—in our case, latent disease—based on statistical modeling. We propose a categorical matrix factorization method to infer latent diseases from EHR data. A latent disease is defined as an unknown biological aberration that causes a set of common symptoms for a group of patients. The proposed approach is based on a novel double feature allocation model which simultaneously allocates features to the rows and the columns of a categorical matrix. Using a Bayesian approach, available prior information on known diseases (e.g., hypertension and diabetes) greatly improves identifiability and interpretability of the latent diseases. We assess the proposed approach by simulation studies including mis-specified models and comparison with sparse latent factor models. In the application to a Chinese EHR dataset, we identify 10 latent diseases, each of which is shared by groups of subjects with specific health traits related to lipid disorder, thrombocytopenia, polycythemia, anemia, bacterial and viral infections, allergy, and malnutrition. The identification of the latent diseases can help healthcare officials better monitor the subjects’ ongoing health conditions and look into potential risk factors and approaches for disease prevention. We cross-check the reported latent diseases with medical literature and find agreement between our discovery and reported findings elsewhere. We provide an R package “dfa” implementing our method and an R shiny web application reporting the findings. Supplementary materials for this article, including a standardized description of the materials available for reproducing the work, are available as an online supplement.

Clinical Named Entity Recognition From Chinese Electronic Health Records via Machine Learning Methods.

BackgroundElectronic health records (EHRs) are important data resources for clinical studies and applications. Physicians or clinicians describe patients’ disorders or treatment procedures in EHRs using free text (unstructured) clinical notes. The narrative information plays an important role in patient treatment and clinical research. However, it is challenging to make machines understand the clinical narratives.ObjectiveThis study aimed to automatically identify Chinese clinical entities from free text in EHRs and make machines semantically understand diagnoses, tests, body parts, symptoms, treatments, and so on.MethodsThe dataset we used for this study is the benchmark dataset with human annotated Chinese EHRs, released by the China Conference on Knowledge Graph and Semantic Computing 2017 clinical named entity recognition challenge task. Overall, 2 machine learning models, the conditional random fields (CRF) method and bidirectional long short-term memory (LSTM)-CRF, were applied to recognize clinical entities from Chinese EHR data. To train the CRF–based model, we selected features such as bag of Chinese characters, part-of-speech tags, character types, and the position of characters. For the bidirectional LSTM-CRF–based model, character embeddings and segmentation information were used as features. In addition, we also employed a dictionary-based approach as the baseline for the purpose of performance evaluation. Precision, recall, and the harmonic average of precision and recall (F1 score) were used to evaluate the performance of the methods.ResultsExperiments on the test set showed that our methods were able to automatically identify types of Chinese clinical entities such as diagnosis, test, symptom, body part, and treatment simultaneously. With regard to overall performance, CRF and bidirectional LSTM-CRF achieved a precision of 0.9203 and 0.9112, recall of 0.8709 and 0.8974, and F1 score of 0.8949 and 0.9043, respectively. The results also indicated that our methods performed well in recognizing each type of clinical entity, in which the “symptom” type achieved the best F1 score of over 0.96. Moreover, as the number of features increased, the F1 score of the CRF model increased from 0.8547 to 0.8949.ConclusionsIn this study, we employed two computational methods to simultaneously identify types of Chinese clinical entities from free text in EHRs. With training, these methods can effectively identify various types of clinical entities (eg, symptom and treatment) with high accuracy. The deep learning model, bidirectional LSTM-CRF, can achieve better performance than the CRF model with little feature engineering. This study contributed to translating human-readable health information into machine-readable information.

Modeling EHR with the openEHR approach: an exploratory study in China

BackgroundThe openEHR approach can improve the interoperability of electronic health record (EHR) through two-level modeling. Developing archetypes for the complete EHR dataset is essential for implementing a large-scale interoperable EHR system with the openEHR approach. Although the openEHR approach has been applied in different domains, the feasibility of archetyping a complete EHR dataset in a hospital has not been reported in academic literature, especially in a country where using openEHR is still in its infancy stage, like China. This paper presents a case study of modeling an EHR in China aiming to investigate the feasibility and challenges of archetyping a complete EHR dataset with the openEHR approach.MethodsWe proposed an archetype modeling method including an iterative process of collecting requirements, normalizing data elements, organizing concepts, searching corresponding archetypes, editing archetypes and reviewing archetypes. Two representative EHR systems from Chinese vendors and the existing Chinese EHR standards have been used as resources to identify the requirements of EHR in China, and a case study of modeling EHR in China has been conducted. Based on the models developed in this case study, we have implemented a clinical data repository (CDR) to verify the feasibility of modeling EHR with archetypes.ResultsSixty four archetypes were developed to represent all requirements of a complete EHR dataset. 59 (91%) archetypes could be found in Clinical Knowledge Manager (CKM), of which 35 could be reused directly without change, and 23 required further development including two revisions, two new versions, 18 extensions and one specialization. Meanwhile, 6 (9%) archetypes were newly developed. The legacy data of the EHR system in hospitals could be integrated into the CDR developed with these archetypes successfully.ConclusionsThe existing archetypes in CKM can faithfully represent most of the EHR requirements in China except customizations for local hospital management. This case study verified the feasibility of modeling EHR with the openEHR approach and identified the fact that the challenges such as localization, tool support, and an agile publishing process still exist for a broader application of the openEHR approach.

Artificial Intelligence Based Data Governance for Chinese Electronic Health Record Analysis

Artificial Intelligence Based Data Governance for Chinese Electronic Health Record Analysis

Artificial Intelligence Based Data Governance for Chinese Electronic Health Record Analysis

Artificial Intelligence Based Data Governance for Chinese Electronic Health Record Analysis

Chinese electronic health record analysis for recommending medical treatment solutions with NLP and unsupervised learning

Electronic health record (EHR) analysis has become increasingly important in improving the quality of human healthcare. To leverage the full insights from the big EHRs, it is very important to define some application scenarios for which the relevant data can be extracted for training machine learning models to accomplish the expected goals. In this paper, we develop a system on how to recommend medical treatment solutions for patients living in the countryside and small cities when they happen to have schizophrenia but the doctors in the local hospitals do not have sufficient expertise to deal with such challenges. In the EHRs, we take the patients’ symptom descriptions as documents and then develop NLP and unsupervised machine learning techniques to analyze such documents to find the relevant and effective treatment solutions provided by medical experts. Extensive experimental results with different vector representations for documents show that the binary keyword vector representation works best to find relevant and effective medical treatment plans and solutions from the EHRs for any input symptom description.

Abstract P166: Multimorbidity and Risk of Mortality in China: Results From the Chinese Electronic Health Records Research in Yinzhou (CHERRY) Study

Introduction: The prevalence of multimorbidity is increasing in developed countries, while corresponding research in Chinese population is limited. Methods: A population-based cohort in Yinzhou located in an eastern coastal area of China (961,008 adults ≥18 years in 2009; latest mortality follow-up: Nov 2015; 22,637 deaths; 6.53 million person-years at risk) was assembled by linking epidemiological surveys, electronic records for chronic disease management, health administrative and medical records databases. Multimorbidity was defined as the presence of two or more following disorders: hypertension, diabetes, cardiovascular disease or cancer. History of selected diseases was extracted from the electronic records for chronic disease management. Follow-up on fatal events is achieved through records linkage to the regional system of death surveillance. Results: Overall age- and sex-standardized prevalence of one, two or ≥3 disorders were 16.16% (16.09% - 16.23%), 4.11% (4.07% - 4.15%) and0.36% (0.35% - 0.38%) respectively, whereas 41.73%, 14.41% and 1.67% were observed in people aged 60 years and older. The all-cause mortality rate adjusted to the age of 60 was 4.77 per 1000 person-years. Comparing to people without any selected diseases, the age- and sexadjusted HRs for all-cause mortality were 1.17 (1.14, 1.21) in those with one disorder, 1.78 (1.72, 1.86) in those with two morbidities, and 2.97 (2.76, 3.19) in those with ≥3 morbidities. The HRs in those with multimorbidites were higher in younger people and were heterogeneous among different combinations of multimorbidites. Population-attributable fractions (PAF) for all-cause mortality due to one, two or ≥3 disorders were 2.63%, 3.04% and 0.67% respectively (1.28%, 7.69% and 2.34% in people aged 60 years and older). Conclusions: The CHERRY study can serve as a valuable big data resource for scientific research in China. Multimorbidity is becoming acommon condition in Chinese population, especially in older population, and is associated with high mortality. A complementary strategy is required for population health interventions.