Natural language processing systems for extracting information from electronic health records about activities of daily living. A systematic review

JAMIA Open, Год журнала: 2024, Номер 7(2)

Опубликована: Апрель 8, 2024

Abstract Objective Natural language processing (NLP) can enhance research on activities of daily living (ADL) by extracting structured information from unstructured electronic health records (EHRs) notes. This review aims to give insight into the state-of-the-art, usability, and performance NLP systems extract ADL EHRs. Materials Methods A systematic was conducted based searches in Pubmed, Embase, Cinahl, Web Science, Scopus. Studies published between 2017 2022 were selected predefined eligibility criteria. Results The identified 22 studies. Most studies (65%) used for classifying EHR data 1 or 2 ADL. Deep learning, combined with a ruled-based method machine approach most commonly used. varied widely terms pre-processing algorithms. Common evaluation methods cross-validation train/test datasets, F1, precision, sensitivity as frequently reported metrics. relativity high overall scores Discussion are valuable extraction However, comparing is difficult due diversity challenges related dataset, including restricted access data, inadequate documentation, lack granularity, small datasets. Conclusion indicates that promising deriving what best-performing system is, depends characteristics question, type

Язык: Английский

---

Health Care Language Models and Their Fine-Tuning for Information Extraction: Scoping Review

JMIR Medical Informatics, Год журнала: 2024, Номер 12, С. e60164 - e60164

Опубликована: Окт. 21, 2024

Background In response to the intricate language, specialized terminology outside everyday life, and frequent presence of abbreviations acronyms inherent in health care text data, domain adaptation techniques have emerged as crucial transformer-based models. This refinement knowledge language models (LMs) allows for a better understanding medical textual which results an improvement downstream tasks, such information extraction (IE). We identified gap literature regarding LMs. Therefore, this study presents scoping review investigating methods transformers care, differentiating between English non-English languages, focusing on Portuguese. Most specifically, we investigated development LMs, with aim comparing Portuguese other more developed languages guide path non–English-language fewer resources. Objective aimed research IE models, regardless understand efficacy what are entities most commonly extracted. Methods was conducted using PRISMA-ScR (Preferred Reporting Items Systematic reviews Meta-Analyses extension Scoping Reviews) methodology Scopus Web Science Core Collection databases. Only studies that mentioned creation LMs or were included, while large (LLMs) excluded. The latest not included since wanted LLMs, architecturally different distinct purposes. Results Our search query retrieved 137 studies, 60 met inclusion criteria, none them systematic reviews. Chinese developed. These already disease-specific others only general–health European does any public LM should take examples from develop, first, general-health then, advanced phase, Regarding used method, named entity recognition popular topic, few mentioning Assertion Status addressing lexical problems. extracted diagnosis, posology, symptoms. Conclusions findings indicate is beneficial, achieving tasks. analysis allowed us use languages. lacks relevant draw develop these drive progress AI. Health professionals could benefit highlighting medically optimizing reading be create patient timelines, allowing profiling.

Язык: Английский

---

Clinical applications of deep learning in neuroinflammatory diseases: A scoping review

Revue Neurologique, Год журнала: 2024, Номер unknown

Опубликована: Май 1, 2024

Deep learning (DL) is an artificial intelligence technology that has aroused much excitement for predictive medicine due to its ability process raw data modalities such as images, text, and time series of signals. Here, we intend give the clinical reader elements understand this technology, taking neuroinflammatory diseases illustrative use case translation efforts. We reviewed scope rapidly evolving field get quantitative insights about which applications concentrate efforts are most commonly used. queried PubMed database articles reporting DL algorithms in radiology.healthairegister.com website commercial algorithms. The review included 148 published between 2018 2024 five could be grouped computer-aided diagnosis, individual prognosis, functional assessment, segmentation radiological structures, optimization acquisition. Our highlighted important discrepancies structures diagnosis currently with overrepresentation imaging. Various model architectures have addressed different applications, relatively low volume data, diverse modalities. report high-level technical characteristics synthesize narratively applications. Predictive performances some common a priori on topic finally discussed. reported position information processing enhancing existing paraclinical investigations bringing perspectives make innovative ones actionable healthcare.

Язык: Английский

---

Direct Clinical Applications of Natural Language Processing in Common Neurological Disorders: Scoping Review

JMIR Neurotechnology, Год журнала: 2024, Номер 3, С. e51822 - e51822

Опубликована: Май 22, 2024

Background Natural language processing (NLP), a branch of artificial intelligence that analyzes unstructured language, is being increasingly used in health care. However, the extent to which NLP has been formally studied neurological disorders remains unclear. Objective We sought characterize studies applied diagnosis, prediction, or treatment common disorders. Methods This review followed PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension Scoping Reviews) standards. The search was conducted using MEDLINE Embase on May 11, 2022. Studies use migraine, Parkinson disease, Alzheimer stroke transient ischemic attack, epilepsy, multiple sclerosis were included. excluded conference abstracts, papers, as well involving heterogeneous clinical populations indirect uses NLP. Study characteristics extracted analyzed descriptive statistics. did not aggregate measurements performance our due high variability study outcomes, main limitation study. Results In total, 916 identified, 41 (4.5%) met all eligibility criteria included final review. Of studies, most frequently represented attack (n=20, 49%), by epilepsy (n=10, 24%), disease (n=6, 15%), (n=5, 12%). found no migraine criteria. objective diagnosis phenotyping (n=17, 41%), prognostication (n=9, 22%), (n=4, 10%). 18 (44%) only machine learning approaches, 6 (15%) rule-based methods, 17 (41%) both. Conclusions commonly implying potential role augmenting diagnostic accuracy settings with limited access expertise. also several gaps research, few addressing certain disorders, may suggest additional areas inquiry. Trial Registration Prospective Register (PROSPERO) CRD42021228703; https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=228703

Язык: Английский

---

Common clinical blood and urine biomarkers for ischemic stroke: an Estonian Electronic Health Records database study

European journal of medical research, Год журнала: 2023, Номер 28(1)

Опубликована: Март 25, 2023

Ischemic stroke (IS) is a major health risk without generally usable effective measures of primary prevention. Early warning signals that are easy to detect and widely available can save lives. Estonia has one nation-wide Electronic Health Record (EHR) database for the storage medical information patients from hospitals care providers.We extracted structured unstructured data EHRs participants Estonian Biobank (EstBB) evaluated different formats input understand how this continuously growing dataset should be prepared best prediction. The utility EHR finding blood- urine-based biomarkers IS was demonstrated by applying analytical machine learning (ML) methods.Several early trends in common clinical laboratory parameter changes (set red blood indices, lymphocyte/neutrophil ratio, etc.) were established developed ML models predicted future occurrence with very high accuracy Random Forests proved as most applicable method data.We conclude factors uncovered valuable resources screening population well constructing disease scores refining prediction ML.

Язык: Английский

---

It’s time to change our documentation philosophy: writing better neurology notes without the burnout

Frontiers in Digital Health, Год журнала: 2022, Номер 4

Опубликована: Ноя. 28, 2022

Succinct clinical documentation is vital to effective twenty-first-century healthcare. Recent changes in outpatient and inpatient evaluation management (E/M) guidelines have allowed neurology practices make that reduce the burden enhance note usability. Despite favorable E/M guidelines, some not moved quickly change their philosophy. We argue favor of design, structure, implementation notes them shorter yet still information-rich. A move from physician-centric team can work for physicians. Changing philosophy "bigger better" "short but sweet" burden, streamline writing reading notes, utility medical decision-making, patient education, research. believe these favorably affect physician well-being without adversely affecting reimbursement.

Язык: Английский

---

Subtypes of relapsing-remitting multiple sclerosis identified by network analysis

Frontiers in Digital Health, Год журнала: 2023, Номер 4

Опубликована: Янв. 11, 2023

We used network analysis to identify subtypes of relapsing-remitting multiple sclerosis subjects based on their cumulative signs and symptoms. The electronic medical records 113 with were reviewed, symptoms mapped classes in a neuro-ontology, collapsed into sixteen superclasses by subsumption. After normalization vectorization the data, bipartite (subject-feature) unipartite (subject-subject) graphs created using NetworkX visualized Gephi. Degree weighted degree calculated for each node. Graphs partitioned communities modularity score. Feature maps differences features community. Network graph yielded higher score (0.49) than (0.25). was five which named fatigue, behavioral, hypertonia/weakness, abnormal gait/sphincter, sensory, feature characteristics. pain, cognitive, gait/weakness/hypertonia features. Although we did not pure (e.g., motor, etc.) this cohort subjects, demonstrated that could partition these different subtype communities. Larger datasets additional partitioning algorithms are needed confirm findings elucidate significance. This study contributes literature investigating combining reduction subsumption analysis.

Язык: Английский

---

Health Care Language Models and Their Fine-Tuning for Information Extraction: Scoping Review (Preprint)

Опубликована: Май 3, 2024

In response to the intricate language, specialized terminology outside everyday life, and frequent presence of abbreviations acronyms inherent in health care text data, domain adaptation techniques have emerged as crucial transformer-based models. This refinement knowledge language models (LMs) allows for a better understanding medical textual which results an improvement downstream tasks, such information extraction (IE). We identified gap literature regarding LMs. Therefore, this study presents scoping review investigating methods transformers care, differentiating between English non-English languages, focusing on Portuguese. Most specifically, we investigated development LMs, with aim comparing Portuguese other more developed languages guide path non–English-language fewer resources. aimed research IE models, regardless understand efficacy what are entities most commonly extracted. was conducted using PRISMA-ScR (Preferred Reporting Items Systematic reviews Meta-Analyses extension Scoping Reviews) methodology Scopus Web Science Core Collection databases. Only studies that mentioned creation LMs or were included, while large (LLMs) excluded. The latest not included since wanted LLMs, architecturally different distinct purposes. Our search query retrieved 137 studies, 60 met inclusion criteria, none them systematic reviews. Chinese developed. These already disease-specific others only general–health European does any public LM should take examples from develop, first, general-health then, advanced phase, Regarding used method, named entity recognition popular topic, few mentioning Assertion Status addressing lexical problems. extracted diagnosis, posology, symptoms. findings indicate is beneficial, achieving tasks. analysis allowed us use languages. lacks relevant draw develop these drive progress AI. Health professionals could benefit highlighting medically optimizing reading be create patient timelines, allowing profiling.

Язык: Английский

---

Clinical and Research Applications of the Electronic Medical Record in Multiple Sclerosis: A Narrative Review of Current Uses and Future Applications

International Journal of MS Care, Год журнала: 2022, Номер 24(6), С. 287 - 294

Опубликована: Ноя. 1, 2022

CE INFORMATION ACTIVITY AVAILABLE ONLINE: To access the article and evaluation online, go to https://www.highmarksce.com/mscare. TARGET AUDIENCE: The target audience for this activity is physicians, advanced practice clinicians, nursing professionals, pharmacists, mental health social workers, other care providers involved in research management of patients with multiple sclerosis (MS). LEARNING OBJECTIVES: Characterize existing EMR platforms designed specifically people MS. Describe relevant variables that are captured allow identification EMR-based cohorts ACCREDITATION: In support improving patient care, has been planned implemented by Consortium Multiple Sclerosis Centers (CMSC) Intellisphere, LLC. CMSC jointly accredited Accreditation Council Continuing Medical Education (ACCME), Pharmacy (ACPE), American Nurses Credentialing Center (ANCC), provide continuing education healthcare team. This was team, learners will receive .5 Interprofessional (IPCE) credit learning change. PHYSICIANS: Physicians: designates journal-based a maximum AMA PRA Category 1 Credit(s)™. Physicians should claim only commensurate extent their participation activity. NURSES: enduring material contact hour professional development (NCPD) (none area pharmacology). PHARMACISTS: knowledge-based (UAN JA4008165-9999-22-033-H01-P) qualifies (.5) (.05 CEUs) pharmacy credit. PSYCHOLOGISTS: awarded 0.5 credits. SOCIAL WORKERS: As Jointly Accredited Organization, approved offer work Association Social Work Boards (ASWB) Approved (ACE) program. Organizations, not individual courses, under State provincial regulatory boards have final authority determine whether an course may be accepted maintains responsibility course. workers completing DISCLOSURES: It policy mitigate all financial disclosures from planners, faculty, persons can affect content For activity, mitigated. Francois Bethoux, MD, editor chief International Journal MS Care (IJMSC), served as physician planner He disclosed no relationships. Alissa Mary Willis, associate IJMSC, Authors Carol Swetlik, Riley Bove, Marisa McGinley, DO, staff at CMSC, LLC who position influence Laurie Scudder, DNP, NP, director reviewer She METHOD OF PARTICIPATION: Release Date: November 1, 2022; Valid Credit through: 2023. order credit, participants must: 1) Review information, including objectives author disclosures.2) Study educational content.3) Complete evaluation, which available Statements upon successful completion evaluation. There fee participate DISCLOSURE UNLABELED USE: contain discussion published and/or investigational uses agents FDA. do recommend use any agent outside labeled indications. opinions expressed those faculty necessarily represent views or DISCLAIMER: Participants implied newly acquired information enhance outcomes own development. presented meant serve guideline management. Any medications, diagnostic procedures, treatments discussed publication used clinicians professionals without first evaluating patients’ conditions, considering possible contraindications risks, reviewing applicable manufacturer’s product comparing therapeutic approach recommendations authorities.

Язык: Английский

---

Claims-based algorithm to estimate the Expanded Disability Status Scale for multiple sclerosis in a German health insurance fund: a validation study using patient medical records

Frontiers in Neurology, Год журнала: 2023, Номер 14

Опубликована: Дек. 7, 2023

Background The Expanded Disability Status Scale (EDSS) quantifies disability and measures disease progression in multiple sclerosis (MS), however is not available administrative claims databases. Objectives To develop a claims-based algorithm for deriving EDSS validate it against clinical dataset capturing true values from medical records. Methods We built unique linked combining data the German AOK PLUS sickness fund records Multiple Sclerosis Management System 3D (MSDS ). Data were deterministically based on insurance numbers. used 69 MS-related diagnostic indicators recorded with ICD-10-GM codes within 3 months before after to estimate proxy (pEDSS). Predictive performance of pEDSS was assessed as an eight-fold (EDSS 1.0–7.0, ≥8.0), three-fold 1.0–3.0, 4.0–5.0, ≥6.0), binary classifier <6.0, ≥6.0). For each classifier, predictive determined, overall summarized using macro F1-score. Finally, we implemented determine among cohort patients MS PLUS, who alive insured 12 prior index diagnosis. Results recruited 100 people by had ≥1 measure MSDS between 01/10/2015 30/06/2019 (620 measurements overall). Patients mean rescaled 3.2 3.0. deviated 1.2 points, resulting squared error prediction 2.6. F1-score 0.25 indicated low performance. Broader severity groupings better performing, classifiers severe achieving 0.68 0.84, respectively. In (3,756 patients, 71.9% female, 51.9 years), older progressive forms those higher comorbidity burden showed pEDSS. Conclusion Generally, underestimated mild-to-moderate symptoms poorly captured across all functional systems. While proxy-based approach may allow granular description disability, broader show good

Язык: Английский

---