Improving the Identification of Diabetic Retinopathy and Related Conditions in the Electronic Health Record Using Natural Language Processing Methods.

Purpose: To compare the performance of 3 phenotyping methods in identifying diabetic retinopathy (DR) and related clinical conditions. Design: Three phenotyping methods were used to identify clinical conditions including unspecified DR, nonproliferative DR (NPDR) (mild, moderate, severe), consolidated NPDR (unspecified DR or any NPDR), proliferative DR, diabetic macular edema (DME), vitreous hemorrhage, retinal detachment (RD) (tractional RD or combined tractional and rhegmatogenous RD), and neovascular glaucoma (NVG). The first method used only International Classification of Diseases, 10th Revision (ICD-10) diagnosis codes (ICD-10 Lookup System). The next 2 methods used a Bidirectional Encoder Representations from Transformers with a dense Multilayer Perceptron output layer natural language processing (NLP) framework. The NLP framework was applied either to free-text of provider notes (Text-Only NLP System) or both free-text and ICD-10 diagnosis codes (Text-and-International Classification of Diseases [ICD] NLP System). Subjects: Adults ≥18 years with diabetes mellitus seen at the Wilmer Eye Institute. Methods: We compared the performance of the 3 phenotyping methods in identifying the DR related conditions with gold standard chart review. We also compared the estimated disease prevalence using each method. Main Outcome Measures: Performance of each method was reported as the macro F1 score. The agreement between the methods was calculated using the kappa statistic. Prevalence estimates were also calculated for each method. Results: A total of 91 097 patients and 692 486 office visits were included in the study. Compared with the gold standard, the Text-and-ICD NLP System had the highest F1 score for most clinical conditions (range 0.39-0.64). The agreement between the ICD-10 Lookup System and Text-Only NLP System varied (kappa of 0.21-0.81). The prevalence of DR and related conditions ranged from 1.1% for NVG to 17.9% for DME (using the Text-and-ICD NLP System). Conclusions: The prevalence of DR and related conditions varied significantly depending on the methodology of identifying cases. The best performing phenotyping method was the Text-and-ICD NLP System that used information in both diagnosis codes as well as free-text notes. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Comparison of Diagnosis Codes to Clinical Notes in Classifying Patients with Diabetic Retinopathy.

Purpose: Electronic health records (EHRs) contain a vast amount of clinical data. Improved automated classification approaches have the potential to accurately and efficiently identify patient cohorts for research. We evaluated if a rule-based natural language processing (NLP) algorithm using clinical notes performed better for classifying proliferative diabetic retinopathy (PDR) and nonproliferative diabetic retinopathy (NPDR) severity compared with International Classification of Diseases, ninth edition (ICD-9) or 10th edition (ICD-10) codes. Design: Cross-sectional study. Subjects: Deidentified EHR data from an academic medical center identified 2366 patients aged ≥18 years, with diabetes mellitus, diabetic retinopathy (DR), and available clinical notes. Methods: From these 2366 patients, 306 random patients (100 training set, 206 test set) underwent chart review by ophthalmologists to establish the gold standard. International Classification of Diseases codes were extracted from the EHR. The notes algorithm identified positive mention of PDR and NPDR severity from clinical notes. Proliferative diabetic retinopathy and NPDR severity classification by ICD codes and the notes algorithm were compared with the gold standard. The entire DR cohort (N = 2366) was then classified as having presence (or absence) of PDR using ICD codes and the notes algorithm. Main Outcome Measures: Sensitivity, specificity, positive predictive value (PPV), negative predictive value, and F1 score for the notes algorithm compared with ICD codes using a gold standard of chart review. Results: For PDR classification of the test set patients, the notes algorithm performed better than ICD codes for all metrics. Specifically, the notes algorithm had significantly higher sensitivity (90.5% [95% confidence interval 85.7, 94.9] vs. 68.4% [60.4, 75.3]), but similar PPV (98.0% [95.4-100] vs. 94.7% [90.3, 98.3]) respectively. The F1 score was 0.941 [0.910, 0.966] for the notes algorithm compared with 0.794 [0.734, 0.842] for ICD codes. For PDR classification, ICD-10 codes performed better than ICD-9 codes (F1 score 0.836 [0.771, 0.878] vs. 0.596 [0.222, 0.692]). For NPDR severity classification, the notes algorithm performed similarly to ICD codes, but performance was limited by small sample size. Conclusions: The notes algorithm outperformed ICD codes for PDR classification. The findings demonstrate the significant potential of applying a rule-based NLP algorithm to clinical notes to increase the efficiency and accuracy of cohort selection for research. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Establishing a Method of Systematic and Reliable Analysis of Psychodynamic Process Notes.

Introduction: Process notes contain unique information concerning core elements of a psychodynamic treatment. These elements may be both conscious and unconscious for the author. One element for study is the tendency to which a therapist writes about providing either supportive or expressive interventions. This study sought to establish a method of systematically and reliably identifying the records of therapists' interventions as supportive or expressive. Methods: Three early-career clinicians were trained in the use of a process note intervention rating scale constructed specifically for this study. Quantitative statistical analyses assessed the scale's reliability and internal consistency. Results: Interrater reliability analysis determined at a p of 0.005 a Fleiss's kappa of 0.24 and an intraclass correlation coefficient of 0.264, suggesting a low but statistically significant reliability between the raters. A Cronbach's alpha of 0.67 and a McDonald's omega of 0.53 suggested questionable internal consistency. Discussion: Early-career clinicians can reliably code the manifestations of interventions in psychodynamic process notes as supportive or expressive. Future studies may improve the reliability and internal consistency of the scale, add measures of interpretation content, and evaluate these data in relation to other core elements of process notes, such as the author's emotional engagement as manifested in language measures and clinical outcome.

Characterizing Myasthenia Gravis Symptoms, Exacerbations, and Crises From Neurologist's Clinical Notes Using Natural Language Processing.

Background Myasthenia gravis (MG) is a rare, autoantibody neuromuscular disorder characterized by fatigable weakness. Real-world evidence based on administrative and structured datasets regarding MG may miss important details related to the clinical encounter. Examination of free-text clinical progress notes has the potential to illuminate aspects of MG care. Objective The primary objective was to examine and characterize neurologist progress notes in the care of individuals with MG regarding the prevalence of documentation of clinical subtypes, antibody status, symptomatology, and MG deteriorations, including exacerbations and crises. The secondary objectives were to categorize MG deteriorations into practical, objective states as well as examine potential sources of clinical inertia in MG care. Methods We performed a retrospective, cross-sectional analysis of de-identified neurologist clinical notes from 2017 to 2022. A qualitative analysis of physician descriptions of MG deteriorations and a discussion of risks in MG care (risk for adverse effects, risk for clinical decompensation, etc.) was performed. Results Of the 3,085 individuals with MG, clinical subtypes and antibody status identified included gMG (n = 400; 13.0%), ocular MG (n = 253; 8.2%), MG unspecified (2,432; 78.8%), seropositivity for acetylcholine receptor antibody (n = 441; 14.3%), and MuSK antibody (n = 29; 0.9%). The most common gMG manifestations were dysphagia (n = 712; 23.0%), dyspnea (n = 626; 20.3%), and dysarthria (n = 514; 16.7%). In MG crisis patients, documentation of difficulties with MG standard therapies was common (n = 62; 45.2%). The qualitative analysis of MG deterioration types includes symptom fluctuation, symptom worsening with treatment intensification, MG deterioration with rescue therapy, and MG crisis. Qualitative analysis of MG-related risks included the toxicity of new therapies and concern for worsening MG because of changing therapies. Conclusions This study of neurologist progress notes demonstrates the potential for real-world evidence generation in the care of individuals with MG. MG patients suffer fluctuating symptomatology and a spectrum of clinical deteriorations. Adverse effects of MG therapies are common, highlighting the need for effective, less toxic treatments.

A Study in Wad Madani, Sudan: Are We Documenting Operation Notes Effectively?

BACKGROUND: Operative notes represent the critical record of a surgical procedure, encompassing comprehensive details encountered throughout the operation. Recognizing the importance of comprehensive documentation, the Royal College of Surgeons (RCS) developed the Good Surgical Practice guidelines, which emphasize accurately recording every procedure and specifying the necessary parameters for each operative note. These guidelines help maintain high standards of surgical care and patient safety. METHODS: A retrospective review of 88 orthopaedic surgery operative notes for fracture neck of femurs was conducted at Gezira Centre for Orthopedic Surgery and Traumatology (GCOST) from March 12 to May 28, 2022. The review assessed 18 parameters against RCS guidelines. Statistical analysis was performed using Statistical Product and Service Solutions (SPSS, version 25.0; IBM SPSS Statistics for Windows, Armonk, NY), which facilitated comprehensive data examination. RESULTS: In 37 cases (42.05%), the operation notes were written by a medical officer. In 29 cases (32.95%), an orthopaedic resident authored the notes. A specialist documented the notes in 21 cases (23.86%), and a consultant wrote the notes in one case (1.14%). Over 90% of the notes included surgeon and assistant names, procedure names, operative diagnoses, operative procedures, prosthesis details, deep vein thrombosis (DVT) and antibiotic prophylaxis, and signatures. The name of the theatre anaesthetist, elective/emergency details, and additional procedures with reasons were absent in all notes. Less than 50% of the notes documented the time of the procedure, type of incision, operative findings, anticipated blood loss, closure technique specifics, and complications. CONCLUSION: The study emphasizes the shortcomings in the operating notes, underscoring the necessity for training initiatives to enhance the recording by medical officers and orthopaedic trainees. Implementing structured templates that adhere to RCS standards can improve the comprehensiveness and consistency of operating notes, effectively resolving existing discrepancies. Regular audits and feedback sessions are essential for identifying and rectifying persistent issues. It is recommended to arrange workshops and seminars to educate medical officials and trainees on the skills of efficient note-taking and thorough documentation procedures.

Assessing the Ability of a Large Language Model to Score Free-Text Medical Student Clinical Notes: Quantitative Study.

Background: Teaching medical students the skills required to acquire, interpret, apply, and communicate clinical information is an integral part of medical education. A crucial aspect of this process involves providing students with feedback regarding the quality of their free-text clinical notes. Objective: The goal of this study was to assess the ability of ChatGPT 3.5, a large language model, to score medical students' free-text history and physical notes. Methods: This is a single-institution, retrospective study. Standardized patients learned a prespecified clinical case and, acting as the patient, interacted with medical students. Each student wrote a free-text history and physical note of their interaction. The students' notes were scored independently by the standardized patients and ChatGPT using a prespecified scoring rubric that consisted of 85 case elements. The measure of accuracy was percent correct. Results: The study population consisted of 168 first-year medical students. There was a total of 14,280 scores. The ChatGPT incorrect scoring rate was 1.0%, and the standardized patient incorrect scoring rate was 7.2%. The ChatGPT error rate was 86%, lower than the standardized patient error rate. The ChatGPT mean incorrect scoring rate of 12 (SD 11) was significantly lower than the standardized patient mean incorrect scoring rate of 85 (SD 74; P=.002). Conclusions: ChatGPT demonstrated a significantly lower error rate compared to standardized patients. This is the first study to assess the ability of a generative pretrained transformer (GPT) program to score medical students' standardized patient-based free-text clinical notes. It is expected that, in the near future, large language models will provide real-time feedback to practicing physicians regarding their free-text notes. GPT artificial intelligence programs represent an important advance in medical education and medical practice.

Bridging the Gap: Can Large Language Models Match Human Expertise in Writing Neurosurgical Operative Notes?

BACKGROUND: Proper documentation is essential for patient care. The popularity of artificial intelligence (AI) offers the potential for improvements in neurosurgical note-writing. The study aimed to assess how AI can optimize documentation in neurosurgical procedures. METHODS: Thirty-six notes were included. All identifiable data were removed. Essential information, such as perioperative data and diagnosis, was sourced from these notes. ChatGPT 4.0 was trained to draft notes from surgical vignettes using each surgeon's note template. One hundred forty-four surveys, with a surgeon or AI note, were shared with three surgeons to evaluate accuracy, content, and organization using a five-point scale. Accuracy was the factual correctness. Content was the comprehensiveness. Organization was the arrangement of the note. Flesch-Kincaid Grade Level (FKGL) and Flesch Reading Ease (FRE) scores quantified each note's readability. RESULTS: The mean AI accuracy (4.44) was not different from the mean surgeon accuracy (4.33, p = 0.512). The mean AI content (3.73) was lower than the mean surgeon content (4.42, p < 0.001). The mean AI organization (4.54) was greater than the mean surgeon organization (4.24, p = 0.064). The mean AI note's FKGL (13.13) was greater than the mean surgeon FKGL (9.99, p <0.001). The mean AI FRE (21.42) was lower than the mean surgeon FRE (41.70, p <0.001). CONCLUSION: AI notes were on par with surgeon notes in accuracy and organization, but lacked in content. Additionally, AI notes utilized language at an advanced reading level. These findings underscore the potential for ChatGPT to enhance the efficiency of neurosurgery documentation.

Similarity Analysis of Clinical Notes for Revealing Writing Styles in an Electronic Medical Record System.

BACKGROUND: One of the disadvantages of medical documentation in Electronic Medical Record (EMR) systems is that records tend to be redundant by "copying and pasting", a writing style to duplicate and revise previous records. In this study, we analyzed the similarity between records to identify the factors affecting the writing style of clinical notes. METHOD: We analyzed 98,038 records of 4,149 patients from two years in the Department of Obstetricians and Gynecology at Kyoto University Hospital, Japan. We observed the correlation between the distribution of the record similarity and string amounts, as well as the disease codes and ratios of outpatient visit. RESULTS: The patient group with high record similarity and large number of strings was the group with reproductive medicine, followed by the group of malignant tumor follow-up or Women's Healthcare. DISCUSSION: In reproductive medicine, physicians have a demand for an overarching evaluation, and in follow-up malignancies or in Women's Healthcare, they have a demand to check for subtle differences from the last time. These facts along with our data insist that the writing style in EMR systems is related to the patient's status. CONCLUSION: We declared that the writing style in EMR systems is affected by the patient's status. The writing style of duplicating and revising is preferred (1) when there is a clinical demand for an overarching evaluation, and (2) when there is a clinical demand to check for subtle differences from the last time.

Insights from the Implementation of Open Notes in Sweden.

The implementation of Open Notes in Sweden, granting patients access to their clinical records, has been a complex and nuanced endeavor, marked by regional variations in strategy and challenges arising from the diverging needs of healthcare providers and patients. This paper presents an interview study with managers about the implementation process in five of the 21 regions in Sweden. The aim of this study is to explore the experiences and strategies of these managers in navigating the implementation challenges. The study sheds light on the prevalent theme of uncertainty throughout the implementation journey and the strategies used to balance conflicting perspectives. The findings contribute to our understanding of Open Notes implementation and offer policymakers and healthcare organizations insights about enhancing the implementation process to optimize patient care.

Accelerating Clinical Text Annotation in Underrepresented Languages: A Case Study on Text De-Identification.

Clinical notes contain valuable information for research and monitoring quality of care. Named Entity Recognition (NER) is the process for identifying relevant pieces of information such as diagnoses, treatments, side effects, etc., and bring them to a more structured form. Although recent advancements in deep learning have facilitated automated recognition, particularly in English, NER can still be challenging due to limited specialized training data. This exacerbated in hospital settings where annotations are costly to obtain without appropriate incentives and often dependent on local specificities. In this work, we study whether this annotation process can be effectively accelerated by combining two practical strategies. First, we convert usually passive annotation tasks into a proactive contest to motivate human annotators in performing a task often considered tedious and time-consuming. Second, we provide pre-annotations for the participants to evaluate how recall and precision of the pre-annotations can boost or deteriorate annotation performance. We applied both strategies to a text de-identification task on French clinical notes and discharge summaries at a large Swiss university hospital. Our results show that proactive contest and average quality pre-annotations can significantly speed up annotation time and increase annotation quality, enabling us to develop a text de-identification model for French clinical notes with high performance (F1 score 0.94).

Role of Intraoperative Documentation: Avoidable Ulnar Nerve Injury During Implant Removal.

Injuries to the ulnar nerve during open reduction and internal fixation of distal humerus fractures are a well-known phenomenon. However, ulnar nerve injury during implant removal has not been well documented. We performed implant removal in a united distal humerus fracture with the aim of improving the elbow's range of motion. Even with proper surgical precautions in place, the ulnar nerve was damaged during dissection. This report aims to provide insight into this rare phenomenon, and the reasons for this injury are examined retrospectively. The importance of operation notes, the surgical approach, anterior transposition of the nerve, and how this and other factors could have helped the surgeons avoid this complication have also been highlighted.

Healthcare Professionals' Documentation in Supported Accommodation for People with Profound Intellectual Disabilities: An Educational Intervention Study.

Good-quality relationships in which individuals with profound intellectual disabilities (intelligence quotient, IQ < 20-25) are recognized by healthcare professionals (HPs) are essential for the quality of healthcare and promoting autonomy. This study examines the impact of an educational intervention on documentation of the interplay between HP and individuals receiving services in supported accommodation in Norway. An educational intervention study was designed to encourage HPs to document their approaches and interplay. The Scale for the Evaluation of Staff-Patient Interactions in Progress Notes (SESPI) was applied to measure documentation before and after the intervention. Journal notes written over a three-month period before the intervention and a three-month period after the intervention were measured. Prior to the intervention, only 23.1% of the journal notes described the resident's experiences, increasing by 5.4% (p = 0.041) post-intervention. Practical solutions to individual experiences increased from 0.9% to 8.5% (p < 0.001). The educational intervention demonstrated a significant increase in the documentation of residents' experiences and the interplay between HPs and residents. Future research should explore the generalizability of these findings. Incomplete documentation of HPs' relational work conceals important aspects of the healthcare provided, potentially resulting in confining autonomy and participation for individuals with intellectual disabilities.

A comparison of written case notes and the delivery of care in dementia specialist mental health wards.

Introduction: Stigmatising language concerning people living with dementia can cause potentially harmful and dehumanising consequences. Language used about people living with dementia in mental health wards may focus on medical perspectives and suggest custodial relationships with patients rather than person-centred accounts of individuals. This language could have a devastating impact on the provision of person-centred care. This study investigated the relationship between accounts of people living with dementia written in healthcare case notes and clinical practice at three dementia specialist wards in Wales, UK. Language guidance was provided to ward staff to assess whether stigmatising language could be reduced and whether this influenced the provision of person-centred care.Methodology: Dementia Care Mapping was adapted to analyse case note entries for enhancing and detracting accounts of people living with dementia at three data collection points. These were compared to the results of routine DCM observations of care across the three wards. The healthcare case notes of 117 people living with dementia, encompassing 4, 522 entries over ten months were analysed. DCM observations of 38 people living with dementia within the three wards were compared against the case note results. Person-centred language guidance was shared with care staff following each data collection point.Results: Following the provision of person-centered language guidance, the use of personally enhancing language was observed to increase across all three wards. Non-person-centred case note entries predominantly focussed on Labelling language, whilst language concerning Invalidation and Objectification also occurred frequently compared to other DCM domains. Person centred language typically concerned Acknowledgement. A relationship between case note entries and practice was evident in some domains although findings were inconsistent.Discussion and Implications: The findings highlight the importance of addressing stigmatising language in healthcare and suggest that further studies to support the anti-stigma agenda in dementia care are required.

Vaginal assisted NOTES hysterectomy in The Netherlands; A prospective cohort study.

Objectives: Vaginal assisted Natural Orifice Transluminal Endoscopic Surgery (NOTES) combines the benefits of vaginal and endoscopic surgery. This study presents the results of the first vaginal assisted NOTES hysterectomies (VANH) in The Netherlands. Study design: A prospective cohort study was performed in two non-academic teaching hospitals in The Netherlands. Data was collected from patients who underwent a VANH for benign indications between August 2019 and April 2023. Baseline characteristics and data of intra- and postoperative surgical outcomes were recorded and analysed. The VANHs were performed by four experienced vaginal and endoscopic gynaecological surgeons. Results: A total of 200 patients underwent a VANH. Indications were dysfunctional menstrual bleeding (61 %; n = 122), abnormal cervical cytology (15.5 %; n = 31), abdominal pain (11.5 %; n = 23), post ablation/sterilization pain syndrome (3.5 %; n = 7), uterine fibroids (5.0 %; n = 10), atypical endometrial hyperplasia (2.5 %; n = 5) and Lynch or BRCA gene mutation carriers (1.0 %, n = 2). The mean surgical time was 61.4 min ( ± 22.8 min) with a mean blood loss of 88 mL ( ± 89 mL) and a mean uterine weight of 150 g ( ± 112 g). In 2.0 % (n = 4) of the cases a conversion was necessary. Same day discharge (SDD) was feasible in 80.2 % (n = 105) of the patients planned in day-care. In 2.0 % (n = 4) an intra-operative complication and in 9.0 % (n = 18) a post-operative complication occurred. Conclusion: This study shows vNOTES to be a safe and feasible surgical technique and can be safely implemented with appropriate patient selection and skilled surgeons. It highlights the importance of surgeon awareness of the challenges inherent in the initial stages of the implementation of a new surgical technique when performing their first vNOTES procedures. Additional randomized clinical trials are needed to show superiority of vNOTES compared to traditional surgery.

Assessing the Proficiency of LLMs with Various Tasks and Evaluators.

Previous studies have been limited to giving one or two tasks to Large Language Models (LLMs) and involved a small number of evaluators within a single domain to evaluate the LLM's answer. We assessed the proficiency of four LLMs by applying eight tasks and evaluating 32 results with 17 evaluators from diverse domains, demonstrating the significance of various tasks and evaluators on LLMs.

[Indications and technique for transoral thyroid gland and parathyroid gland surgery]. / Indikation und Technik der transoralen Schilddrüsen­/Nebenschilddrüsenchirurgie.

The conventional Kocher collar incision is the standard access to the thyroid and parathyroid glands. Although the incision length has been significantly shortened in recent years with this approach, there is increasing interest among patients in a surgical technique without visible scars in the décolleté. Transoral endoscopic thyroid gland surgery via the vestibular approach (TOETVA) is a modern technique that can be learned relatively quickly and leaves no visible scars because it is carried out exclusively through a natural orifice (natural orifice transluminal endoscopic surgery, NOTES). For retrieval of larger specimens, the transoral approach can be combined with a retroauricular access and thus covers a larger range of indications. The indications must be strictly followed, analogous to conventional surgery. Once the transoral access has been established, the operation is carried out as in open surgery but strictly from cranial to caudal. The classical complications are comparable to the results of conventional surgery. Specific complications include perioral, mandibular or cervical dysesthesia and hypesthesia.

Video Documentation in Thyroidectomy and an Evaluation of Operative Notes.

Introduction Accurate and detailed documentation of surgical operation notes is crucial for post-operative care, research and academic purposes, and medico-legal clarity. Several studies have shown their defiency and inaccuracy sometimes, and some methods have been proposed to make them more objective. This study aimed to evaluate the completeness of thyroidectomy operative notes in a tertiary center and to assess the adequacy of video documentation by comparing it to the corresponding operative notes. Methods A retrospective review of thyroidectomy operative notes from 2010 to 2020 at King Abdulaziz University Hospital, Jeddah, Saudi Arabia, was performed to ensure completeness. Subsequently, 15 thyroidectomies were video recorded, and their notes were compared to the corresponding written operative notes. The completeness score was calculated based on an item list that included items that had to be included in an operative note. An independent samples t-test was used to compare the completeness score means between the two groups. One-way analysis of variance was used to compare the completeness score means between two or more groups. Result A total of 385 thyroidectomy-operative notes were retrospectively reviewed. The completeness scores ranged between 6% and 89% for the various items that had to be documented, with a mean of 54.47%. The mean score of the video-documented operative record was 83.86%±12.84%, which was significantly higher than the corresponding written operative notes (47.53%±18.06%) (p <0.001). Conclusion Video documentation showed significant improvement compared to the corresponding written and retrospective operative notes. Video recording can also be a valuable tool when teaching anatomy and surgical skills and conducting research.

Extraction of sleep information from clinical notes of Alzheimer's disease patients using natural language processing.

OBJECTIVES: Alzheimer's disease (AD) is the most common form of dementia in the United States. Sleep is one of the lifestyle-related factors that has been shown critical for optimal cognitive function in old age. However, there is a lack of research studying the association between sleep and AD incidence. A major bottleneck for conducting such research is that the traditional way to acquire sleep information is time-consuming, inefficient, non-scalable, and limited to patients' subjective experience. We aim to automate the extraction of specific sleep-related patterns, such as snoring, napping, poor sleep quality, daytime sleepiness, night wakings, other sleep problems, and sleep duration, from clinical notes of AD patients. These sleep patterns are hypothesized to play a role in the incidence of AD, providing insight into the relationship between sleep and AD onset and progression. MATERIALS AND METHODS: A gold standard dataset is created from manual annotation of 570 randomly sampled clinical note documents from the adSLEEP, a corpus of 192 000 de-identified clinical notes of 7266 AD patients retrieved from the University of Pittsburgh Medical Center (UPMC). We developed a rule-based natural language processing (NLP) algorithm, machine learning models, and large language model (LLM)-based NLP algorithms to automate the extraction of sleep-related concepts, including snoring, napping, sleep problem, bad sleep quality, daytime sleepiness, night wakings, and sleep duration, from the gold standard dataset. RESULTS: The annotated dataset of 482 patients comprised a predominantly White (89.2%), older adult population with an average age of 84.7 years, where females represented 64.1%, and a vast majority were non-Hispanic or Latino (94.6%). Rule-based NLP algorithm achieved the best performance of F1 across all sleep-related concepts. In terms of positive predictive value (PPV), the rule-based NLP algorithm achieved the highest PPV scores for daytime sleepiness (1.00) and sleep duration (1.00), while the machine learning models had the highest PPV for napping (0.95) and bad sleep quality (0.86), and LLAMA2 with finetuning had the highest PPV for night wakings (0.93) and sleep problem (0.89). DISCUSSION: Although sleep information is infrequently documented in the clinical notes, the proposed rule-based NLP algorithm and LLM-based NLP algorithms still achieved promising results. In comparison, the machine learning-based approaches did not achieve good results, which is due to the small size of sleep information in the training data. CONCLUSION: The results show that the rule-based NLP algorithm consistently achieved the best performance for all sleep concepts. This study focused on the clinical notes of patients with AD but could be extended to general sleep information extraction for other diseases.

Historical Field Records Reveal Habitat as an Ecological Correlate of Locomotor Phenotypic Diversity in the Radiation of Neotropical Geophagini Fishes.

AbstractPhenotypic macroevolutionary studies provide insight into how ecological processes shape biodiversity. However, the complexity of phenotype-ecology relationships underscores the importance of also validating phenotype-based ecological inference with direct evidence of resource use. Unfortunately, macroevolutionary-scale ecological studies are often hindered by the challenges of acquiring taxonomically and spatially representative ecological data for large and widely distributed clades. The South American cichlid fish tribe Geophagini represents a continentally distributed radiation whose early locomotor morphological divergence suggests habitat as one ecological correlate of diversification, but an association between locomotor traits and habitat preference has not been corroborated. Field notes accumulated over decades of collecting across South America provide firsthand environmental records that can be mined for habitat data in support of macroevolutionary ecological research. In this study, we applied a newly developed method to transform descriptive field note information into quantitative habitat data and used it to assess habitat preference and its relationship to locomotor morphology in Geophagini. Field note-derived data shed light on geophagine habitat use patterns and reinforced habitat as an ecological correlate of locomotor morphological diversity. Our work emphasizes the rich data potential of museum collections, including often-overlooked material such as field notes, for evolutionary and ecological research.

Identification of an ANCA-Associated Vasculitis Cohort Using Deep Learning and Electronic Health Records.

Background: ANCA-associated vasculitis (AAV) is a rare but serious disease. Traditional case-identification methods using claims data can be time-intensive and may miss important subgroups. We hypothesized that a deep learning model analyzing electronic health records (EHR) can more accurately identify AAV cases. Methods: We examined the Mass General Brigham (MGB) repository of clinical documentation from 12/1/1979 to 5/11/2021, using expert-curated keywords and ICD codes to identify a large cohort of potential AAV cases. Three labeled datasets (I, II, III) were created, each containing note sections. We trained and evaluated a range of machine learning and deep learning algorithms for note-level classification, using metrics like positive predictive value (PPV), sensitivity, F-score, area under the receiver operating characteristic curve (AUROC), and area under the precision and recall curve (AUPRC). The deep learning model was further evaluated for its ability to classify AAV cases at the patient-level, compared with rule-based algorithms in 2,000 randomly chosen samples. Results: Datasets I, II, and III comprised 6,000, 3,008, and 7,500 note sections, respectively. Deep learning achieved the highest AUROC in all three datasets, with scores of 0.983, 0.991, and 0.991. The deep learning approach also had among the highest PPVs across the three datasets (0.941, 0.954, and 0.800, respectively). In a test cohort of 2,000 cases, the deep learning model achieved a PPV of 0.262 and an estimated sensitivity of 0.975. Compared to the best rule-based algorithm, the deep learning model identified six additional AAV cases, representing 13% of the total. Conclusion: The deep learning model effectively classifies clinical note sections for AAV diagnosis. Its application to EHR notes can potentially uncover additional cases missed by traditional rule-based methods.