Natural language processing for structuring clinical text data on depression using UK-CRIS

Annotation of epilepsy clinic letters for natural language processing

BACKGROUND

Natural language processing (NLP) is increasingly being used to extract structured information from unstructured text to assist clinical decision-making and aid healthcare research. The availability of expert-annotated documents for the development and validation of NLP applications is limited. We created synthetic clinical documents to address this, and to validate the Extraction of Epilepsy Clinical Text version 2 (ExECTv2) NLP pipeline.

METHODS

We created 200 synthetic clinic letters based on hospital outpatient consultations with epilepsy specialists. The letters were double annotated by trained clinicians and researchers according to agreed guidelines. We used the annotation tool, Markup, with an epilepsy concept list based on the Unified Medical Language System ontology. All annotations were reviewed, and a gold standard set of annotations was agreed and used to validate the performance of ExECTv2.

RESULTS

The overall inter-annotator agreement (IAA) between the two sets of annotations produced a per item F1 score of 0.73. Validating ExECTv2 using the gold standard gave an overall F1 score of 0.87 per item, and 0.90 per letter.

CONCLUSION

The synthetic letters, annotations, and annotation guidelines have been made freely available. To our knowledge, this is the first publicly available set of annotated epilepsy clinic letters and guidelines that can be used for NLP researchers with minimum epilepsy knowledge. The IAA results show that clinical text annotation tasks are difficult and require a gold standard to be arranged by researcher consensus. The results for ExECTv2, our automated epilepsy NLP pipeline, extracted detailed epilepsy information from unstructured epilepsy letters with more accuracy than human annotators, further confirming the utility of NLP for clinical and research applications.

Identification of risk factors for the onset of delirium associated with COVID-19 by mining nursing records

COVID-19 has a range of complications, from no symptoms to severe pneumonia. It can also affect multiple organs including the nervous system. COVID-19 affects the brain, leading to neurological symptoms such as delirium. Delirium, a sudden change in consciousness, can increase the risk of death and prolong the hospital stay. However, research on delirium prediction in patients with COVID-19 is insufficient. This study aimed to identify new risk factors that could predict the onset of delirium in patients with COVID-19 using machine learning (ML) applied to nursing records. This retrospective cohort study used natural language processing and ML to develop a model for classifying the nursing records of patients with delirium. We extracted the features of each word from the model and grouped similar words. To evaluate the usefulness of word groups in predicting the occurrence of delirium in patients with COVID-19, we analyzed the temporal changes in the frequency of occurrence of these word groups before and after the onset of delirium. Moreover, the sensitivity, specificity, and odds ratios were calculated. We identified (1) elimination-related behaviors and conditions and (2) abnormal patient behavior and conditions as risk factors for delirium. Group 1 had the highest sensitivity (0.603), whereas group 2 had the highest specificity and odds ratio (0.938 and 6.903, respectively). These results suggest that these parameters may be useful in predicting delirium in these patients. The risk factors for COVID-19-associated delirium identified in this study were more specific but less sensitive than the ICDSC (Intensive Care Delirium Screening Checklist) and CAM-ICU (Confusion Assessment Method for the Intensive Care Unit). However, they are superior to the ICDSC and CAM-ICU because they can predict delirium without medical staff and at no cost.

Use of Natural Language Processing to Identify Sexual and Reproductive Health Information in Clinical Text

OBJECTIVES

This study aimed to enable clinical researchers without expertise in natural language processing (NLP) to extract and analyze information about sexual and reproductive health (SRH), or other sensitive health topics, from large sets of clinical notes.

METHODS

(1) We retrieved text from the electronic health record as individual notes. (2) We segmented notes into sentences using one of scispaCy's NLP toolkits. (3) We exported sentences to the labeling application Watchful and annotated subsets of these as relevant or irrelevant to various SRH categories by applying a combination of regular expressions and manual annotation. (4) The labeled sentences served as training data to create machine learning models for classifying text; specifically, we used spaCy's default text classification ensemble, comprising a bag-of-words model and a neural network with attention. (5) We applied each model to unlabeled sentences to identify additional references to SRH with novel relevant vocabulary. We used this information and repeated steps 3 to 5 iteratively until the models identified no new relevant sentences for each topic. Finally, we aggregated the labeled data for analysis.

RESULTS

This methodology was applied to 3,663 Child Neurology notes for 971 female patients. Our search focused on six SRH categories. We validated the approach using two subject matter experts, who independently labeled a sample of 400 sentences. Cohen's kappa values were calculated for each category between the reviewers (menstruation: 1, sexual activity: 0.9499, contraception: 0.9887, folic acid: 1, teratogens: 0.8864, pregnancy: 0.9499). After removing the sentences on which reviewers did not agree, we compared the reviewers' labels to those produced via our methodology, again using Cohen's kappa (menstruation: 1, sexual activity: 1, contraception: 0.9885, folic acid: 1, teratogens: 0.9841, pregnancy: 0.9871).

CONCLUSION

Our methodology is reproducible, enables analysis of large amounts of text, and has produced results that are highly comparable to subject matter expert manual review.

Natural language processing for mental health interventions: a systematic review and research framework

Neuropsychiatric disorders pose a high societal cost, but their treatment is hindered by lack of objective outcomes and fidelity metrics. AI technologies and specifically Natural Language Processing (NLP) have emerged as tools to study mental health interventions (MHI) at the level of their constituent conversations. However, NLP's potential to address clinical and research challenges remains unclear. We therefore conducted a pre-registered systematic review of NLP-MHI studies using PRISMA guidelines (osf.io/s52jh) to evaluate their models, clinical applications, and to identify biases and gaps. Candidate studies (n = 19,756), including peer-reviewed AI conference manuscripts, were collected up to January 2023 through PubMed, PsycINFO, Scopus, Google Scholar, and ArXiv. A total of 102 articles were included to investigate their computational characteristics (NLP algorithms, audio features, machine learning pipelines, outcome metrics), clinical characteristics (clinical ground truths, study samples, clinical focus), and limitations. Results indicate a rapid growth of NLP MHI studies since 2019, characterized by increased sample sizes and use of large language models. Digital health platforms were the largest providers of MHI data. Ground truth for supervised learning models was based on clinician ratings (n = 31), patient self-report (n = 29) and annotations by raters (n = 26). Text-based features contributed more to model accuracy than audio markers. Patients' clinical presentation (n = 34), response to intervention (n = 11), intervention monitoring (n = 20), providers' characteristics (n = 12), relational dynamics (n = 14), and data preparation (n = 4) were commonly investigated clinical categories. Limitations of reviewed studies included lack of linguistic diversity, limited reproducibility, and population bias. A research framework is developed and validated (NLPxMHI) to assist computational and clinical researchers in addressing the remaining gaps in applying NLP to MHI, with the goal of improving clinical utility, data access, and fairness.

Few-shot learning for medical text: A review of advances, trends, and opportunities

BACKGROUND

Few-shot learning (FSL) is a class of machine learning methods that require small numbers of labeled instances for training. With many medical topics having limited annotated text-based data in practical settings, FSL-based natural language processing (NLP) holds substantial promise. We aimed to conduct a review to explore the current state of FSL methods for medical NLP.

METHODS

We searched for articles published between January 2016 and October 2022 using PubMed/Medline, Embase, ACL Anthology, and IEEE Xplore Digital Library. We also searched the preprint servers (e.g., arXiv, medRxiv, and bioRxiv) via Google Scholar to identify the latest relevant methods. We included all articles that involved FSL and any form of medical text. We abstracted articles based on the data source, target task, training set size, primary method(s)/approach(es), and evaluation metric(s).

RESULTS

Fifty-one articles met our inclusion criteria-all published after 2018, and most since 2020 (42/51; 82%). Concept extraction/named entity recognition was the most frequently addressed task (21/51; 41%), followed by text classification (16/51; 31%). Thirty-two (61%) articles reconstructed existing datasets to fit few-shot scenarios, and MIMIC-III was the most frequently used dataset (10/51; 20%). 77% of the articles attempted to incorporate prior knowledge to augment the small datasets available for training. Common methods included FSL with attention mechanisms (20/51; 39%), prototypical networks (11/51; 22%), meta-learning (7/51; 14%), and prompt-based learning methods, the latter being particularly popular since 2021. Benchmarking experiments demonstrated relative underperformance of FSL methods on biomedical NLP tasks.

CONCLUSION

Despite the potential for FSL in biomedical NLP, progress has been limited. This may be attributed to the rarity of specialized data, lack of standardized evaluation criteria, and the underperformance of FSL methods on biomedical topics. The creation of publicly-available specialized datasets for biomedical FSL may aid method development by facilitating comparative analyses.

Translating clinical notes into quantitative measures-a real-world observation on the response to cholinesterase inhibitors or selective serotonin reuptake inhibitors prescribed to outpatients with dementia using electronic medical records

Objective: Cholinesterase inhibitors (CEIs) are prescribed for dementia to maintain or improve memory. Selective serotonin reuptake inhibitors (SSRIs) are also prescribed to manage psychiatric symptoms seen in dementia. What proportion of outpatients actually responds to these drugs is still unclear. Our objective was to investigate the responder rates of these medications in an outpatient setting using the electronic medical record (EMR). Methods: We used the Johns Hopkins EMR system to identify patients with dementia who were prescribed a CEI or SSRI for the first time between 2010 and 2021. Treatment effects were assessed through routinely documented clinical notes and free-text entries in which healthcare providers record clinical findings and impressions of patients. Responses were scored using a three-point Likert scale named the NOte-based evaluation method for Treatment Efficacy (NOTE) in addition to the Clinician's Interview-Based Impression of Change Plus caregiver input (CIBIC-plus), a seven-point Likert scale used in clinical trials. To validate NOTE, the relationships between NOTE and CIBIC-plus and between NOTE and change in MMSE (Mini-Mental State Examination) before and after medication were examined. Inter-rater reliability was evaluated using Krippendorff's alpha. The responder rates were calculated. Results: NOTE showed excellent inter-rater reliability and correlated well with CIBIC-plus and changes in MMSEs. Out of 115 CEI cases, 27.0% reported improvement and 34.8% reported stable symptoms in cognition; out of 225 SSRI cases, 69.3% reported an improvement in neuropsychiatric symptoms. Conclusion: NOTE showed high validity in measuring the pharmacotherapy effects based on unstructured clinical entries. Although our real-world observation included various types of dementia, the results were remarkably similar to what was reported in controlled clinical trials of Alzheimer's disease and its related neuropsychiatric symptoms.

Demystifying the Role of Natural Language Processing (NLP) in Smart City Applications: Background, Motivation, Recent Advances, and Future Research Directions

Smart cities provide an efficient infrastructure for the enhancement of the quality of life of the people by aiding in fast urbanization and resource management through sustainable and scalable innovative solutions. The penetration of Information and Communication Technology (ICT) in smart cities has been a major contributor to keeping up with the agility and pace of their development. In this paper, we have explored Natural Language Processing (NLP) which is one such technical discipline that has great potential in optimizing ICT processes and has so far been kept away from the limelight. Through this study, we have established the various roles that NLP plays in building smart cities after thoroughly analyzing its architecture, background, and scope. Subsequently, we present a detailed description of NLP's recent applications in the domain of smart healthcare, smart business, and industry, smart community, smart media, smart research, and development as well as smart education accompanied by NLP's open challenges at the very end. This work aims to throw light on the potential of NLP as one of the pillars in assisting the technical advancement and realization of smart cities.

Natural Language Processing in Electronic Health Records in relation to healthcare decision-making: A systematic review

BACKGROUND

Natural Language Processing (NLP) is widely used to extract clinical insights from Electronic Health Records (EHRs). However, the lack of annotated data, automated tools, and other challenges hinder the full utilisation of NLP for EHRs. Various Machine Learning (ML), Deep Learning (DL) and NLP techniques are studied and compared to understand the limitations and opportunities in this space comprehensively.

METHODOLOGY

After screening 261 articles from 11 databases, we included 127 papers for full-text review covering seven categories of articles: (1) medical note classification, (2) clinical entity recognition, (3) text summarisation, (4) deep learning (DL) and transfer learning architecture, (5) information extraction, (6) Medical language translation and (7) other NLP applications. This study follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

RESULT AND DISCUSSION

EHR was the most commonly used data type among the selected articles, and the datasets were primarily unstructured. Various ML and DL methods were used, with prediction or classification being the most common application of ML or DL. The most common use cases were: the International Classification of Diseases, Ninth Revision (ICD-9) classification, clinical note analysis, and named entity recognition (NER) for clinical descriptions and research on psychiatric disorders.

CONCLUSION

We find that the adopted ML models were not adequately assessed. In addition, the data imbalance problem is quite important, yet we must find techniques to address this underlining problem. Future studies should address key limitations in studies, primarily identifying Lupus Nephritis, Suicide Attempts, perinatal self-harmed and ICD-9 classification.

Application of Natural Language Processing (NLP) in Detecting and Preventing Suicide Ideation: A Systematic Review

(1) Introduction: Around a million people are reported to die by suicide every year, and due to the stigma associated with the nature of the death, this figure is usually assumed to be an underestimate. Machine learning and artificial intelligence such as natural language processing has the potential to become a major technique for the detection, diagnosis, and treatment of people. (2) Methods: PubMed, EMBASE, MEDLINE, PsycInfo, and Global Health databases were searched for studies that reported use of NLP for suicide ideation or self-harm. (3) Result: The preliminary search of 5 databases generated 387 results. Removal of duplicates resulted in 158 potentially suitable studies. Twenty papers were finally included in this review. (4) Discussion: Studies show that combining structured and unstructured data in NLP data modelling yielded more accurate results than utilizing either alone. Additionally, to reduce suicides, people with mental problems must be continuously and passively monitored. (5) Conclusions: The use of AI&ML opens new avenues for considerably guiding risk prediction and advancing suicide prevention frameworks. The review's analysis of the included research revealed that the use of NLP may result in low-cost and effective alternatives to existing resource-intensive methods of suicide prevention.

Automated Identification of Clinical Procedures in Free-Text Electronic Clinical Records with a Low-Code Named Entity Recognition Workflow

INTRODUCTION

 Clinical procedures are often performed in outpatient clinics without prior scheduling at the administrative level, and documentation of the procedure often occurs solely in free-text clinical electronic notes. Natural language processing (NLP), particularly named entity recognition (NER), may provide a solution to extracting procedure data from free-text electronic notes.

METHODS

 Free-text notes from outpatient ophthalmology visits were collected from the electronic clinical records at a single institution over 3 months. The Prodigy low-code annotation tool was used to create an annotation dataset and train a custom NER model for clinical procedures. Clinical procedures were extracted from the entire set of clinical notes.

RESULTS

 There were a total of 5,098 clinic notes extracted for the study period; 1,923 clinic notes were used to build the NER model, which included a total of 231 manual annotations. The NER model achieved an F-score of 0.767, a precision of 0.810, and a recall of 0.729. The most common procedures performed included intravitreal injections of therapeutic substances, removal of corneal foreign bodies, and epithelial debridement of corneal ulcers.

CONCLUSIONS

 The use of a low-code annotation software tool allows the rapid creation of a custom annotation dataset to train a NER model to identify clinical procedures stored in free-text electronic clinical notes. This enables clinicians to rapidly gather previously unidentified procedural data for quality improvement and auditing purposes. Low-code annotation tools may reduce time and coding barriers to clinician participation in NLP research.

Multiscale neural signatures of major depressive, anxiety, and stress-related disorders

The extent of shared and distinct neural mechanisms underlying major depressive disorder (MDD), anxiety, and stress-related disorders is still unclear. We compared the neural signatures of these disorders in 5,405 UK Biobank patients and 21,727 healthy controls. We found the greatest case–control differences in resting-state functional connectivity and cortical thickness in MDD, followed by anxiety and stress-related disorders. Neural signatures for MDD and anxiety disorders were highly concordant, whereas stress-related disorders showed a distinct pattern. Controlling for cross-disorder genetic risk somewhat decreased the similarity between functional neural signatures of stress-related disorders and both MDD and anxiety disorders. Among cases and healthy controls, reduced within-network and increased between-network frontoparietal and default mode connectivity were associated with poorer cognitive performance (processing speed, attention, associative learning, and fluid intelligence). These results provide evidence for distinct neural circuit function impairments in MDD and anxiety disorders compared to stress disorders, yet cognitive impairment appears unrelated to diagnosis and varies with circuit function.

The burden associated with, and management of, difficult-to-treat depression in patients under specialist psychiatric care in the United Kingdom

BACKGROUND

Major depressive disorder (MDD) is common and often has sub-optimal response to treatment. Difficult-to-treat depression (DTD) is a new concept that describes 'depression that continues to cause significant burden despite usual treatment efforts'.

AIMS

To identify patients with likely DTD in UK secondary care and examine demographic, disease and treatment data as compared with 'non-DTD' MDD patients.

METHODS

Anonymised electronic health records (EHRs) of five specialist mental health National Health Service (NHS) Trusts in the United Kingdom were analysed using a natural language processing model. Data on disease characteristics, comorbidities and treatment histories were extracted from structured fields and using natural language algorithms from unstructured fields. Patients with MDD aged ⩾18 years were included in the analysis; those with presumed DTD were identified on the basis of MDD history (duration and recurrence) and number of treatments prescribed.

RESULTS

In a sample of 28,184 patients with MDD, 19% met criteria for DTD. Compared to the non-DTD group, patients with DTD were more likely to have severe depression, suicidal ideation, and comorbid psychiatric and/or physical illness, as well as higher rates of hospitalisation. They were also more likely to be in receipt of unemployment and sickness/disability benefits. More intensive treatment strategies were used in the DTD group, including higher rates of combination therapy, augmentation, psychotherapy and electroconvulsive therapy.

CONCLUSION

This study demonstrates the feasibility of identifying patients with probable DTD from EHRs and highlights the increased burden associated with MDD in these patients.

Exploring the Association of Cancer and Depression in Electronic Health Records: Combining Encoded Diagnosis and Mining Free Text Clinical Notes (Preprint)

Background

A cancer diagnosis is a source of psychological and emotional stress, which are often maintained for sustained periods of time that may lead to depressive disorders. Depression is one of the most common psychological conditions in patients with cancer. According to the Global Cancer Observatory, breast and colorectal cancers are the most prevalent cancers in both sexes and across all age groups in Spain.

Objective

This study aimed to compare the prevalence of depression in patients before and after the diagnosis of breast or colorectal cancer, as well as to assess the usefulness of the analysis of free-text clinical notes in 2 languages (Spanish or Catalan) for detecting depression in combination with encoded diagnoses.

Methods

We carried out an analysis of the electronic health records from a general hospital by considering the different sources of clinical information related to depression in patients with breast and colorectal cancer. This analysis included ICD-9-CM (International Classification of Diseases, Ninth Revision, Clinical Modification) diagnosis codes and unstructured information extracted by mining free-text clinical notes via natural language processing tools based on Systematized Nomenclature of Medicine Clinical Terms that mentions symptoms and drugs used for the treatment of depression.

Results

We observed that the percentage of patients diagnosed with depressive disorders significantly increased after cancer diagnosis in the 2 types of cancer considered—breast and colorectal cancers. We managed to identify a higher number of patients with depression by mining free-text clinical notes than the group selected exclusively on ICD-9-CM codes, increasing the number of patients diagnosed with depression by 34.8% (441/1269). In addition, the number of patients with depression who received chemotherapy was higher than those who did not receive this treatment, with significant differences (P<.001).

Conclusions

This study provides new clinical evidence of the depression-cancer comorbidity and supports the use of natural language processing for extracting and analyzing free-text clinical notes from electronic health records, contributing to the identification of additional clinical data that complements those provided by coded data to improve the management of these patients.

Technology-Based Mental Health Interventions for Domestic Violence Victims Amid COVID-19

INTRODUCTION

Domestic violence is a threat to human dignity and public health. Mounting evidence shows that domestic violence erodes personal and public health, spawning issues such as lifelong mental health challenges. To further compound the situation, COVID-19 and societies' poor response to the pandemic have not only worsened the domestic violence crisis but also disrupted mental health services for domestic violence victims. While technology-based health solutions can overcome physical constraints posed by the pandemic and offer timely support to address domestic violence victims' mental health issues, there is a dearth of research in the literature. To bridge the research gap, in this study, we aim to examine technology-based mental health solutions for domestic violence victims amid COVID-19.

METHODS

A literature review was conducted to examine solutions that domestic violence victims can utilize to safeguard and improve their mental health amid COVID-19. Databases including PubMed, PsycINFO, and Scopus were utilized for the literature search. The search was focused on four themes: domestic violence, mental health, technology-based interventions, and COVID-19. A reverse search of pertinent references was conducted in Google Scholar. The social ecological model was utilized to systematically structure the review findings.

RESULTS

The findings show that a wide array of technology-based solutions has been proposed to address mental health challenges faced by domestic violence victims amid COVID-19. However, none of these proposals is based on empirical evidence amid COVID-19. In terms of social and ecological levels of influence, most of the interventions were developed on the individual level, as opposed to the community level or social level, effectively placing the healthcare responsibility on the victims rather than government and health officials. Furthermore, most of the articles failed to address risks associated with utilizing technology-based interventions (e.g., privacy issues) or navigating the online environment (e.g., cyberstalking).

CONCLUSION

Overall, our findings highlight the need for greater research endeavors on the research topic. Although technology-based interventions have great potential in resolving domestic violence victims' mental health issues, risks associated with these health solutions should be comprehensively acknowledged and addressed.

Validation of UK Biobank data for mental health outcomes: A pilot study using secondary care electronic health records

UK Biobank (UKB) is widely employed to investigate mental health disorders and related exposures; however, its applicability and relevance in a clinical setting and the assumptions required have not been sufficiently and systematically investigated. Here, we present the first validation study using secondary care mental health data with linkage to UKB from Oxford - Clinical Record Interactive Search (CRIS) focusing on comparison of demographic information, diagnostic outcome, medication record and cognitive test results, with missing data and the implied bias from both resources depicted. We applied a natural language processing model to extract information embedded in unstructured text from clinical notes and attachments. Using a contingency table we compared the demographic information recorded in UKB and CRIS. We calculated the positive predictive value (PPV, proportion of true positives cases detected) for mental health diagnosis and relevant medication. Amongst the cohort of 854 subjects, PPVs for any mental health diagnosis for dementia, depression, bipolar disorder and schizophrenia were 41.6%, and were 59.5%, 12.5%, 50.0% and 52.6%, respectively. Self-reported medication records in UKB had general PPV of 47.0%, with the prevalence of frequently prescribed medicines to each typical mental health disorder considerably different from the information provided by CRIS. UKB is highly multimodal, but with limited follow-up records, whereas CRIS offers a longitudinal high-resolution clinical picture with more than ten years of observations. The linkage of both datasets will reduce the self-report bias and synergistically augment diverse modalities into a unified resource to facilitate more robust research in mental health.

Personalised treatment for cognitive impairment in dementia: development and validation of an artificial intelligence model

BACKGROUND

Donepezil, galantamine, rivastigmine and memantine are potentially effective interventions for cognitive impairment in dementia, but the use of these drugs has not been personalised to individual patients yet. We examined whether artificial intelligence-based recommendations can identify the best treatment using routinely collected patient-level information.

METHODS

Six thousand eight hundred four patients aged 59-102 years with a diagnosis of dementia from two National Health Service (NHS) Foundation Trusts in the UK were used for model training/internal validation and external validation, respectively. A personalised prescription model based on the Recurrent Neural Network machine learning architecture was developed to predict the Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA) scores post-drug initiation. The drug that resulted in the smallest decline in cognitive scores between prescription and the next visit was selected as the treatment of choice. Change of cognitive scores up to 2 years after treatment initiation was compared for model evaluation.

RESULTS

Overall, 1343 patients with MMSE scores were identified for internal validation and 285 [21.22%] took the drug recommended. After 2 years, the reduction of mean [standard deviation] MMSE score in this group was significantly smaller than the remaining 1058 [78.78%] patients (0.60 [0.26] vs 2.80 [0.28]; P = 0.02). In the external validation cohort (N = 1772), 222 [12.53%] patients took the drug recommended and reported a smaller MMSE reduction compared to the 1550 [87.47%] patients who did not (1.01 [0.49] vs 4.23 [0.60]; P = 0.01). A similar performance gap was seen when testing the model on patients prescribed with AChEIs only.

CONCLUSIONS

It was possible to identify the most effective drug for the real-world treatment of cognitive impairment in dementia at an individual patient level. Routine care patients whose prescribed medications were the best fit according to the model had better cognitive performance after 2 years.

Natural language processing in clinical neuroscience and psychiatry: A review

Natural language processing (NLP) is rapidly becoming an important topic in the medical community. The ability to automatically analyze any type of medical document could be the key factor to fully exploit the data it contains. Cutting-edge artificial intelligence (AI) architectures, particularly machine learning and deep learning, have begun to be applied to this topic and have yielded promising results. We conducted a literature search for 1,024 papers that used NLP technology in neuroscience and psychiatry from 2010 to early 2022. After a selection process, 115 papers were evaluated. Each publication was classified into one of three categories: information extraction, classification, and data inference. Automated understanding of clinical reports in electronic health records has the potential to improve healthcare delivery. Overall, the performance of NLP applications is high, with an average F1-score and AUC above 85%. We also derived a composite measure in the form of Z-scores to better compare the performance of NLP models and their different classes as a whole. No statistical differences were found in the unbiased comparison. Strong asymmetry between English and non-English models, difficulty in obtaining high-quality annotated data, and train biases causing low generalizability are the main limitations. This review suggests that NLP could be an effective tool to help clinicians gain insights from medical reports, clinical research forms, and more, making NLP an effective tool to improve the quality of healthcare services.

Play the Pain: A Digital Strategy for Play-Oriented Research and Action

The value of understanding patients' illness experience and social contexts for advancing medicine and clinical care is widely acknowledged. However, methodologies for rigorous and inclusive data gathering and integrative analysis of biomedical, cultural, and social factors are limited. In this paper, we propose a digital strategy for large-scale qualitative health research, using play (as a state of being, a communication mode or context, and a set of imaginative, expressive, and game-like activities) as a research method for recursive learning and action planning. Our proposal builds on Gregory Bateson's cybernetic approach to knowledge production. Using chronic pain as an example, we show how pragmatic, structural and cultural constraints that define the relationship of patients to the healthcare system can give rise to conflicted messaging that impedes inclusive health research. We then review existing literature to illustrate how different types of play including games, chatbots, virtual worlds, and creative art making can contribute to research in chronic pain. Inspired by Frederick Steier's application of Bateson's theory to designing a science museum, we propose DiSPORA (Digital Strategy for Play-Oriented Research and Action), a virtual citizen science laboratory which provides a framework for delivering health information, tools for play-based experimentation, and data collection capacity, but is flexible in allowing participants to choose the mode and the extent of their interaction. Combined with other data management platforms used in epidemiological studies of neuropsychiatric illness, DiSPORA offers a tool for large-scale qualitative research, digital phenotyping, and advancing personalized medicine.

A Computational Framework to Analyze the Associations Between Symptoms and Cancer Patient Attributes Post Chemotherapy Using EHR Data

Patients with cancer, such as breast and colorectal cancer, often experience different symptoms post-chemotherapy. The symptoms could be fatigue, gastrointestinal (nausea, vomiting, lack of appetite), psychoneurological symptoms (depressive symptoms, anxiety), or other types. Previous research focused on understanding the symptoms using survey data. In this research, we propose to utilize the data within the Electronic Health Record (EHR). A computational framework is developed to use a natural language processing (NLP) pipeline to extract the clinician-documented symptoms from clinical notes. Then, a patient clustering method is based on the symptom severity levels to group the patient in clusters. The association rule mining is used to analyze the associations between symptoms and patient attributes (smoking history, number of comorbidities, diabetes status, age at diagnosis) in the patient clusters. The results show that the various symptom types and severity levels have different associations between breast and colorectal cancers and different timeframes post-chemotherapy. The results also show that patients with breast or colorectal cancers, who smoke and have severe fatigue, likely have severe gastrointestinal symptoms six months after the chemotherapy. Our framework can be generalized to analyze symptoms or symptom clusters of other chronic diseases where symptom management is critical.

Med7: A transferable clinical natural language processing model for electronic health records

Electronic health record systems are ubiquitous and the majority of patients' data are now being collected electronically in the form of free text. Deep learning has significantly advanced the field of natural language processing and the self-supervised representation learning and the transfer learning have become the methods of choice in particular when the high quality annotated data are limited. Identification of medical concepts and information extraction is a challenging task, yet important ingredient for parsing unstructured data into structured and tabulated format for downstream analytical tasks. In this work we introduced a named-entity recognition (NER) model for clinical natural language processing. The model is trained to recognise seven categories: drug names, route of administration, frequency, dosage, strength, form, duration. The model was first pre-trained on the task of predicting the next word, using a collection of 2 million free-text patients' records from MIMIC-III corpora followed by fine-tuning on the named-entity recognition task. The model achieved a micro-averaged F1 score of 0.957 across all seven categories. Additionally, we evaluated the transferability of the developed model using the data from the Intensive Care Unit in the US to secondary care mental health records (CRIS) in the UK. A direct application of the trained NER model to CRIS data resulted in reduced performance of F1 = 0.762, however after fine-tuning on a small sample from CRIS, the model achieved a reasonable performance of F1 = 0.944. This demonstrated that despite a close similarity between the data sets and the NER tasks, it is essential to fine-tune the target domain data in order to achieve more accurate results. The resulting model and the pre-trained embeddings are available at https://github.com/kormilitzin/med7.

The role of machine learning in clinical research: transforming the future of evidence generation

Background

Interest in the application of machine learning (ML) to the design, conduct, and analysis of clinical trials has grown, but the evidence base for such applications has not been surveyed. This manuscript reviews the proceedings of a multi-stakeholder conference to discuss the current and future state of ML for clinical research. Key areas of clinical trial methodology in which ML holds particular promise and priority areas for further investigation are presented alongside a narrative review of evidence supporting the use of ML across the clinical trial spectrum.

Results

Conference attendees included stakeholders, such as biomedical and ML researchers, representatives from the US Food and Drug Administration (FDA), artificial intelligence technology and data analytics companies, non-profit organizations, patient advocacy groups, and pharmaceutical companies. ML contributions to clinical research were highlighted in the pre-trial phase, cohort selection and participant management, and data collection and analysis. A particular focus was paid to the operational and philosophical barriers to ML in clinical research. Peer-reviewed evidence was noted to be lacking in several areas.

Conclusions

ML holds great promise for improving the efficiency and quality of clinical research, but substantial barriers remain, the surmounting of which will require addressing significant gaps in evidence.

Analyzing the symptoms in colorectal and breast cancer patients with or without type 2 diabetes using EHR data

This research extracted patient-reported symptoms from free-text EHR notes of colorectal and breast cancer patients and studied the correlation of the symptoms with comorbid type 2 diabetes, race, and smoking status. An NLP framework was developed first to use UMLS MetaMap to extract all symptom terms from the 366,398 EHR clinical notes of 1694 colorectal cancer (CRC) patients and 3458 breast cancer (BC) patients. Semantic analysis and clustering algorithms were then developed to categorize all the relevant symptoms into eight symptom clusters defined by seed terms. After all the relevant symptoms were extracted from the EHR clinical notes, the frequency of the symptoms reported from colorectal cancer (CRC) and breast cancer (BC) patients over three time-periods post-chemotherapy was calculated. Logistic regression (LR) was performed with each symptom cluster as the response variable while controlling for diabetes, race, and smoking status. The results show that the CRC and BC patients with Type 2 Diabetes (T2D) were more likely to report symptoms than CRC and BC without T2D over three time-periods in the cancer trajectory. We also found that current smokers were more likely to report anxiety (CRC, BC), neuropathic symptoms (CRC, BC), anxiety (BC), and depression (BC) than non-smokers.

Effectiveness of a Smartphone App With a Wearable Activity Tracker in Preventing the Recurrence of Mood Disorders: Prospective Case-Control Study

BACKGROUND

Smartphones and wearable devices can be used to obtain diverse daily log data related to circadian rhythms. For patients with mood disorders, giving feedback via a smartphone app with appropriate behavioral correction guides could play an important therapeutic role in the real world.

OBJECTIVE

We aimed to evaluate the effectiveness of a smartphone app named Circadian Rhythm for Mood (CRM), which was developed to prevent mood episodes based on a machine learning algorithm that uses passive digital phenotype data of circadian rhythm behaviors obtained with a wearable activity tracker. The feedback intervention for the CRM app consisted of a trend report of mood prediction, H-score feedback with behavioral guidance, and an alert system triggered when trending toward a high-risk state.

METHODS

In total, 73 patients with a major mood disorder were recruited and allocated in a nonrandomized fashion into 2 groups: the CRM group (14 patients) and the non-CRM group (59 patients). After the data qualification process, 10 subjects in the CRM group and 33 subjects in the non-CRM group were evaluated over 12 months. Both groups were treated in a similar manner. Patients took their usual medications, wore a wrist-worn activity tracker, and checked their eMoodChart daily. Patients in the CRM group were provided with daily feedback on their mood prediction and health scores based on the algorithm. For the CRM group, warning alerts were given when irregular life patterns were observed. However, these alerts were not given to patients in the non-CRM group. Every 3 months, mood episodes that had occurred in the previous 3 months were assessed based on the completed daily eMoodChart for both groups. The clinical course and prognosis, including mood episodes, were evaluated via face-to-face interviews based on the completed daily eMoodChart. For a 1-year prospective period, the number and duration of mood episodes were compared between the CRM and non-CRM groups using a generalized linear model.

RESULTS

The CRM group had 96.7% fewer total depressive episodes (n/year; exp β=0.033, P=.03), 99.5% shorter depressive episodes (total; exp β=0.005, P<.001), 96.1% shorter manic or hypomanic episodes (exp β=0.039, P<.001), 97.4% fewer total mood episodes (exp β=0.026, P=.008), and 98.9% shorter mood episodes (total; exp β=0.011, P<.001) than the non-CRM group. Positive changes in health behaviors due to the alerts and in wearable device adherence rates were observed in the CRM group.

CONCLUSIONS

The CRM app with a wearable activity tracker was found to be effective in preventing and reducing the recurrence of mood disorders, improving prognosis, and promoting better health behaviors. Patients appeared to develop a regular habit of using the CRM app.

TRIAL REGISTRATION

ClinicalTrials.gov NCT03088657; https://clinicaltrials.gov/ct2/show/NCT03088657.

A Big Data Platform for Real Time Analysis of Signs of Depression in Social Media

In this paper we propose a scalable platform for real-time processing of Social Media data. The platform ingests huge amounts of contents, such as Social Media posts or comments, and can support Public Health surveillance tasks. The processing and analytical needs of multiple screening tasks can easily be handled by incorporating user-defined execution graphs. The design is modular and supports different processing elements, such as crawlers to extract relevant contents or classifiers to categorise Social Media. We describe here an implementation of a use case built on the platform that monitors Social Media users and detects early signs of depression.

Maximizing the use of social and behavioural information from secondary care mental health electronic health records

PURPOSE

The contribution of social and behavioural factors in the development of mental health conditions and treatment effectiveness is widely supported, yet there are weak population level data sources on social and behavioural determinants of mental health. Enriching these data gaps will be crucial to accelerating precision medicine. Some have suggested the broader use of electronic health records (EHR) as a source of non-clinical determinants, although social and behavioural information are not systematically collected metrics in EHRs, internationally.

OBJECTIVE

In this commentary, we highlight the nature and quality of key available structured and unstructured social and behavioural data using a case example of value counts from secondary mental health data available in the UK from the UK Clinical Record Interactive Search (CRIS) database; highlight the methodological challenges in the use of such data; and possible solutions and opportunities involving the use of natural language processing (NLP) of unstructured EHR text.

CONCLUSIONS

Most structured non-clinical data fields within secondary care mental health EHR data have too much missing data for adequate use. The utility of other non-clinical fields reported semi-consistently (e.g., ethnicity and marital status) is entirely dependent on treating them appropriately in analyses, quantifying the many reasons behind missingness in consideration of selection biases. Advancements in NLP offer new opportunities in the exploitation of unstructured text from secondary care EHR data particularly given that clinical notes and attachments are available in large volumes of patients and are more routinely completed by clinicians. Tackling ways to re-use, harmonize, and improve our existing and future secondary care mental health data, leveraging advanced analytics such as NLP is worth the effort in an attempt to fill the data gap on social and behavioural contributors to mental health conditions and will be necessary to fulfill all of the domains needed to inform personalized interventions.

Predicting treatment effects in unipolar depression: A meta-review

There is increasing interest in clinical prediction models in psychiatry, which focus on developing multivariate algorithms to guide personalized diagnostic or management decisions. The main target of these models is the prediction of treatment response to different antidepressant therapies. This is because the ability to predict response based on patients' personal data may allow clinicians to make improved treatment decisions, and to provide more efficacious or more tolerable medications to the right patient. We searched the literature for systematic reviews about treatment prediction in the context of existing treatment modalities for adult unipolar depression, until July 2019. Treatment effect is defined broadly to include efficacy, safety, tolerability and acceptability outcomes. We first focused on the identification of individual predictor variables that might predict treatment response, and second, we considered multivariate clinical prediction models. Our meta-review included a total of 10 systematic reviews; seven (from 2014 to 2018) focusing on individual predictor variables and three focusing on clinical prediction models. These identified a number of sociodemographic, phenomenological, clinical, neuroimaging, remote monitoring, genetic and serum marker variables as possible predictor variables for treatment response, alongside statistical and machine-learning approaches to clinical prediction model development. Effect sizes for individual predictor variables were generally small and clinical prediction models had generally not been validated in external populations. There is a need for rigorous model validation in large external data-sets to prove the clinical utility of models. We also discuss potential future avenues in the field of personalized psychiatry, particularly the combination of multiple sources of data and the emerging field of artificial intelligence and digital mental health to identify new individual predictor variables.

Identifying Predictors of Suicide in Severe Mental Illness: A Feasibility Study of a Clinical Prediction Rule (Oxford Mental Illness and Suicide Tool or OxMIS)

BACKGROUND

Oxford Mental Illness and Suicide tool (OxMIS) is a brief, scalable, freely available, structured risk assessment tool to assess suicide risk in patients with severe mental illness (schizophrenia-spectrum disorders or bipolar disorder). OxMIS requires further external validation, but a lack of large-scale cohorts with relevant variables makes this challenging. Electronic health records provide possible data sources for external validation of risk prediction tools. However, they contain large amounts of information within free-text that is not readily extractable. In this study, we examined the feasibility of identifying suicide predictors needed to validate OxMIS in routinely collected electronic health records.

METHODS

In study 1, we manually reviewed electronic health records of 57 patients with severe mental illness to calculate OxMIS risk scores. In study 2, we examined the feasibility of using natural language processing to scale up this process. We used anonymized free-text documents from the Clinical Record Interactive Search database to train a named entity recognition model, a machine learning technique which recognizes concepts in free-text. The model identified eight concepts relevant for suicide risk assessment: medication (antidepressant/antipsychotic treatment), violence, education, self-harm, benefits receipt, drug/alcohol use disorder, suicide, and psychiatric admission. We assessed model performance in terms of precision (similar to positive predictive value), recall (similar to sensitivity) and F1 statistic (an overall performance measure).

RESULTS

In study 1, we estimated suicide risk for all patients using the OxMIS calculator, giving a range of 12 month risk estimates from 0.1-3.4%. For 13 out of 17 predictors, there was no missing information in electronic health records. For the remaining 4 predictors missingness ranged from 7-26%; to account for these missing variables, it was possible for OxMIS to estimate suicide risk using a range of scores. In study 2, the named entity recognition model had an overall precision of 0.77, recall of 0.90 and F1 score of 0.83. The concept with the best precision and recall was medication (precision 0.84, recall 0.96) and the weakest were suicide (precision 0.37), and drug/alcohol use disorder (recall 0.61).

CONCLUSIONS

It is feasible to estimate suicide risk with the OxMIS tool using predictors identified in routine clinical records. Predictors could be extracted using natural language processing. However, electronic health records differ from other data sources, particularly for family history variables, which creates methodological challenges.