Segment convolutional neural networks (Seg-CNNs) for classifying relations in clinical notes

Northwestern University resource and education development initiatives to advance collaborative artificial intelligence across the learning health system

Introduction

The rapid development of artificial intelligence (AI) in healthcare has exposed the unmet need for growing a multidisciplinary workforce that can collaborate effectively in the learning health systems. Maximizing the synergy among multiple teams is critical for Collaborative AI in Healthcare.

Methods

We have developed a series of data, tools, and educational resources for cultivating the next generation of multidisciplinary workforce for Collaborative AI in Healthcare. We built bulk-natural language processing pipelines to extract structured information from clinical notes and stored them in common data models. We developed multimodal AI/machine learning (ML) tools and tutorials to enrich the toolbox of the multidisciplinary workforce to analyze multimodal healthcare data. We have created a fertile ground to cross-pollinate clinicians and AI scientists and train the next generation of AI health workforce to collaborate effectively.

Results

Our work has democratized access to unstructured health information, AI/ML tools and resources for healthcare, and collaborative education resources. From 2017 to 2022, this has enabled studies in multiple clinical specialties resulting in 68 peer-reviewed publications. In 2022, our cross-discipline efforts converged and institutionalized into the Center for Collaborative AI in Healthcare.

Conclusions

Our Collaborative AI in Healthcare initiatives has created valuable educational and practical resources. They have enabled more clinicians, scientists, and hospital administrators to successfully apply AI methods in their daily research and practice, develop closer collaborations, and advanced the institution-level learning health system.

Optimizing healthcare system by amalgamation of text processing and deep learning: a systematic review

The explosion of clinical textual data has drawn the attention of researchers. Owing to the abundance of clinical data, it is becoming difficult for healthcare professionals to take real-time measures. The tools and methods are lacking when compared to the amount of clinical data generated every day. This review aims to survey the text processing pipeline with deep learning methods such as CNN, RNN, LSTM, and GRU in the healthcare domain and discuss various applications such as clinical concept detection and extraction, medically aware dialogue systems, sentiment analysis of drug reviews shared online, clinical trial matching, and pharmacovigilance. In addition, we highlighted the major challenges in deploying text processing with deep learning to clinical textual data and identified the scope of research in this domain. Furthermore, we have discussed various resources that can be used in the future to optimize the healthcare domain by amalgamating text processing and deep learning.

Disease- and Drug-Related Knowledge Extraction for Health Management from Online Health Communities Based on BERT-BiGRU-ATT

Knowledge extraction from rich text in online health communities can supplement and improve the existing knowledge base, supporting evidence-based medicine and clinical decision making. The extracted time series health management data of users can help users with similar conditions when managing their health. By annotating four relationships, this study constructed a deep learning model, BERT-BiGRU-ATT, to extract disease-medication relationships. A Chinese-pretrained BERT model was used to generate word embeddings for the question-and-answer data from online health communities in China. In addition, the bidirectional gated recurrent unit, combined with an attention mechanism, was employed to capture sequence context features and then to classify text related to diseases and drugs using a softmax classifier and to obtain the time series data provided by users. By using various word embedding training experiments and comparisons with classical models, the superiority of our model in relation to extraction was verified. Based on the knowledge extraction, the evolution of a user's disease progression was analyzed according to the time series data provided by users to further analyze the evolution of the user's disease progression. BERT word embedding, GRU, and attention mechanisms in our research play major roles in knowledge extraction. The knowledge extraction results obtained are expected to supplement and improve the existing knowledge base, assist doctors' diagnosis, and help users with dynamic lifecycle health management, such as user disease treatment management. In future studies, a co-reference resolution can be introduced to further improve the effect of extracting the relationships among diseases, drugs, and drug effects.

Machine Learning in Causal Inference: Application in Pharmacovigilance

Monitoring adverse drug events or pharmacovigilance has been promoted by the World Health Organization to assure the safety of medicines through a timely and reliable information exchange regarding drug safety issues. We aim to discuss the application of machine learning methods as well as causal inference paradigms in pharmacovigilance. We first reviewed data sources for pharmacovigilance. Then, we examined traditional causal inference paradigms, their applications in pharmacovigilance, and how machine learning methods and causal inference paradigms were integrated to enhance the performance of traditional causal inference paradigms. Finally, we summarized issues with currently mainstream correlation-based machine learning models and how the machine learning community has tried to address these issues by incorporating causal inference paradigms. Our literature search revealed that most existing data sources and tasks for pharmacovigilance were not designed for causal inference. Additionally, pharmacovigilance was lagging in adopting machine learning-causal inference integrated models. We highlight several currently trending directions or gaps to integrate causal inference with machine learning in pharmacovigilance research. Finally, our literature search revealed that the adoption of causal paradigms can mitigate known issues with machine learning models. We foresee that the pharmacovigilance domain can benefit from the progress in the machine learning field.

Automatic segmentation of cardiac magnetic resonance images based on multi-input fusion network

PURPOSE

Using computer-assisted means to process a large amount of heart image data in order to speed up the diagnosis efficiency and accuracy of medical doctors has become a research worthy of investigation.

METHOD

Based on the U-Net model, this paper proposes a multi-input fusion network (MIFNet) model based on multi-scale input and feature fusion, which automatically extracts and fuses features of different input scales to realize the detection of Cardiac Magnetic Resonance Images (CMRI). The MIFNet model is trained and verified on the public data set, and then compared with the segmentation models, namely the Fully Convolutional Network (FCN) and DeepLab v1.

RESULTS

MIFNet model segmentation of CMRI significantly improved the segmentation accuracy, and the Dice value reached 97.238%. Compared with FCN and DeepLab v1, the average Hausdorff distance (HD) was reduced by 16.425%. The capacity parameter of FCN is 124.86% of MIFNet, DeepLab v1 is 103.22% of MIFNet.

CONCLUSION

Our proposed MIFNet model reduces the amount of parameters and improves the training speed while ensuring the simultaneous segmentation of overlapping targets. It can help clinicians to more quickly check the patient's CMRI focus area, and thereby improving the efficiency of diagnosis.

Modern Clinical Text Mining: A Guide and Review

Electronic health records (EHRs) are becoming a vital source of data for healthcare quality improvement, research, and operations. However, much of the most valuable information contained in EHRs remains buried in unstructured text. The field of clinical text mining has advanced rapidly in recent years, transitioning from rule-based approaches to machine learning and, more recently, deep learning. With new methods come new challenges, however, especially for those new to the field. This review provides an overview of clinical text mining for those who are encountering it for the first time (e.g., physician researchers, operational analytics teams, machine learning scientists from other domains). While not a comprehensive survey, this review describes the state of the art, with a particular focus on new tasks and methods developed over the past few years. It also identifies key barriers between these remarkable technical advances and the practical realities of implementation in health systems and in industry.

Relation Classification for Bleeding Events From Electronic Health Records Using Deep Learning Systems: An Empirical Study

BACKGROUND

Accurate detection of bleeding events from electronic health records (EHRs) is crucial for identifying and characterizing different common and serious medical problems. To extract such information from EHRs, it is essential to identify the relations between bleeding events and related clinical entities (eg, bleeding anatomic sites and lab tests). With the advent of natural language processing (NLP) and deep learning (DL)-based techniques, many studies have focused on their applicability for various clinical applications. However, no prior work has utilized DL to extract relations between bleeding events and relevant entities.

OBJECTIVE

In this study, we aimed to evaluate multiple DL systems on a novel EHR data set for bleeding event-related relation classification.

METHODS

We first expert annotated a new data set of 1046 deidentified EHR notes for bleeding events and their attributes. On this data set, we evaluated three state-of-the-art DL architectures for the bleeding event relation classification task, namely, convolutional neural network (CNN), attention-guided graph convolutional network (AGGCN), and Bidirectional Encoder Representations from Transformers (BERT). We used three BERT-based models, namely, BERT pretrained on biomedical data (BioBERT), BioBERT pretrained on clinical text (Bio+Clinical BERT), and BioBERT pretrained on EHR notes (EhrBERT).

RESULTS

Our experiments showed that the BERT-based models significantly outperformed the CNN and AGGCN models. Specifically, BioBERT achieved a macro F1 score of 0.842, outperforming both the AGGCN (macro F1 score, 0.828) and CNN models (macro F1 score, 0.763) by 1.4% (P<.001) and 7.9% (P<.001), respectively.

CONCLUSIONS

In this comprehensive study, we explored and compared different DL systems to classify relations between bleeding events and other medical concepts. On our corpus, BERT-based models outperformed other DL models for identifying the relations of bleeding-related entities. In addition to pretrained contextualized word representation, BERT-based models benefited from the use of target entity representation over traditional sequence representation.

Constructing High-Fidelity Phenotype Knowledge Graphs for Infectious Diseases With a Fine-Grained Semantic Information Model: Development and Usability Study

BACKGROUND

Phenotypes characterize the clinical manifestations of diseases and provide important information for diagnosis. Therefore, the construction of phenotype knowledge graphs for diseases is valuable to the development of artificial intelligence in medicine. However, phenotype knowledge graphs in current knowledge bases such as WikiData and DBpedia are coarse-grained knowledge graphs because they only consider the core concepts of phenotypes while neglecting the details (attributes) associated with these phenotypes.

OBJECTIVE

To characterize the details of disease phenotypes for clinical guidelines, we proposed a fine-grained semantic information model named PhenoSSU (semantic structured unit of phenotypes).

METHODS

PhenoSSU is an "entity-attribute-value" model by its very nature, and it aims to capture the full semantic information underlying phenotype descriptions with a series of attributes and values. A total of 193 clinical guidelines for infectious diseases from Wikipedia were selected as the study corpus, and 12 attributes from SNOMED-CT were introduced into the PhenoSSU model based on the co-occurrences of phenotype concepts and attribute values. The expressive power of the PhenoSSU model was evaluated by analyzing whether PhenoSSU instances could capture the full semantics underlying the descriptions of the corresponding phenotypes. To automatically construct fine-grained phenotype knowledge graphs, a hybrid strategy that first recognized phenotype concepts with the MetaMap tool and then predicted the attribute values of phenotypes with machine learning classifiers was developed.

RESULTS

Fine-grained phenotype knowledge graphs of 193 infectious diseases were manually constructed with the BRAT annotation tool. A total of 4020 PhenoSSU instances were annotated in these knowledge graphs, and 3757 of them (89.5%) were found to be able to capture the full semantics underlying the descriptions of the corresponding phenotypes listed in clinical guidelines. By comparison, other information models, such as the clinical element model and the HL7 fast health care interoperability resource model, could only capture the full semantics underlying 48.4% (2034/4020) and 21.8% (914/4020) of the descriptions of phenotypes listed in clinical guidelines, respectively. The hybrid strategy achieved an F1-score of 0.732 for the subtask of phenotype concept recognition and an average weighted accuracy of 0.776 for the subtask of attribute value prediction.

CONCLUSIONS

PhenoSSU is an effective information model for the precise representation of phenotype knowledge for clinical guidelines, and machine learning can be used to improve the efficiency of constructing PhenoSSU-based knowledge graphs. Our work will potentially shift the focus of medical knowledge engineering from a coarse-grained level to a more fine-grained level.

Extracting Drug Names and Associated Attributes From Discharge Summaries: Text Mining Study

Background

Drug prescriptions are often recorded in free-text clinical narratives; making this information available in a structured form is important to support many health-related tasks. Although several natural language processing (NLP) methods have been proposed to extract such information, many challenges remain.

Objective

This study evaluates the feasibility of using NLP and deep learning approaches for extracting and linking drug names and associated attributes identified in clinical free-text notes and presents an extensive error analysis of different methods. This study initiated with the participation in the 2018 National NLP Clinical Challenges (n2c2) shared task on adverse drug events and medication extraction.

Methods

The proposed system (DrugEx) consists of a named entity recognizer (NER) to identify drugs and associated attributes and a relation extraction (RE) method to identify the relations between them. For NER, we explored deep learning-based approaches (ie, bidirectional long-short term memory with conditional random fields [BiLSTM-CRFs]) with various embeddings (ie, word embedding, character embedding [CE], and semantic-feature embedding) to investigate how different embeddings influence the performance. A rule-based method was implemented for RE and compared with a context-aware long-short term memory (LSTM) model. The methods were trained and evaluated using the 2018 n2c2 shared task data.

Results

The experiments showed that the best model (BiLSTM-CRFs with pretrained word embeddings [PWE] and CE) achieved lenient micro F-scores of 0.921 for NER, 0.927 for RE, and 0.855 for the end-to-end system. NER, which relies on the pretrained word and semantic embeddings, performed better on most individual entity types, but NER with PWE and CE had the highest classification efficiency among the proposed approaches. Extracting relations using the rule-based method achieved higher accuracy than the context-aware LSTM for most relations. Interestingly, the LSTM model performed notably better in the reason-drug relations, the most challenging relation type.

Conclusions

The proposed end-to-end system achieved encouraging results and demonstrated the feasibility of using deep learning methods to extract medication information from free-text data.

The impact of learning Unified Medical Language System knowledge embeddings in relation extraction from biomedical texts

OBJECTIVE

We explored how knowledge embeddings (KEs) learned from the Unified Medical Language System (UMLS) Metathesaurus impact the quality of relation extraction on 2 diverse sets of biomedical texts.

MATERIALS AND METHODS

Two forms of KEs were learned for concepts and relation types from the UMLS Metathesaurus, namely lexicalized knowledge embeddings (LKEs) and unlexicalized KEs. A knowledge embedding encoder (KEE) enabled learning either LKEs or unlexicalized KEs as well as neural models capable of producing LKEs for mentions of biomedical concepts in texts and relation types that are not encoded in the UMLS Metathesaurus. This allowed us to design the relation extraction with knowledge embeddings (REKE) system, which incorporates either LKEs or unlexicalized KEs produced for relation types of interest and their arguments.

RESULTS

The incorporation of either LKEs or unlexicalized KE in REKE advances the state of the art in relation extraction on 2 relation extraction datasets: the 2010 i2b2/VA dataset and the 2013 Drug-Drug Interaction Extraction Challenge corpus. Moreover, the impact of LKEs is superior, achieving F1 scores of 78.2 and 82.0, respectively.

DISCUSSION

REKE not only highlights the importance of incorporating knowledge encoded in the UMLS Metathesaurus in a novel way, through 2 possible forms of KEs, but it also showcases the subtleties of incorporating KEs in relation extraction systems.

CONCLUSIONS

Incorporating LKEs informed by the UMLS Metathesaurus in a relation extraction system operating on biomedical texts shows significant promise. We present the REKE system, which establishes new state-of-the-art results for relation extraction on 2 datasets when using LKEs.

Using UMLS for electronic health data standardization and database design

OBJECTIVE

Patients that undergo medical transfer represent 1 patient population that remains infrequently studied due to challenges in aggregating data across multiple domains and sources that are necessary to capture the entire episode of patient care. To facilitate access to and secondary use of transport patient data, we developed the Transport Data Repository that combines data from 3 separate domains and many sources within our health system.

METHODS

The repository is a relational database anchored by the Unified Medical Language System unique concept identifiers to integrate, map, and standardize the data into a common data model. Primary data domains included sending and receiving hospital encounters, medical transport record, and custom hospital transport log data. A 4-step mapping process was developed: 1) automatic source code match, 2) exact text match, 3) fuzzy matching, and 4) manual matching.

RESULTS

431 090 total mappings were generated in the Transport Data Repository, consisting of 69 010 unique concepts with 77% of the data being mapped automatically. Transport Source Data yielded significantly lower mapping results with only 8% of data entities automatically mapped and a significant amount (43%) remaining unmapped.

DISCUSSION

The multistep mapping process resulted in a majority of data been automatically mapped. Poor matching of transport medical record data is due to the third-party vendor data being generated and stored in a nonstandardized format.

CONCLUSION

The multistep mapping process developed and implemented is necessary to normalize electronic health data from multiple domains and sources into a common data model to support secondary use of data.

Extracting entities with attributes in clinical text via joint deep learning

OBJECTIVE

Extracting clinical entities and their attributes is a fundamental task of natural language processing (NLP) in the medical domain. This task is typically recognized as 2 sequential subtasks in a pipeline, clinical entity or attribute recognition followed by entity-attribute relation extraction. One problem of pipeline methods is that errors from entity recognition are unavoidably passed to relation extraction. We propose a novel joint deep learning method to recognize clinical entities or attributes and extract entity-attribute relations simultaneously.

MATERIALS AND METHODS

The proposed method integrates 2 state-of-the-art methods for named entity recognition and relation extraction, namely bidirectional long short-term memory with conditional random field and bidirectional long short-term memory, into a unified framework. In this method, relation constraints between clinical entities and attributes and weights of the 2 subtasks are also considered simultaneously. We compare the method with other related methods (ie, pipeline methods and other joint deep learning methods) on an existing English corpus from SemEval-2015 and a newly developed Chinese corpus.

RESULTS

Our proposed method achieves the best F1 of 74.46% on entity recognition and the best F1 of 50.21% on relation extraction on the English corpus, and 89.32% and 88.13% on the Chinese corpora, respectively, which outperform the other methods on both tasks.

CONCLUSIONS

The joint deep learning-based method could improve both entity recognition and relation extraction from clinical text in both English and Chinese, indicating that the approach is promising.

Predicting mortality in critically ill patients with diabetes using machine learning and clinical notes

BACKGROUND

Diabetes mellitus is a prevalent metabolic disease characterized by chronic hyperglycemia. The avalanche of healthcare data is accelerating precision and personalized medicine. Artificial intelligence and algorithm-based approaches are becoming more and more vital to support clinical decision-making. These methods are able to augment health care providers by taking away some of their routine work and enabling them to focus on critical issues. However, few studies have used predictive modeling to uncover associations between comorbidities in ICU patients and diabetes. This study aimed to use Unified Medical Language System (UMLS) resources, involving machine learning and natural language processing (NLP) approaches to predict the risk of mortality.

METHODS

We conducted a secondary analysis of Medical Information Mart for Intensive Care III (MIMIC-III) data. Different machine learning modeling and NLP approaches were applied. Domain knowledge in health care is built on the dictionaries created by experts who defined the clinical terminologies such as medications or clinical symptoms. This knowledge is valuable to identify information from text notes that assert a certain disease. Knowledge-guided models can automatically extract knowledge from clinical notes or biomedical literature that contains conceptual entities and relationships among these various concepts. Mortality classification was based on the combination of knowledge-guided features and rules. UMLS entity embedding and convolutional neural network (CNN) with word embeddings were applied. Concept Unique Identifiers (CUIs) with entity embeddings were utilized to build clinical text representations.

RESULTS

The best configuration of the employed machine learning models yielded a competitive AUC of 0.97. Machine learning models along with NLP of clinical notes are promising to assist health care providers to predict the risk of mortality of critically ill patients.

CONCLUSION

UMLS resources and clinical notes are powerful and important tools to predict mortality in diabetic patients in the critical care setting. The knowledge-guided CNN model is effective (AUC = 0.97) for learning hidden features.

Annotating social determinants of health using active learning, and characterizing determinants using neural event extraction

Social determinants of health (SDOH) affect health outcomes, and knowledge of SDOH can inform clinical decision-making. Automatically extracting SDOH information from clinical text requires data-driven information extraction models trained on annotated corpora that are heterogeneous and frequently include critical SDOH. This work presents a new corpus with SDOH annotations, a novel active learning framework, and the first extraction results on the new corpus. The Social History Annotation Corpus (SHAC) includes 4480 social history sections with detailed annotation for 12 SDOH characterizing the status, extent, and temporal information of 18K distinct events. We introduce a novel active learning framework that selects samples for annotation using a surrogate text classification task as a proxy for a more complex event extraction task. The active learning framework successfully increases the frequency of health risk factors and improves automatic extraction of these events over undirected annotation. An event extraction model trained on SHAC achieves high extraction performance for substance use status (0.82-0.93 F1), employment status (0.81-0.86 F1), and living status type (0.81-0.93 F1) on data from three institutions.

Comparison of deep learning models for natural language processing-based classification of non-English head CT reports

PURPOSE

Natural language processing (NLP) can be used for automatic flagging of radiology reports. We assessed deep learning models for classifying non-English head CT reports.

METHODS

We retrospectively collected head CT reports (2011-2018). Reports were signed in Hebrew. Emergency department (ED) reports of adult patients from January to February for each year (2013-2018) were manually labeled. All other reports were used to pre-train an embedding layer. We explored two use cases: (1) general labeling use case, in which reports were labeled as normal vs. pathological; (2) specific labeling use case, in which reports were labeled as with and without intra-cranial hemorrhage. We tested long short-term memory (LSTM) and LSTM-attention (LSTM-ATN) networks for classifying reports. We also evaluated the improvement of adding Word2Vec word embedding. Deep learning models were compared with a bag-of-words (BOW) model.

RESULTS

We retrieved 176,988 head CT reports for pre-training. We manually labeled 7784 reports as normal (46.3%) or pathological (53.7%), and 7.1% with intra-cranial hemorrhage. For the general labeling, LSTM-ATN-Word2Vec showed the best results (AUC = 0.967 ± 0.006, accuracy 90.8% ± 0.01). For the specific labeling, all methods showed similar accuracies between 95.0 and 95.9%. Both LSTM-ATN-Word2Vec and BOW had the highest AUC (0.970).

CONCLUSION

For a general use case, word embedding using a large cohort of non-English head CT reports and ATN improves NLP performance. For a more specific task, BOW and deep learning showed similar results. Models should be explored and tailored to the NLP task.

Clinical Text Data in Machine Learning: Systematic Review

Background

Clinical narratives represent the main form of communication within health care, providing a personalized account of patient history and assessments, and offering rich information for clinical decision making. Natural language processing (NLP) has repeatedly demonstrated its feasibility to unlock evidence buried in clinical narratives. Machine learning can facilitate rapid development of NLP tools by leveraging large amounts of text data.

Objective

The main aim of this study was to provide systematic evidence on the properties of text data used to train machine learning approaches to clinical NLP. We also investigated the types of NLP tasks that have been supported by machine learning and how they can be applied in clinical practice.

Methods

Our methodology was based on the guidelines for performing systematic reviews. In August 2018, we used PubMed, a multifaceted interface, to perform a literature search against MEDLINE. We identified 110 relevant studies and extracted information about text data used to support machine learning, NLP tasks supported, and their clinical applications. The data properties considered included their size, provenance, collection methods, annotation, and any relevant statistics.

Results

The majority of datasets used to train machine learning models included only hundreds or thousands of documents. Only 10 studies used tens of thousands of documents, with a handful of studies utilizing more. Relatively small datasets were utilized for training even when much larger datasets were available. The main reason for such poor data utilization is the annotation bottleneck faced by supervised machine learning algorithms. Active learning was explored to iteratively sample a subset of data for manual annotation as a strategy for minimizing the annotation effort while maximizing the predictive performance of the model. Supervised learning was successfully used where clinical codes integrated with free-text notes into electronic health records were utilized as class labels. Similarly, distant supervision was used to utilize an existing knowledge base to automatically annotate raw text. Where manual annotation was unavoidable, crowdsourcing was explored, but it remains unsuitable because of the sensitive nature of data considered. Besides the small volume, training data were typically sourced from a small number of institutions, thus offering no hard evidence about the transferability of machine learning models. The majority of studies focused on text classification. Most commonly, the classification results were used to support phenotyping, prognosis, care improvement, resource management, and surveillance.

Conclusions

We identified the data annotation bottleneck as one of the key obstacles to machine learning approaches in clinical NLP. Active learning and distant supervision were explored as a way of saving the annotation efforts. Future research in this field would benefit from alternatives such as data augmentation and transfer learning, or unsupervised learning, which do not require data annotation.

Relation Extraction from Clinical Narratives Using Pre-trained Language Models

Natural language processing (NLP) is useful for extracting information from clinical narratives, and both traditional machine learning methods and more-recent deep learning methods have been successful in various clinical NLP tasks. These methods often depend on traditional word embeddings that are outputs of language models (LMs). Recently, methods that are directly based on pre-trained language models themselves, followed by fine-tuning on the LMs (e.g. the Bidirectional Encoder Representations from Transformers (BERT)), have achieved state-of-the-art performance on many NLP tasks. Despite their success in the open domain and biomedical literature, these pre-trained LMs have not yet been applied to the clinical relation extraction (RE) task. In this study, we developed two different implementations of the BERT model for clinical RE tasks. Our results show that our tuned LMs outperformed previous state-of-the-art RE systems in two shared tasks, which demonstrates the potential of LM-based methods on the RE task.

Classifying relations in clinical narratives using segment graph convolutional and recurrent neural networks (Seg-GCRNs)

We propose to use segment graph convolutional and recurrent neural networks (Seg-GCRNs), which use only word embedding and sentence syntactic dependencies, to classify relations from clinical notes without manual feature engineering. In this study, the relations between 2 medical concepts are classified by simultaneously learning representations of text segments in the context of sentence syntactic dependency: preceding, concept1, middle, concept2, and succeeding segments. Seg-GCRN was systematically evaluated on the i2b2/VA relation classification challenge datasets. Experiments show that Seg-GCRN attains state-of-the-art micro-averaged F-measure for all 3 relation categories: 0.692 for classifying medical treatment-problem relations, 0.827 for medical test-problem relations, and 0.741 for medical problem-medical problem relations. Comparison with the previous state-of-the-art segment convolutional neural network (Seg-CNN) suggests that adding syntactic dependency information helps refine medical word embedding and improves concept relation classification without manual feature engineering. Seg-GCRN can be trained efficiently for the i2b2/VA dataset on a GPU platform.

Artificial Intelligence in Health in 2018: New Opportunities, Challenges, and Practical Implications

Objective : To summarize significant research contributions to the field of artificial intelligence (AI) in health in 2018.

Methods : Ovid MEDLINE ^® and Web of Science ^® databases were searched to identify original research articles that were published in the English language during 2018 and presented advances in the science of AI applied in health. Queries employed Medical Subject Heading (MeSH ^® ) terms and keywords representing AI methodologies and limited results to health applications. Section editors selected 15 best paper candidates that underwent peer review by internationally renowned domain experts. Final best papers were selected by the editorial board of the 2018 International Medical Informatics Association (IMIA) Yearbook.

Results : Database searches returned 1,480 unique publications. Best papers employed innovative AI techniques that incorporated domain knowledge or explored approaches to support distributed or federated learning. All top-ranked papers incorporated novel approaches to advance the science of AI in health and included rigorous evaluations of their methodologies.

Conclusions : Performance of state-of-the-art AI machine learning algorithms can be enhanced by approaches that employ a multidisciplinary biomedical informatics pipeline to incorporate domain knowledge and can overcome challenges such as sparse, missing, or inconsistent data. Innovative training heuristics and encryption techniques may support distributed learning with preservation of privacy.

AI in Health: State of the Art, Challenges, and Future Directions

INTRODUCTION

Artificial intelligence (AI) technologies continue to attract interest from a broad range of disciplines in recent years, including health. The increase in computer hardware and software applications in medicine, as well as digitization of health-related data together fuel progress in the development and use of AI in medicine. This progress provides new opportunities and challenges, as well as directions for the future of AI in health.

OBJECTIVE

The goals of this survey are to review the current state of AI in health, along with opportunities, challenges, and practical implications. This review highlights recent developments over the past five years and directions for the future.

METHODS

Publications over the past five years reporting the use of AI in health in clinical and biomedical informatics journals, as well as computer science conferences, were selected according to Google Scholar citations. Publications were then categorized into five different classes, according to the type of data analyzed.

RESULTS

The major data types identified were multi-omics, clinical, behavioral, environmental and pharmaceutical research and development (R&D) data. The current state of AI related to each data type is described, followed by associated challenges and practical implications that have emerged over the last several years. Opportunities and future directions based on these advances are discussed.

CONCLUSION

Technologies have enabled the development of AI-assisted approaches to healthcare. However, there remain challenges. Work is currently underway to address multi-modal data integration, balancing quantitative algorithm performance and qualitative model interpretability, protection of model security, federated learning, and model bias.

An investigation of single-domain and multidomain medication and adverse drug event relation extraction from electronic health record notes using advanced deep learning models

OBJECTIVE

We aim to evaluate the effectiveness of advanced deep learning models (eg, capsule network [CapNet], adversarial training [ADV]) for single-domain and multidomain relation extraction from electronic health record (EHR) notes.

MATERIALS AND METHODS

We built multiple deep learning models with increased complexity, namely a multilayer perceptron (MLP) model and a CapNet model for single-domain relation extraction and fully shared (FS), shared-private (SP), and adversarial training (ADV) modes for multidomain relation extraction. Our models were evaluated in 2 ways: first, we compared our models using our expert-annotated cancer (the MADE1.0 corpus) and cardio corpora; second, we compared our models with the systems in the MADE1.0 and i2b2 challenges.

RESULTS

Multidomain models outperform single-domain models by 0.7%-1.4% in F1 (t test P < .05), but the results of FS, SP, and ADV modes are mixed. Our results show that the MLP model generally outperforms the CapNet model by 0.1%-1.0% in F1. In the comparisons with other systems, the CapNet model achieves the state-of-the-art result (87.2% in F1) in the cancer corpus and the MLP model generally outperforms MedEx in the cancer, cardiovascular diseases, and i2b2 corpora.

CONCLUSIONS

Our MLP or CapNet model generally outperforms other state-of-the-art systems in medication and adverse drug event relation extraction. Multidomain models perform better than single-domain models. However, neither the SP nor the ADV mode can always outperform the FS mode significantly. Moreover, the CapNet model is not superior to the MLP model for our corpora.

Distant supervision for treatment relation extraction by leveraging MeSH subheadings

The growing body of knowledge in biomedicine is too vast for human consumption. Hence there is a need for automated systems able to navigate and distill the emerging wealth of information. One fundamental task to that end is relation extraction, whereby linguistic expressions of semantic relationships between biomedical entities are recognized and extracted. In this study, we propose a novel distant supervision approach for relation extraction of binary treatment relationships such that high quality positive/negative training examples are generated from PubMed abstracts by leveraging associated MeSH subheadings. The quality of generated examples is assessed based on the quality of supervised models they induce; that is, the mean performance of trained models (derived via bootstrapped ensembling) on a gold standard test set is used as a proxy for data quality. We show that our approach is preferable to traditional distant supervision for treatment relations and is closer to human crowd annotations in terms of annotation quality. For treatment relations, our generated training data performs at 81.38%, compared to traditional distant supervision at 64.33% and crowd-sourced annotations at 90.57% on the model-wide PR-AUC metric. We also demonstrate that examples generated using our method can be used to augment crowd-sourced datasets. Augmented models improve over non-augmented models by more than two absolute points on the more established F1 metric. We lastly demonstrate that performance can be further improved by implementing a classification loss that is resistant to label noise.

The Revival of the Notes Field: Leveraging the Unstructured Content in Electronic Health Records

Problem: Clinical practice requires the production of a time- and resource-consuming great amount of notes. They contain relevant information, but their secondary use is almost impossible, due to their unstructured nature. Researchers are trying to address this problems, with traditional and promising novel techniques. Application in real hospital settings seems not to be possible yet, though, both because of relatively small and dirty dataset, and for the lack of language-specific pre-trained models. Aim: Our aim is to demonstrate the potential of the above techniques, but also raise awareness of the still open challenges that the scientific communities of IT and medical practitioners must jointly address to realize the full potential of unstructured content that is daily produced and digitized in hospital settings, both to improve its data quality and leverage the insights from data-driven predictive models. Methods: To this extent, we present a narrative literature review of the most recent and relevant contributions to leverage the application of Natural Language Processing techniques to the free-text content electronic patient records. In particular, we focused on four selected application domains, namely: data quality, information extraction, sentiment analysis and predictive models, and automated patient cohort selection. Then, we will present a few empirical studies that we undertook at a major teaching hospital specializing in musculoskeletal diseases. Results: We provide the reader with some simple and affordable pipelines, which demonstrate the feasibility of reaching literature performance levels with a single institution non-English dataset. In such a way, we bridged literature and real world needs, performing a step further toward the revival of notes fields.

Identifying Breast Cancer Distant Recurrences from Electronic Health Records Using Machine Learning

Accurately identifying distant recurrences in breast cancer from the Electronic Health Records (EHR) is important for both clinical care and secondary analysis. Although multiple applications have been developed for computational phenotyping in breast cancer, distant recurrence identification still relies heavily on manual chart review. In this study, we aim to develop a model that identifies distant recurrences in breast cancer using clinical narratives and structured data from EHR. We applied MetaMap to extract features from clinical narratives and also retrieved structured clinical data from EHR. Using these features, we trained a support vector machine model to identify distant recurrences in breast cancer patients. We trained the model using 1,396 double-annotated subjects and validated the model using 599 double-annotated subjects. In addition, we validated the model on a set of 4,904 single-annotated subjects as a generalization test. In the held-out test and generalization test, we obtained F-measure scores of 0.78 and 0.74, area under curve (AUC) scores of 0.95 and 0.93, respectively. To explore the representation learning utility of deep neural networks, we designed multiple convolutional neural networks and multilayer neural networks to identify distant recurrences. Using the same test set and generalizability test set, we obtained F-measure scores of 0.79 ± 0.02 and 0.74 ± 0.004, AUC scores of 0.95 ± 0.002 and 0.95 ± 0.01, respectively. Our model can accurately and efficiently identify distant recurrences in breast cancer by combining features extracted from unstructured clinical narratives and structured clinical data.

Clinical text classification with rule-based features and knowledge-guided convolutional neural networks

Background

Clinical text classification is an fundamental problem in medical natural language processing. Existing studies have cocnventionally focused on rules or knowledge sources-based feature engineering, but only a limited number of studies have exploited effective representation learning capability of deep learning methods.

Methods

In this study, we propose a new approach which combines rule-based features and knowledge-guided deep learning models for effective disease classification. Critical Steps of our method include recognizing trigger phrases, predicting classes with very few examples using trigger phrases and training a convolutional neural network (CNN) with word embeddings and Unified Medical Language System (UMLS) entity embeddings.

Results

We evaluated our method on the 2008 Integrating Informatics with Biology and the Bedside (i2b2) obesity challenge. The results demonstrate that our method outperforms the state-of-the-art methods.

Conclusion

We showed that CNN model is powerful for learning effective hidden features, and CUIs embeddings are helpful for building clinical text representations. This shows integrating domain knowledge into CNN models is promising.

Integrating shortest dependency path and sentence sequence into a deep learning framework for relation extraction in clinical text

Background

Extracting relations between important clinical entities is critical but very challenging for natural language processing (NLP) in the medical domain. Researchers have applied deep learning-based approaches to clinical relation extraction; but most of them consider sentence sequence only, without modeling syntactic structures. The aim of this study was to utilize a deep neural network to capture the syntactic features and further improve the performances of relation extraction in clinical notes.

Methods

We propose a novel neural approach to model shortest dependency path (SDP) between target entities together with the sentence sequence for clinical relation extraction. Our neural network architecture consists of three modules: (1) sentence sequence representation module using bidirectional long short-term memory network (Bi-LSTM) to capture the features in the sentence sequence; (2) SDP representation module implementing the convolutional neural network (CNN) and Bi-LSTM network to capture the syntactic context for target entities using SDP information; and (3) classification module utilizing a fully-connected layer with Softmax function to classify the relation type between target entities.

Results

Using the 2010 i2b2/VA relation extraction dataset, we compared our approach with other baseline methods. Our experimental results show that the proposed approach achieved significant improvements over comparable existing methods, demonstrating the effectiveness of utilizing syntactic structures in deep learning-based relation extraction. The F-measure of our method reaches 74.34% which is 2.5% higher than the method without using syntactic features.

Conclusions

We propose a new neural network architecture by modeling SDP along with sentence sequence to extract multi-relations from clinical text. Our experimental results show that the proposed approach significantly improve the performances on clinical notes, demonstrating the effectiveness of syntactic structures in deep learning-based relation extraction.

Natural Language Processing for EHR-Based Computational Phenotyping

This article reviews recent advances in applying natural language processing (NLP) to Electronic Health Records (EHRs) for computational phenotyping. NLP-based computational phenotyping has numerous applications including diagnosis categorization, novel phenotype discovery, clinical trial screening, pharmacogenomics, drug-drug interaction (DDI), and adverse drug event (ADE) detection, as well as genome-wide and phenome-wide association studies. Significant progress has been made in algorithm development and resource construction for computational phenotyping. Among the surveyed methods, well-designed keyword search and rule-based systems often achieve good performance. However, the construction of keyword and rule lists requires significant manual effort, which is difficult to scale. Supervised machine learning models have been favored because they are capable of acquiring both classification patterns and structures from data. Recently, deep learning and unsupervised learning have received growing attention, with the former favored for its performance and the latter for its ability to find novel phenotypes. Integrating heterogeneous data sources have become increasingly important and have shown promise in improving model performance. Often, better performance is achieved by combining multiple modalities of information. Despite these many advances, challenges and opportunities remain for NLP-based computational phenotyping, including better model interpretability and generalizability, and proper characterization of feature relations in clinical narratives.

A general approach for improving deep learning-based medical relation extraction using a pre-trained model and fine-tuning

The automatic extraction of meaningful relations from biomedical literature or clinical records is crucial in various biomedical applications. Most of the current deep learning approaches for medical relation extraction require large-scale training data to prevent overfitting of the training model. We propose using a pre-trained model and a fine-tuning technique to improve these approaches without additional time-consuming human labeling. Firstly, we show the architecture of Bidirectional Encoder Representations from Transformers (BERT), an approach for pre-training a model on large-scale unstructured text. We then combine BERT with a one-dimensional convolutional neural network (1d-CNN) to fine-tune the pre-trained model for relation extraction. Extensive experiments on three datasets, namely the BioCreative V chemical disease relation corpus, traditional Chinese medicine literature corpus and i2b2 2012 temporal relation challenge corpus, show that the proposed approach achieves state-of-the-art results (giving a relative improvement of 22.2, 7.77, and 38.5% in F1 score, respectively, compared with a traditional 1d-CNN classifier). The source code is available at https://github.com/chentao1999/MedicalRelationExtraction.

A Frame-Based NLP System for Cancer-Related Information Extraction

We propose a frame-based natural language processing (NLP) method that extracts cancer-related information from clinical narratives. We focus on three frames: cancer diagnosis, cancer therapeutic procedure, and tumor description. We utilize a deep learning-based approach, bidirectional Long Short-term Memory (LSTM) Conditional Random Field (CRF), which uses both character and word embeddings. The system consists of two constituent sequence classifiers: a frame identification (lexical unit) classifier and a frame element classifier. The classifier achieves an F₁ of 93.70 for cancer diagnosis, 96.33 for therapeutic procedure, and 87.18 for tumor description. These represent improvements of 10.72, 0.85, and 8.04 over a baseline heuristic, respectively. Additionally, we demonstrate that the combination of both GloVe and MIMIC-III embeddings has the best representational effect. Overall, this study demonstrates the effectiveness of deep learning methods to extract frame semantic information from clinical narratives.

Deep Generative Classifiers for Thoracic Disease Diagnosis with Chest X-ray Images

Thoracic diseases are very serious health problems that plague a large number of people. Chest X-ray is currently one of the most popular methods to diagnose thoracic diseases, playing an important role in the healthcare workflow. However, reading the chest X-ray images and giving an accurate diagnosis remain challenging tasks for expert radiologists. With the success of deep learning in computer vision, a growing number of deep neural network architectures were applied to chest X-ray image classification. However, most of the previous deep neural network classifiers were based on deterministic architectures which are usually very noise-sensitive and are likely to aggravate the overfitting issue. In this paper, to make a deep architecture more robust to noise and to reduce overfitting, we propose using deep generative classifiers to automatically diagnose thorax diseases from the chest X-ray images. Unlike the traditional deterministic classifier, a deep generative classifier has a distribution middle layer in the deep neural network. A sampling layer then draws a random sample from the distribution layer and input it to the following layer for classification. The classifier is generative because the class label is generated from samples of a related distribution. Through training the model with a certain amount of randomness, the deep generative classifiers are expected to be robust to noise and can reduce overfitting and then achieve good performances. We implemented our deep generative classifiers based on a number of well-known deterministic neural network architectures, and tested our models on the chest X-ray14 dataset. The results demonstrated the superiority of deep generative classifiers compared with the corresponding deep deterministic classifiers.

Implementing a Portable Clinical NLP System with a Common Data Model - a Lisp Perspective

This paper presents a Lisp architecture for a portable NLP system, termed LAPNLP, for processing clinical notes. LAPNLP integrates multiple standard, customized and in-house developed NLP tools. Our system facilitates portability across different institutions and data systems by incorporating an enriched Common Data Model (CDM) to standardize necessary data elements. It utilizes UMLS to perform domain adaptation when integrating generic domain NLP tools. It also features stand-off annotations that are specified by positional reference to the original document. We built an interval tree based search engine to efficiently query and retrieve the stand-off annotations by specifying positional requirements. We also developed a utility to convert an inline annotation format to stand-off annotations to enable the reuse of clinical text datasets with in-line annotations. We experimented with our system on several NLP facilitated tasks including computational phenotyping for lymphoma patients and semantic relation extraction for clinical notes. These experiments showcased the broader applicability and utility of LAPNLP.

Extraction of Information Related to Adverse Drug Events from Electronic Health Record Notes: Design of an End-to-End Model Based on Deep Learning

BACKGROUND

Pharmacovigilance and drug-safety surveillance are crucial for monitoring adverse drug events (ADEs), but the main ADE-reporting systems such as Food and Drug Administration Adverse Event Reporting System face challenges such as underreporting. Therefore, as complementary surveillance, data on ADEs are extracted from electronic health record (EHR) notes via natural language processing (NLP). As NLP develops, many up-to-date machine-learning techniques are introduced in this field, such as deep learning and multi-task learning (MTL). However, only a few studies have focused on employing such techniques to extract ADEs.

OBJECTIVE

We aimed to design a deep learning model for extracting ADEs and related information such as medications and indications. Since extraction of ADE-related information includes two steps-named entity recognition and relation extraction-our second objective was to improve the deep learning model using multi-task learning between the two steps.

METHODS

We employed the dataset from the Medication, Indication and Adverse Drug Events (MADE) 1.0 challenge to train and test our models. This dataset consists of 1089 EHR notes of cancer patients and includes 9 entity types such as Medication, Indication, and ADE and 7 types of relations between these entities. To extract information from the dataset, we proposed a deep-learning model that uses a bidirectional long short-term memory (BiLSTM) conditional random field network to recognize entities and a BiLSTM-Attention network to extract relations. To further improve the deep-learning model, we employed three typical MTL methods, namely, hard parameter sharing, parameter regularization, and task relation learning, to build three MTL models, called HardMTL, RegMTL, and LearnMTL, respectively.

RESULTS

Since extraction of ADE-related information is a two-step task, the result of the second step (ie, relation extraction) was used to compare all models. We used microaveraged precision, recall, and F1 as evaluation metrics. Our deep learning model achieved state-of-the-art results (F1=65.9%), which is significantly higher than that (F1=61.7%) of the best system in the MADE1.0 challenge. HardMTL further improved the F1 by 0.8%, boosting the F1 to 66.7%, whereas RegMTL and LearnMTL failed to boost the performance.

CONCLUSIONS

Deep learning models can significantly improve the performance of ADE-related information extraction. MTL may be effective for named entity recognition and relation extraction, but it depends on the methods, data, and other factors. Our results can facilitate research on ADE detection, NLP, and machine learning.

Opportunities and challenges in developing deep learning models using electronic health records data: a systematic review

Objective

To conduct a systematic review of deep learning models for electronic health record (EHR) data, and illustrate various deep learning architectures for analyzing different data sources and their target applications. We also highlight ongoing research and identify open challenges in building deep learning models of EHRs.

Design/method

We searched PubMed and Google Scholar for papers on deep learning studies using EHR data published between January 1, 2010, and January 31, 2018. We summarize them according to these axes: types of analytics tasks, types of deep learning model architectures, special challenges arising from health data and tasks and their potential solutions, as well as evaluation strategies.

Results

We surveyed and analyzed multiple aspects of the 98 articles we found and identified the following analytics tasks: disease detection/classification, sequential prediction of clinical events, concept embedding, data augmentation, and EHR data privacy. We then studied how deep architectures were applied to these tasks. We also discussed some special challenges arising from modeling EHR data and reviewed a few popular approaches. Finally, we summarized how performance evaluations were conducted for each task.

Discussion

Despite the early success in using deep learning for health analytics applications, there still exist a number of issues to be addressed. We discuss them in detail including data and label availability, the interpretability and transparency of the model, and ease of deployment.

Using machine learning to identify health outcomes from electronic health record data

Purpose of review

Electronic health records (EHRs) contain valuable data for identifying health outcomes, but these data also present numerous challenges when creating computable phenotyping algorithms. Machine learning methods could help with some of these challenges. In this review, we discuss four common scenarios that researchers may find helpful for thinking critically about when and for what tasks machine learning may be used to identify health outcomes from EHR data.

Recent findings

We first consider the conditions in which machine learning may be especially useful with respect to two dimensions of a health outcome: 1) the characteristics of its diagnostic criteria, and 2) the format in which its diagnostic data are usually stored within EHR systems. In the first dimension, we propose that for health outcomes with diagnostic criteria involving many clinical factors, vague definitions, or subjective interpretations, machine learning may be useful for modeling the complex diagnostic decision-making process from a vector of clinical inputs to identify individuals with the health outcome. In the second dimension, we propose that for health outcomes where diagnostic information is largely stored in unstructured formats such as free text or images, machine learning may be useful for extracting and structuring this information as part of a natural language processing system or an image recognition task. We then consider these two dimensions jointly to define four common scenarios of health outcomes. For each scenario, we discuss the potential uses for machine learning - first assuming accurate and complete EHR data and then relaxing these assumptions to accommodate the limitations of real-world EHR systems. We illustrate these four scenarios using concrete examples and describe how recent studies have used machine learning to identify these health outcomes from EHR data.

Summary

Machine learning has great potential to improve the accuracy and efficiency of health outcome identification from EHR systems, especially under certain conditions. To promote the use of machine learning in EHR-based phenotyping tasks, future work should prioritize efforts to increase the transportability of machine learning algorithms for use in multi-site settings.

Computer-Aided Multi-Target Management of Emergent Alzheimer's Disease

Alzheimer's disease (AD) represents an enormous global health burden in terms of human suffering and economic cost. AD management requires a shift from the prevailing paradigm targeting pathogenesis to design and develop effective drugs with adequate success in clinical trials. Therefore, it is of interest to report a review on amyloid beta (Aβ) effects and other multi-targets including cholinesterase, NFTs, tau protein and TNF associated with brain cell death to be neuro-protective from AD. It should be noted that these molecules have been generated either by target-based or phenotypic methods. Hence, the use of recent advancements in nanomedicine and other natural compounds screening tools as a feasible alternative for circumventing specific liabilities is realized. We review recent developments in the design and identification of neuro-degenerative compounds against AD generated using current advancements in computational multi-target modeling algorithms reflected by theragnosis (combination of diagnostic tests and therapy) concern.

An end-to-end deep learning architecture for extracting protein-protein interactions affected by genetic mutations

The BioCreative VI Track IV (mining protein interactions and mutations for precision medicine) challenge was organized in 2017 with the goal of applying biomedical text mining methods to support advancements in precision medicine approaches. As part of the challenge, a new dataset was introduced for the purpose of building a supervised relation extraction model capable of taking a test article and returning a list of interacting protein pairs identified by their Entrez Gene IDs. Specifically, such pairs represent proteins participating in a binary protein–protein interaction relation where the interaction is additionally affected by a genetic mutation—referred to as a PPIm relation. In this study, we explore an end-to-end approach for PPIm relation extraction by deploying a three-component pipeline involving deep learning-based named-entity recognition and relation classification models along with a knowledge-based approach for gene normalization. We propose several recall-focused improvements to our original challenge entry that placed second when matching on Entrez Gene ID (exact matching) and on HomoloGene ID. On exact matching, the improved system achieved new competitive test results of 37.78% micro-F1 with a precision of 38.22% and recall of 37.34% that corresponds to an improvement from the prior best system by approximately three micro-F1 points. When matching on HomoloGene IDs, we report similarly competitive test results at 46.17% micro-F1 with a precision and recall of 46.67 and 45.59%, respectively, corresponding to an improvement of more than eight micro-F1 points over the prior best result. The code for our deep learning system is made publicly available at https://github.com/bionlproc/biocppi_extraction.