A systematic comparison of feature space effects on disease classifier performance for phenotype identification of five diseases

Clinical Decision Support and Natural Language Processing in Medicine: Systematic Literature Review

BACKGROUND

Ensuring access to accurate and verified information is essential for effective patient treatment and diagnosis. Although health workers rely on the internet for clinical data, there is a need for a more streamlined approach.

OBJECTIVE

This systematic review aims to assess the current state of artificial intelligence (AI) and natural language processing (NLP) techniques in health care to identify their potential use in electronic health records and automated information searches.

METHODS

A search was conducted in the PubMed, Embase, ScienceDirect, Scopus, and Web of Science online databases for articles published between January 2000 and April 2023. The only inclusion criteria were (1) original research articles and studies on the application of AI-based medical clinical decision support using NLP techniques and (2) publications in English. A Critical Appraisal Skills Programme tool was used to assess the quality of the studies.

RESULTS

The search yielded 707 articles, from which 26 studies were included (24 original articles and 2 systematic reviews). Of the evaluated articles, 21 (81%) explained the use of NLP as a source of data collection, 18 (69%) used electronic health records as a data source, and a further 8 (31%) were based on clinical data. Only 5 (19%) of the articles showed the use of combined strategies for NLP to obtain clinical data. In total, 16 (62%) articles presented stand-alone data review algorithms. Other studies (n=9, 35%) showed that the clinical decision support system alternative was also a way of displaying the information obtained for immediate clinical use.

CONCLUSIONS

The use of NLP engines can effectively improve clinical decision systems' accuracy, while biphasic tools combining AI algorithms and human criteria may optimize clinical diagnosis and treatment flows.

TRIAL REGISTRATION

PROSPERO CRD42022373386; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=373386.

Using Natural Language Processing to Measure and Improve Quality of Diabetes Care: A Systematic Review

BACKGROUND

Real-world evidence research plays an increasingly important role in diabetes care. However, a large fraction of real-world data are "locked" in narrative format. Natural language processing (NLP) technology offers a solution for analysis of narrative electronic data.

METHODS

We conducted a systematic review of studies of NLP technology focused on diabetes. Articles published prior to June 2020 were included.

RESULTS

We included 38 studies in the analysis. The majority (24; 63.2%) described only development of NLP tools; the remainder used NLP tools to conduct clinical research. A large fraction (17; 44.7%) of studies focused on identification of patients with diabetes; the rest covered a broad range of subjects that included hypoglycemia, lifestyle counseling, diabetic kidney disease, insulin therapy and others. The mean F₁ score for all studies where it was available was 0.882. It tended to be lower (0.817) in studies of more linguistically complex concepts. Seven studies reported findings with potential implications for improving delivery of diabetes care.

CONCLUSION

Research in NLP technology to study diabetes is growing quickly, although challenges (e.g. in analysis of more linguistically complex concepts) remain. Its potential to deliver evidence on treatment and improving quality of diabetes care is demonstrated by a number of studies. Further growth in this area would be aided by deeper collaboration between developers and end-users of natural language processing tools as well as by broader sharing of the tools themselves and related resources.

Evaluating shallow and deep learning strategies for the 2018 n2c2 shared task on clinical text classification

Objective

Automated clinical phenotyping is challenging because word-based features quickly turn it into a high-dimensional problem, in which the small, privacy-restricted, training datasets might lead to overfitting. Pretrained embeddings might solve this issue by reusing input representation schemes trained on a larger dataset. We sought to evaluate shallow and deep learning text classifiers and the impact of pretrained embeddings in a small clinical dataset.

Materials and Methods

We participated in the 2018 National NLP Clinical Challenges (n2c2) Shared Task on cohort selection and received an annotated dataset with medical narratives of 202 patients for multilabel binary text classification. We set our baseline to a majority classifier, to which we compared a rule-based classifier and orthogonal machine learning strategies: support vector machines, logistic regression, and long short-term memory neural networks. We evaluated logistic regression and long short-term memory using both self-trained and pretrained BioWordVec word embeddings as input representation schemes.

Results

Rule-based classifier showed the highest overall micro F₁ score (0.9100), with which we finished first in the challenge. Shallow machine learning strategies showed lower overall micro F₁ scores, but still higher than deep learning strategies and the baseline. We could not show a difference in classification efficiency between self-trained and pretrained embeddings.

Discussion

Clinical context, negation, and value-based criteria hindered shallow machine learning approaches, while deep learning strategies could not capture the term diversity due to the small training dataset.

Conclusion

Shallow methods for clinical phenotyping can still outperform deep learning methods in small imbalanced data, even when supported by pretrained embeddings.

BioHackathon 2015: Semantics of data for life sciences and reproducible research

We report on the activities of the 2015 edition of the BioHackathon, an annual event that brings together researchers and developers from around the world to develop tools and technologies that promote the reusability of biological data. We discuss issues surrounding the representation, publication, integration, mining and reuse of biological data and metadata across a wide range of biomedical data types of relevance for the life sciences, including chemistry, genotypes and phenotypes, orthology and phylogeny, proteomics, genomics, glycomics, and metabolomics. We describe our progress to address ongoing challenges to the reusability and reproducibility of research results, and identify outstanding issues that continue to impede the progress of bioinformatics research. We share our perspective on the state of the art, continued challenges, and goals for future research and development for the life sciences Semantic Web.

Cohort Selection for Clinical Trials From Longitudinal Patient Records: Text Mining Approach

Background

Clinical trials are an important step in introducing new interventions into clinical practice by generating data on their safety and efficacy. Clinical trials need to ensure that participants are similar so that the findings can be attributed to the interventions studied and not to some other factors. Therefore, each clinical trial defines eligibility criteria, which describe characteristics that must be shared by the participants. Unfortunately, the complexities of eligibility criteria may not allow them to be translated directly into readily executable database queries. Instead, they may require careful analysis of the narrative sections of medical records. Manual screening of medical records is time consuming, thus negatively affecting the timeliness of the recruitment process.

Objective

Track 1 of the 2018 National Natural Language Processing Clinical Challenge focused on the task of cohort selection for clinical trials, aiming to answer the following question: Can natural language processing be applied to narrative medical records to identify patients who meet eligibility criteria for clinical trials? The task required the participating systems to analyze longitudinal patient records to determine if the corresponding patients met the given eligibility criteria. We aimed to describe a system developed to address this task.

Methods

Our system consisted of 13 classifiers, one for each eligibility criterion. All classifiers used a bag-of-words document representation model. To prevent the loss of relevant contextual information associated with such representation, a pattern-matching approach was used to extract context-sensitive features. They were embedded back into the text as lexically distinguishable tokens, which were consequently featured in the bag-of-words representation. Supervised machine learning was chosen wherever a sufficient number of both positive and negative instances was available to learn from. A rule-based approach focusing on a small set of relevant features was chosen for the remaining criteria.

Results

The system was evaluated using microaveraged F measure. Overall, 4 machine algorithms, including support vector machine, logistic regression, naïve Bayesian classifier, and gradient tree boosting (GTB), were evaluated on the training data using 10–fold cross-validation. Overall, GTB demonstrated the most consistent performance. Its performance peaked when oversampling was used to balance the training data. The final evaluation was performed on previously unseen test data. On average, the F measure of 89.04% was comparable to 3 of the top ranked performances in the shared task (91.11%, 90.28%, and 90.21%). With an F measure of 88.14%, we significantly outperformed these systems (81.03%, 78.50%, and 70.81%) in identifying patients with advanced coronary artery disease.

Conclusions

The holdout evaluation provides evidence that our system was able to identify eligible patients for the given clinical trial with high accuracy. Our approach demonstrates how rule-based knowledge infusion can improve the performance of machine learning algorithms even when trained on a relatively small dataset.

Feature extraction for phenotyping from semantic and knowledge resources

OBJECTIVE

Phenotyping algorithms can efficiently and accurately identify patients with a specific disease phenotype and construct electronic health records (EHR)-based cohorts for subsequent clinical or genomic studies. Previous studies have introduced unsupervised EHR-based feature selection methods that yielded algorithms with high accuracy. However, those selection methods still require expert intervention to tweak the parameter settings according to the EHR data distribution for each phenotype. To further accelerate the development of phenotyping algorithms, we propose a fully automated and robust unsupervised feature selection method that leverages only publicly available medical knowledge sources, instead of EHR data.

METHODS

SEmantics-Driven Feature Extraction (SEDFE) collects medical concepts from online knowledge sources as candidate features and gives them vector-form distributional semantic representations derived with neural word embedding and the Unified Medical Language System Metathesaurus. A number of features that are semantically closest and that sufficiently characterize the target phenotype are determined by a linear decomposition criterion and are selected for the final classification algorithm.

RESULTS

SEDFE was compared with the EHR-based SAFE algorithm and domain experts on feature selection for the classification of five phenotypes including coronary artery disease, rheumatoid arthritis, Crohn's disease, ulcerative colitis, and pediatric pulmonary arterial hypertension using both supervised and unsupervised approaches. Algorithms yielded by SEDFE achieved comparable accuracy to those yielded by SAFE and expert-curated features. SEDFE is also robust to the input semantic vectors.

CONCLUSION

SEDFE attains satisfying performance in unsupervised feature selection for EHR phenotyping. Both fully automated and EHR-independent, this method promises efficiency and accuracy in developing algorithms for high-throughput phenotyping.

Bias of Inaccurate Disease Mentions in Electronic Health Record-based Phenotyping

OBJECTIVES

Electronic health record (EHR)-based phenotyping is an automated technique for identifying patients diagnosed with a particular disease using EHR data. However, EHR-based phenotyping has difficulties in achieving satisfactorily high performance because clinical notes include disease mentions that ultimately signify something other than the patient's diagnosis (such as differential diagnosis or screening). Our objective is to quantify the influence of such disease mentions on EHR-based phenotyping performance.

METHODS

Physicians manually reviewed whether the disease mentions indicated the patients' diseases in 487,300 clinical notes of 4,430 patients. Particular focus was placed on disease mentions that did not signify the patient's diagnosis even though they did not have any syntactic modifier or indicator in the same sentences. Patients were then classified according to whether their clinical notes included such disease mentions.

RESULTS

Among the patients whose clinical notes included disease mentions without any modifier or indicator, the proportion of patients whose disease mentions signified the patients' diagnosis was 78.1% (on average). This value can be interpreted as the bias of disease mentions that did not signify the patient's diagnosis on the precision of EHR-based phenotyping by extracting disease mentions from clinical notes.

CONCLUSION

This study quantified the bias occurred owing to disease mentions that incorrectly signify a patient's diagnosis in the value of precision of EHR-based phenotyping from four dataset types. The results of this study will help researchers in diverse research environments with different available data types.

Whale optimized mixed kernel function of support vector machine for colorectal cancer diagnosis

Microarray technique is a prevalent method for the classification and prediction of colorectal cancer (CRC). Nevertheless, microarray data suffers from the curse of dimensionality when selecting feature genes of the disease based on imbalance samples, thus causing low prediction accuracy. Hence, it is of vital significance to build proper models that can avoid the above problems and predict the CRC more accurately. In this paper, we use an ensemble model to classify samples into healthy and CRC groups and improve prediction performance. The proposed model is composed of three functional modules. The first module mainly performs the function of removing redundant genes. The main feature genes are selected using minimum redundancy maximum relevance (mRMR) method to reduce the dimensionality of features thereby increasing the prediction results. The second module aims to solve the problem caused by imbalanced data using hybrid sampling algorithm RUSBoost. The third module focuses on the classification algorithm optimization. We use mixed kernel function (MKF) based support vector machine (SVM) model to classify an unknown sample into healthy individuals and CRC patients, and then, the Whale Optimization Algorithm (WOA) is applied to find most optimal parameters of the proposed MKF-SVM. The final results show that the proposed model achieves higher G-means than other comparable models. The conclusion comes to show that RUSBoost wrapping WOA + MKF-SVM model can be applied to improve the predictive performance of colorectal cancer based on the imbalanced data.

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.

Development and validation of various phenotyping algorithms for Diabetes Mellitus using data from electronic health records

BACKGROUND AND OBJECTIVE

Recent progression towards precision medicine has encouraged the use of electronic health records (EHRs) as a source for large amounts of data, which is required for studying the effect of treatments or risk factors in more specific subpopulations. Phenotyping algorithms allow to automatically classify patients according to their particular electronic phenotype thus facilitating the setup of retrospective cohorts. Our objective is to compare the performance of different classification strategies (only using standardized problems, rule-based algorithms, statistical learning algorithms (six learners) and stacked generalization (five versions)), for the categorization of patients according to their diabetic status (diabetics, not diabetics and inconclusive; Diabetes of any type) using information extracted from EHRs.

METHODS

Patient information was extracted from the EHR at Hospital Italiano de Buenos Aires, Buenos Aires, Argentina. For the derivation and validation datasets, two probabilistic samples of patients from different years (2005: n = 1663; 2015: n = 800) were extracted. The only inclusion criterion was age (≥40 & <80 years). Four researchers manually reviewed all records and classified patients according to their diabetic status (diabetic: diabetes registered as a health problem or fulfilling the ADA criteria; non-diabetic: not fulfilling the ADA criteria and having at least one fasting glycemia below 126 mg/dL; inconclusive: no data regarding their diabetic status or only one abnormal value). The best performing algorithms within each strategy were tested on the validation set.

RESULTS

The standardized codes algorithm achieved a Kappa coefficient value of 0.59 (95% CI 0.49, 0.59) in the validation set. The Boolean logic algorithm reached 0.82 (95% CI 0.76, 0.88). A slightly higher value was achieved by the Feedforward Neural Network (0.9, 95% CI 0.85, 0.94). The best performing learner was the stacked generalization meta-learner that reached a Kappa coefficient value of 0.95 (95% CI 0.91, 0.98).

CONCLUSIONS

The stacked generalization strategy and the feedforward neural network showed the best classification metrics in the validation set. The implementation of these algorithms enables the exploitation of the data of thousands of patients accurately.

Ensembles of NLP Tools for Data Element Extraction from Clinical Notes

Natural Language Processing (NLP) is essential for concept extraction from narrative text in electronic health records (EHR). To extract numerous and diverse concepts, such as data elements (i.e., important concepts related to a certain medical condition), a plausible solution is to combine various NLP tools into an ensemble to improve extraction performance. However, it is unclear to what extent ensembles of popular NLP tools improve the extraction of numerous and diverse concepts. Therefore, we built an NLP ensemble pipeline to synergize the strength of popular NLP tools using seven ensemble methods, and to quantify the improvement in performance achieved by ensembles in the extraction of data elements for three very different cohorts. Evaluation results show that the pipeline can improve the performance of NLP tools, but there is high variability depending on the cohort.

Natural Language Processing Based Instrument for Classification of Free Text Medical Records

According to the Ministry of Labor, Health and Social Affairs of Georgia a new health management system has to be introduced in the nearest future. In this context arises the problem of structuring and classifying documents containing all the history of medical services provided. The present work introduces the instrument for classification of medical records based on the Georgian language. It is the first attempt of such classification of the Georgian language based medical records. On the whole 24.855 examination records have been studied. The documents were classified into three main groups (ultrasonography, endoscopy, and X-ray) and 13 subgroups using two well-known methods: Support Vector Machine (SVM) and K-Nearest Neighbor (KNN). The results obtained demonstrated that both machine learning methods performed successfully, with a little supremacy of SVM. In the process of classification a “shrink” method, based on features selection, was introduced and applied. At the first stage of classification the results of the “shrink” case were better; however, on the second stage of classification into subclasses 23% of all documents could not be linked to only one definite individual subclass (liver or binary system) due to common features characterizing these subclasses. The overall results of the study were successful.