Clinical Information Extraction Research Articles

Potential Roles of Large Language Models in the Production of Systematic Reviews and Meta-Analyses.

Large language models (LLMs) such as ChatGPT have become widely applied in the field of medical research. In the process of conducting systematic reviews, similar tools can be used to expedite various steps, including defining clinical questions, performing the literature search, document screening, information extraction, and language refinement, thereby conserving resources and enhancing efficiency. However, when using LLMs, attention should be paid to transparent reporting, distinguishing between genuine and false content, and avoiding academic misconduct. In this viewpoint, we highlight the potential roles of LLMs in the creation of systematic reviews and meta-analyses, elucidating their advantages, limitations, and future research directions, aiming to provide insights and guidance for authors planning systematic reviews and meta-analyses.

An Empirical Evaluation of Prompting Strategies for Large Language Models in Zero-Shot Clinical Natural Language Processing: Algorithm Development and Validation Study.

Large language models (LLMs) have shown remarkable capabilities in natural language processing (NLP), especially in domains where labeled data are scarce or expensive, such as the clinical domain. However, to unlock the clinical knowledge hidden in these LLMs, we need to design effective prompts that can guide them to perform specific clinical NLP tasks without any task-specific training data. This is known as in-context learning, which is an art and science that requires understanding the strengths and weaknesses of different LLMs and prompt engineering approaches. The objective of this study is to assess the effectiveness of various prompt engineering techniques, including 2 newly introduced types-heuristic and ensemble prompts, for zero-shot and few-shot clinical information extraction using pretrained language models. This comprehensive experimental study evaluated different prompt types (simple prefix, simple cloze, chain of thought, anticipatory, heuristic, and ensemble) across 5 clinical NLP tasks: clinical sense disambiguation, biomedical evidence extraction, coreference resolution, medication status extraction, and medication attribute extraction. The performance of these prompts was assessed using 3 state-of-the-art language models: GPT-3.5 (OpenAI), Gemini (Google), and LLaMA-2 (Meta). The study contrasted zero-shot with few-shot prompting and explored the effectiveness of ensemble approaches. The study revealed that task-specific prompt tailoring is vital for the high performance of LLMs for zero-shot clinical NLP. In clinical sense disambiguation, GPT-3.5 achieved an accuracy of 0.96 with heuristic prompts and 0.94 in biomedical evidence extraction. Heuristic prompts, alongside chain of thought prompts, were highly effective across tasks. Few-shot prompting improved performance in complex scenarios, and ensemble approaches capitalized on multiple prompt strengths. GPT-3.5 consistently outperformed Gemini and LLaMA-2 across tasks and prompt types. This study provides a rigorous evaluation of prompt engineering methodologies and introduces innovative techniques for clinical information extraction, demonstrating the potential of in-context learning in the clinical domain. These findings offer clear guidelines for future prompt-based clinical NLP research, facilitating engagement by non-NLP experts in clinical NLP advancements. To the best of our knowledge, this is one of the first works on the empirical evaluation of different prompt engineering approaches for clinical NLP in this era of generative artificial intelligence, and we hope that it will inspire and inform future research in this area.

Natural language processing in urology: Automated extraction of clinical information from histopathology reports of uro-oncology procedures

ObjectivesWe aimed to automate routine extraction of clinically relevant unstructured information from uro-oncological histopathology reports by applying rule-based and machine learning (ML)/deep learning (DL) methods to develop an oncology focused natural language processing (NLP) algorithm. MethodsOur algorithm employs a combination of a rule-based approach and support vector machines/neural networks (BioBert/Clinical BERT), and is optimised for accuracy. We randomly extracted 5772 uro-oncological histology reports from 2008 to 2018 from electronic health records (EHRs) and split the data into training and validation datasets in an 80:20 ratio. The training dataset was annotated by medical professionals and reviewed by cancer registrars. The validation dataset was annotated by cancer registrars and defined as the gold standard with which the algorithm outcomes were compared. The accuracy of NLP-parsed data was matched against these human annotation results. We defined an accuracy rate of >95% as “acceptable” by professional human extraction, as per our cancer registry definition. ResultsThere were 11 extraction variables in 268 free-text reports. We achieved an accuracy rate of between 61.2% and 99.0% using our algorithm. Of the 11 data fields, a total of 8 data fields met the acceptable accuracy standard, while another 3 data fields had an accuracy rate between 61.2% and 89.7%. Noticeably, the rule-based approach was shown to be more effective and robust in extracting variables of interest. On the other hand, ML/DL models had poorer predictive performances due to highly imbalanced data distribution and variable writing styles between different reports and data used for domain-specific pre-trained models. ConclusionWe designed an NLP algorithm that can automate clinical information extraction accurately from histopathology reports with an overall average micro accuracy of 93.3%.

Clinical concept recognition: Evaluation of existing systems on EHRs.

The adoption of electronic health records (EHRs) has produced enormous amounts of data, creating research opportunities in clinical data sciences. Several concept recognition systems have been developed to facilitate clinical information extraction from these data. While studies exist that compare the performance of many concept recognition systems, they are typically developed internally and may be biased due to different internal implementations, parameters used, and limited number of systems included in the evaluations. The goal of this research is to evaluate the performance of existing systems to retrieve relevant clinical concepts from EHRs. We investigated six concept recognition systems, including CLAMP, cTAKES, MetaMap, NCBO Annotator, QuickUMLS, and ScispaCy. Clinical concepts extracted included procedures, disorders, medications, and anatomical location. The system performance was evaluated on two datasets: the 2010 i2b2 and the MIMIC-III. Additionally, we assessed the performance of these systems in five challenging situations, including negation, severity, abbreviation, ambiguity, and misspelling. For clinical concept extraction, CLAMP achieved the best performance on exact and inexact matching, with an F-score of 0.70 and 0.94, respectively, on i2b2; and 0.39 and 0.50, respectively, on MIMIC-III. Across the five challenging situations, ScispaCy excelled in extracting abbreviation information (F-score: 0.86) followed by NCBO Annotator (F-score: 0.79). CLAMP outperformed in extracting severity terms (F-score 0.73) followed by NCBO Annotator (F-score: 0.68). CLAMP outperformed other systems in extracting negated concepts (F-score 0.63). Several concept recognition systems exist to extract clinical information from unstructured data. This study provides an external evaluation by end-users of six commonly used systems across different extraction tasks. Our findings suggest that CLAMP provides the most comprehensive set of annotations for clinical concept extraction tasks and associated challenges. Comparing standard extraction tasks across systems provides guidance to other clinical researchers when selecting a concept recognition system relevant to their clinical information extraction task.

Moving toward a standardized diagnostic statement of pituitary adenoma using an information extraction model: a real-world study based on electronic medical records

PurposeDiagnostic statements for pituitary adenomas (PAs) are complex and unstandardized. We aimed to determine the most commonly used elements contained in the statements and their combination patterns and variations in real-world clinical practice, with the ultimate goal of promoting standardized diagnostic recording and establishing an efficient element extraction process.MethodsPatient medical records from 2012 to 2020 that included PA among the first three diagnoses were included. After manually labeling the elements in the diagnostic texts, we obtained element types and training sets, according to which an information extraction model was constructed based on the word segmentation model “Jieba” to extract information contained in the remaining diagnostic texts.ResultsA total of 576 different diagnostic statements from 4010 texts of 3770 medical records were enrolled in the analysis. The first ten diagnostic elements related to PA were histopathology, tumor location, endocrine status, tumor size, invasiveness, recurrence, diagnostic confirmation, Knosp grade, residual tumor, and refractoriness. The automated extraction model achieved F1-scores that reached 100% for all ten elements in the second round and 97.3–100.0% in the test set consisting of an additional 532 diagnostic texts. Tumor location, endocrine status, histopathology, and tumor size were the most commonly used elements, and diagnoses composed of the above elements were the most frequent. Endocrine status had the greatest expression variability, followed by Knosp grade. Among all the terms, the percentage of loss of tumor size was among the highest (21%). Among statements where the principal diagnoses were PAs, 18.6% did not have information on tumor size, while for those with other diagnoses, this percentage rose to 48% (P < 0.001).ConclusionStandardization of the diagnostic statement for PAs is unsatisfactory in real-world clinical practice. This study could help standardize a structured pattern for PA diagnosis and establish a foundation for research-friendly, high-quality clinical information extraction.

Temporal Events Detector for Pregnancy Care (TED-PC): A rule-based algorithm to infer gestational age and delivery date from electronic health records of pregnant women with and without COVID-19.

Identifying the time of SARS-CoV-2 viral infection relative to specific gestational weeks is critical for delineating the role of viral infection timing in adverse pregnancy outcomes. However, this task is difficult when it comes to Electronic Health Records (EHR). In combating the COVID-19 pandemic for maternal health, we sought to develop and validate a clinical information extraction algorithm to detect the time of clinical events relative to gestational weeks. We used EHR from the National COVID Cohort Collaborative (N3C), in which the EHR are normalized by the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM). We performed EHR phenotyping, resulting in 270,897 pregnant women (June 1st, 2018 to May 31st, 2021). We developed a rule-based algorithm and performed a multi-level evaluation to test content validity and clinical validity, and extreme length of gestation (<150 or >300). The algorithm identified 296,194 pregnancies (16,659 COVID-19, 174,744 without COVID-19) in 270,897 pregnant women. For inferring gestational age, 95% cases (n = 40) have moderate-high accuracy (Cohen's Kappa = 0.62); 100% cases (n = 40) have moderate-high granularity of temporal information (Cohen's Kappa = 1). For inferring delivery dates, the accuracy is 100% (Cohen's Kappa = 1). The accuracy of gestational age detection for the extreme length of gestation is 93.3% (Cohen's Kappa = 1). Mothers with COVID-19 showed higher prevalence in obesity or overweight (35.1% vs. 29.5%), diabetes (17.8% vs. 17.0%), chronic obstructive pulmonary disease (0.2% vs. 0.1%), respiratory distress syndrome or acute respiratory failure (1.8% vs. 0.2%). We explored the characteristics of pregnant women by different gestational weeks of SARS-CoV-2 infection with our algorithm. TED-PC is the first to infer the exact gestational week linked with every clinical event from EHR and detect the timing of SARS-CoV-2 infection in pregnant women. The algorithm shows excellent clinical validity in inferring gestational age and delivery dates, which supports multiple EHR cohorts on N3C studying the impact of COVID-19 on pregnancy.

Using natural language processing to automatically classify written self-reported narratives by patients with migraine or cluster headache

BackgroundHeadache medicine is largely based on detailed history taking by physicians analysing patients’ descriptions of headache. Natural language processing (NLP) structures and processes linguistic data into quantifiable units. In this study, we apply these digital techniques on self-reported narratives by patients with headache disorders to research the potential of analysing and automatically classifying human-generated text and information extraction in clinical contexts.MethodsA prospective cross-sectional clinical trial collected self-reported narratives on headache disorders from participants with either migraine or cluster headache. NLP was applied for the analysis of lexical, semantic and thematic properties of the texts. Machine learning (ML) algorithms were applied to classify the descriptions of headache attacks from individual participants into their correct group (migraine versus cluster headache).ResultsOne-hundred and twenty-one patients (81 participants with migraine and 40 participants with cluster headache) provided a self-reported narrative on their headache disorder. Lexical analysis of this text corpus resulted in several specific key words per diagnostic group (cluster headache: Dutch (nl): “oog” | English (en): “eye”, nl: “pijn” | en: “pain” and nl: “terug” | en: “back/to come back”; migraine: nl: “hoofdpijn” | en: “headache”, nl: “stress” | en: “stress” and nl: “misselijkheid” | en: “nausea”). Thematic and sentiment analysis of text revealed largely negative sentiment in texts by both patients with migraine and cluster headache. Logistic regression and support vector machine algorithms with different feature groups performed best for the classification of attack descriptions (with F1-scores for detecting cluster headache varying between 0.82 and 0.86) compared to naïve Bayes classifiers.ConclusionsDifferences in lexical choices between patients with migraine and cluster headache are detected with NLP and are congruent with domain expert knowledge of the disorders. Our research shows that ML algorithms have potential to classify patients’ self-reported narratives of migraine or cluster headache with good performance. NLP shows its capability to discern relevant linguistic aspects in narratives from patients with different headache disorders and demonstrates relevance in clinical information extraction. The potential benefits on the classification performance of larger datasets and neural NLP methods can be investigated in the future.Trial registrationThis study was registered with clinicaltrials.gov with ID NCT05377437.

Application of Deep Learning in Generating Structured Radiology Reports: A Transformer-Based Technique.

Since radiology reports needed for clinical practice and research are written and stored in free-text narrations, extraction of relative information for further analysis is difficult. In these circumstances, natural language processing (NLP) techniques can facilitate automatic information extraction and transformation of free-text formats to structured data. In recent years, deep learning (DL)-based models have been adapted for NLP experiments with promising results. Despite the significant potential of DL models based on artificial neural networks (ANN) and convolutional neural networks (CNN), the models face some limitations to implement in clinical practice. Transformers, another new DL architecture, have been increasingly applied to improve the process. Therefore, in this study, we propose a transformer-based fine-grained named entity recognition (NER) architecture for clinical information extraction. We collected 88 abdominopelvic sonography reports in free-text formats and annotated them based on our developed information schema. The text-to-text transfer transformer model (T5) and Scifive, a pre-trained domain-specific adaptation of the T5 model, were applied for fine-tuning to extract entities and relations and transform the input into a structured format. Our transformer-based model in this study outperformed previously applied approaches such as ANN and CNN models based on ROUGE-1, ROUGE-2, ROUGE-L, and BLEU scores of 0.816, 0.668, 0.528, and 0.743, respectively, while providing an interpretable structured report.

A Novel Patient Similarity Network (PSN) Framework Based on Multi-Model Deep Learning for Precision Medicine.

Precision medicine can be defined as the comparison of a new patient with existing patients that have similar characteristics and can be referred to as patient similarity. Several deep learning models have been used to build and apply patient similarity networks (PSNs). However, the challenges related to data heterogeneity and dimensionality make it difficult to use a single model to reduce data dimensionality and capture the features of diverse data types. In this paper, we propose a multi-model PSN that considers heterogeneous static and dynamic data. The combination of deep learning models and PSN allows ample clinical evidence and information extraction against which similar patients can be compared. We use the bidirectional encoder representations from transformers (BERT) to analyze the contextual data and generate word embedding, where semantic features are captured using a convolutional neural network (CNN). Dynamic data are analyzed using a long-short-term-memory (LSTM)-based autoencoder, which reduces data dimensionality and preserves the temporal features of the data. We propose a data fusion approach combining temporal and clinical narrative data to estimate patient similarity. The experiments we conducted proved that our model provides a higher classification accuracy in determining various patient health outcomes when compared with other traditional classification algorithms.

Optimal dynamic treatment regime estimation using information extraction from unstructured clinical text.

The wide-scale adoption of electronic health records (EHRs) provides extensive information to support precision medicine and personalized health care. In addition to structured EHRs, we leverage free-text clinical information extraction (IE) techniques to estimate optimal dynamic treatment regimes (DTRs), a sequence of decision rules that dictate how to individualize treatments to patients based on treatment and covariate history. The proposed IE of patient characteristics closely resembles "The clinical Text Analysis and Knowledge Extraction System" and employs named entity recognition, boundary detection, and negation annotation. It also utilizes regular expressions to extract numerical information. Combining the proposed IE with optimal DTR estimation, we extract derived patient characteristics and use tree-based reinforcement learning (T-RL) to estimate multistage optimal DTRs. IE significantly improved the estimation in counterfactual outcome models compared to using structured EHR data alone, which often include incomplete data, data entry errors, and other potentially unobserved risk factors. Moreover, including IE in optimal DTR estimation provides larger study cohorts and a broader pool of candidate tailoring variables. We demonstrate the performance of our proposed method via simulations and an application using clinical records to guide blood pressure control treatments among critically ill patients with severe acute hypertension. This joint estimation approach improves the accuracy of identifying the optimal treatment sequence by 14-24% compared to traditional inference without using IE, based on our simulations over various scenarios. In the blood pressure control application, we successfully extracted significant blood pressure predictors that are unobserved or partially missing from structuredEHR.

Combining data augmentation and domain information with TENER model for Clinical Event Detection

BackgroundIn recent years, with the development of artificial intelligence, the use of deep learning technology for clinical information extraction has become a new trend. Clinical Event Detection (CED) as its subtask has attracted the attention from academia and industry. However, directly applying the advancements in deep learning to CED task often yields unsatisfactory results. The main reasons are due to the following two points: (1) A great number of obscure professional terms in the electronic medical record leads to poor recognition performance of model. (2) The scarcity of datasets required for the task leads to poor model robustness. Therefore, it is urgent to solve these two problems to improve model performance.MethodsThis paper proposes a combining data augmentation and domain information with TENER Model for Clinical Event Detection.ResultsWe use two evaluation metrics to compare the overall performance of the proposed model with the existing model on the 2012 i2b2 challenge dataset. Experimental results demonstrate that our proposed model achieves the best F1-score of 80.26%, type accuracy of 93% and Span F1-score of 90.33%, and outperforms the state-of-the-art approaches.ConclusionsThis paper proposes a multi-granularity information fusion encoder-decoder framework, which applies the TENER model to the CED task for the first time. It uses the pre-trained language model (BioBERT) to generate word-level features, solving the problem of a great number of obscure professional terms in the electronic medical record lead to poor recognition performance of model. In addition, this paper proposes a new data augmentation method for sequence labeling tasks, solving the problem of the scarcity of datasets required for the task leads to poor model robustness.

MT-clinical BERT: scaling clinical information extraction with multitask learning.

Clinical notes contain an abundance of important, but not-readily accessible, information about patients. Systems that automatically extract this information rely on large amounts of training data of which there exists limited resources to create. Furthermore, they are developed disjointly, meaning that no information can be shared among task-specific systems. This bottleneck unnecessarily complicates practical application, reduces the performance capabilities of each individual solution, and associates the engineering debt of managing multiple information extraction systems. We address these challenges by developing Multitask-Clinical BERT: a single deep learning model that simultaneously performs 8 clinical tasks spanning entity extraction, personal health information identification, language entailment, and similarity by sharing representations among tasks. We compare the performance of our multitasking information extraction system to state-of-the-art BERT sequential fine-tuning baselines. We observe a slight but consistent performance degradation in MT-Clinical BERT relative to sequential fine-tuning. These results intuitively suggest that learning a general clinical text representation capable of supporting multiple tasks has the downside of losing the ability to exploit dataset or clinical note-specific properties when compared to a single, task-specific model. We find our single system performs competitively with all state-the-art task-specific systems while also benefiting from massive computational benefits at inference.

Pulse of the pandemic: Iterative topic filtering for clinical information extraction from social media

The rapid evolution of the COVID-19 pandemic has underscored the need to quickly disseminate the latest clinical knowledge during a public-health emergency. One surprisingly effective platform for healthcare professionals (HCPs) to share knowledge and experiences from the front lines has been social media (for example, the “#medtwitter” community on Twitter). However, identifying clinically-relevant content in social media without manual labeling is a challenge because of the sheer volume of irrelevant data. We present an unsupervised, iterative approach to mine clinically relevant information from social media data, which begins by heuristically filtering for HCP-authored texts and incorporates topic modeling and concept extraction with MetaMap. This approach identifies granular topics and tweets with high clinical relevance from a set of about 52 million COVID-19-related tweets from January to mid-June 2020. We also show that because the technique does not require manual labeling, it can be used to identify emerging topics on a week-to-week basis. Our method can aid in future public-health emergencies by facilitating knowledge transfer among healthcare workers in a rapidly-changing information environment, and by providing an efficient and unsupervised way of highlighting potential areas for clinical research.

Chinese Clinical Named Entity Recognition in Electronic Medical Records: Development of a Lattice Long Short-Term Memory Model With Contextualized Character Representations.

BackgroundClinical named entity recognition (CNER), whose goal is to automatically identify clinical entities in electronic medical records (EMRs), is an important research direction of clinical text data mining and information extraction. The promotion of CNER can provide support for clinical decision making and medical knowledge base construction, which could then improve overall medical quality. Compared with English CNER, and due to the complexity of Chinese word segmentation and grammar, Chinese CNER was implemented later and is more challenging.ObjectiveWith the development of distributed representation and deep learning, a series of models have been applied in Chinese CNER. Different from the English version, Chinese CNER is mainly divided into character-based and word-based methods that cannot make comprehensive use of EMR information and cannot solve the problem of ambiguity in word representation.MethodsIn this paper, we propose a lattice long short-term memory (LSTM) model combined with a variant contextualized character representation and a conditional random field (CRF) layer for Chinese CNER: the Embeddings from Language Models (ELMo)-lattice-LSTM-CRF model. The lattice LSTM model can effectively utilize the information from characters and words in Chinese EMRs; in addition, the variant ELMo model uses Chinese characters as input instead of the character-encoding layer of the ELMo model, so as to learn domain-specific contextualized character embeddings.ResultsWe evaluated our method using two Chinese CNER datasets from the China Conference on Knowledge Graph and Semantic Computing (CCKS): the CCKS-2017 CNER dataset and the CCKS-2019 CNER dataset. We obtained F1 scores of 90.13% and 85.02% on the test sets of these two datasets, respectively.ConclusionsOur results show that our proposed method is effective in Chinese CNER. In addition, the results of our experiments show that variant contextualized character representations can significantly improve the performance of the model.

Clinical information extraction for preterm birth risk prediction.

This paper contributes to the pursuit of leveraging unstructured medical notes to structured clinical decision making. In particular, we present a pipeline for clinical information extraction from medical notes related to preterm birth, and discuss the main challenges as well as its potential for clinical practice. A large collection of medical notes, created by staff during hospitalizations of patients who were at risk of delivering preterm, was gathered and analyzed. Based on an annotated collection of notes, we trained and evaluated information extraction components to discover clinical entities such as symptoms, events, anatomical sites and procedures, as well as attributes linked to these clinical entities. In a retrospective study, we show that these are highly informative for clinical decision support models that are trained to predict whether delivery is likely to occur within specific time windows, in combination with structured information from electronic health records.

The Use of Ontology in Clinical Information Extraction

Extracting clinical data from medical or clinical reports is a crucial effort. These records contain the most valuable pieces of evidence of treatments in humans. Integration of information extraction (IE) and ontology can produce a great tool for clinical concept extraction. The aim of this paper is to present a quick overview of the research work which has applied IE and ontology approaches in medical or clinical concepts extraction. This paper also presents our proposed framework based on the integration of both approaches mentioned above for extracting clinical concepts.

Adapting and evaluating a deep learning language model for clinical why-question answering.

ObjectivesTo adapt and evaluate a deep learning language model for answering why-questions based on patient-specific clinical text.Materials and MethodsBidirectional encoder representations from transformers (BERT) models were trained with varying data sources to perform SQuAD 2.0 style why-question answering (why-QA) on clinical notes. The evaluation focused on: (1) comparing the merits from different training data and (2) error analysis.ResultsThe best model achieved an accuracy of 0.707 (or 0.760 by partial match). Training toward customization for the clinical language helped increase 6% in accuracy.DiscussionThe error analysis suggested that the model did not really perform deep reasoning and that clinical why-QA might warrant more sophisticated solutions.ConclusionThe BERT model achieved moderate accuracy in clinical why-QA and should benefit from the rapidly evolving technology. Despite the identified limitations, it could serve as a competent proxy for question-driven clinical information extraction.

Ontology-based clinical information extraction from physician's free-text notes.

Documenting clinical notes in electronic health records might affect physician's workflow. In this paper, an Ontology-based clinical information extraction system, OB-CIE, has been developed. OB-CIE system provides a method for extracting clinical concepts from physician's free-text notes and converts the unstructured clinical notes to structured information to be accessed in electronic health records. OB-CIE system can help physicians to document visit notes without changing their workflow. For recognizing named entities of clinical concepts, ontology concepts have been used to construct a dictionary of semantic categories, then, exact dictionary matching method has been used to match noun phrases to their semantic categories. A rule-based approach has been used to classify clinical sentences to their predefined categories. The system evaluation results have achieved an F-measure of 94.90% and 97.80% for concepts classification and sentences classification, respectively. The results have showed that OB-CIE system performed well on extracting clinical concepts compared with data mining techniques. The system can be used in another field by adapting its ontology and extraction rule set.

Adapting State-of-the-Art Deep Language Models to Clinical Information Extraction Systems: Potentials, Challenges, and Solutions.

BackgroundDeep learning (DL) has been widely used to solve problems with success in speech recognition, visual object recognition, and object detection for drug discovery and genomics. Natural language processing has achieved noticeable progress in artificial intelligence. This gives an opportunity to improve on the accuracy and human-computer interaction of clinical informatics. However, due to difference of vocabularies and context between a clinical environment and generic English, transplanting language models directly from up-to-date methods to real-world health care settings is not always satisfactory. Moreover, the legal restriction on using privacy-sensitive patient records hinders the progress in applying machine learning (ML) to clinical language processing.ObjectiveThe aim of this study was to investigate 2 ways to adapt state-of-the-art language models to extracting patient information from free-form clinical narratives to populate a handover form at a nursing shift change automatically for proofing and revising by hand: first, by using domain-specific word representations and second, by using transfer learning models to adapt knowledge from general to clinical English. We have described the practical problem, composed it as an ML task known as information extraction, proposed methods for solving the task, and evaluated their performance.MethodsFirst, word representations trained from different domains served as the input of a DL system for information extraction. Second, the transfer learning model was applied as a way to adapt the knowledge learned from general text sources to the task domain. The goal was to gain improvements in the extraction performance, especially for the classes that were topically related but did not have a sufficient amount of model solutions available for ML directly from the target domain. A total of 3 independent datasets were generated for this task, and they were used as the training (101 patient reports), validation (100 patient reports), and test (100 patient reports) sets in our experiments.ResultsOur system is now the state-of-the-art in this task. Domain-specific word representations improved the macroaveraged F1 by 3.4%. Transferring the knowledge from general English corpora to the task-specific domain contributed a further 7.1% improvement. The best performance in populating the handover form with 37 headings was the macroaveraged F1 of 41.6% and F1 of 81.1% for filtering out irrelevant information. Performance differences between this system and its baseline were statistically significant (P<.001; Wilcoxon test).ConclusionsTo our knowledge, our study is the first attempt to transfer models from general deep models to specific tasks in health care and gain a significant improvement. As transfer learning shows its advantage over other methods, especially on classes with a limited amount of training data, less experts’ time is needed to annotate data for ML, which may enable good results even in resource-poor domains.

Extraction of Temporal Information from Clinical Narratives

The existence of massive quantity of clinical text in electronic medical records (EMRs) has created significant demand for clinical text processing and information extraction in the field of health care and medical research. Detailed clinical observations of patients are typically recorded chronologically. Temporal information in such clinical texts consist of three elements: temporal expressions, temporal events, and temporal relations. Due to the implicit expression of temporal information, lack of writing quality, and domain-specific nature in the clinical text, extraction of temporal information is much more complex than for newswire texts. In spite of these difficulties, to extract temporal information using the annotated corpora, few research works reported rule-based, machine-learning, and hybrid methods. On the other hand, creating the annotated corpora is expensive, time-consuming, and demands significant human effort; the processing quality is inevitably affected by the small size of corpora. Motivated by this issue, in this research work, we present a novel method to effectively extract the temporal information from EMR clinical texts. The essential idea of this method is first to build a feature set appropriately for clinical expressions, followed by the development of a semi-supervised framework for temporal event extraction, and finally detection of temporal relations among events with a newly formulated hypothesis. Comparative experimental evaluation on the I2B2 data set has clearly shown improved performance of the proposed methods. Specifically, temporal event and relation extraction is possible with an F-measure 89.98 and 67.1% respectively.