A5057184477- Respiratory Support and Mechanisms
- Machine Learning in Healthcare
- Artificial Intelligence in Healthcare and Education
- Topic Modeling
- Sepsis Diagnosis and Treatment
- Chronic Obstructive Pulmonary Disease (COPD) Research
- Frailty in Older Adults
- Neonatal Respiratory Health Research
- Advanced Text Analysis Techniques
- Electronic Health Records Systems
- Palliative Care and End-of-Life Issues
- Bacterial Identification and Susceptibility Testing
- Intensive Care Unit Cognitive Disorders
- Text and Document Classification Technologies
- Colorectal Cancer Screening and Detection
- Cardiac Arrest and Resuscitation
- Antibiotic Use and Resistance
- Hemodynamic Monitoring and Therapy
- Antimicrobial Resistance in Staphylococcus
- Clostridium difficile and Clostridium perfringens research
- Biomedical and Engineering Education
- Natural Language Processing Techniques
- Emergency and Acute Care Studies
- Text Readability and Simplification
- Biomedical Text Mining and Ontologies
New York University
2016-2025
NYU Langone Health
2016-2024
Center for Innovation
2022
Hinge Health
2021
University of Canterbury
2014-2016
Christian Medical College
1992
Christian Medical College & Hospital
1992
Importance Virtual patient-physician communications have increased since 2020 and negatively impacted primary care physician (PCP) well-being. Generative artificial intelligence (GenAI) drafts of patient messages could potentially reduce health professional (HCP) workload improve communication quality, but only if the are considered useful. Objectives To assess PCPs’ perceptions GenAI to examine linguistic characteristics associated with equity perceived empathy. Design, Setting,...
Real-time patient respiratory mechanics estimation can be used to guide mechanical ventilation settings, particularly, positive end-expiratory pressure (PEEP). This work presents a software, Clinical Utilisation of Respiratory Elastance (CURE Soft), using time-varying elastance model offer this ability aid in treatment.CURE Soft is desktop application developed JAVA. It has two modes operation, 1) Online real-time monitoring decision support and, 2) Offline for user education purposes,...
The COVID-19 pandemic has challenged front-line clinical decision-making, leading to numerous published prognostic tools. However, few models have been prospectively validated and none report implementation in practice. Here, we use 3345 retrospective 474 prospective hospitalizations develop validate a parsimonious model identify patients with favorable outcomes within 96 h of prediction, based on real-time lab values, vital signs, oxygen support variables. In validation, the achieves high...
This descriptive study evaluates the association between nursing reports of sepsis overalerting and alert volume by quantifying number alerts generated Epic Sepsis Model at 24 US hospitals before during COVID-19 pandemic.
Abstract Objectives To evaluate the proficiency of a HIPAA-compliant version GPT-4 in identifying actionable, incidental findings from unstructured radiology reports Emergency Department patients. assess appropriateness artificial intelligence (AI)-generated, patient-facing summaries these findings. Materials and Methods Radiology extracted electronic health record large academic medical center were manually reviewed to identify non-emergent, with high likelihood requiring follow-up, further...
BACKGROUND: Myocardial injury detected after percutaneous coronary intervention (PCI) is associated with increased mortality. Predictors of post-PCI myocardial are not well established. The long-term prognostic relevance remains uncertain. METHODS: Consecutive adults aged ≥18 years stable ischemic heart disease who underwent elective PCI at NYU Langone Health between 2011 and 2020 were included in a retrospective, observational study. Patients acute infarction or creatinine kinase band...
Abstract Importance Large language models (LLMs) are crucial for medical tasks. Ensuring their reliability is vital to avoid false results. Our study assesses two state-of-the-art LLMs (ChatGPT and LlaMA-2) extracting clinical information, focusing on cognitive tests like MMSE CDR. Objective Evaluate ChatGPT LlaMA-2 performance in CDR scores, including associated dates. Methods data consisted of 135,307 notes (Jan 12th, 2010 May 24th, 2023) mentioning MMSE, CDR, or MoCA. After applying...
Ensuring reliability of Large Language Models (LLMs) in clinical tasks is crucial. Our study assesses two state-of-the-art LLMs (ChatGPT and LlaMA-2) for extracting information, focusing on cognitive tests like MMSE CDR. data consisted 135,307 notes (Jan 12th, 2010 to May 24th, 2023) mentioning MMSE, CDR, or MoCA. After applying inclusion criteria 34,465 remained, which 765 underwent ChatGPT (GPT-4) LlaMA-2, 22 experts reviewed the responses. successfully extracted CDR instances with dates...
Background Healthcare crowdsourcing events (e.g. hackathons) facilitate interdisciplinary collaboration and encourage innovation. Peer-reviewed research has not yet considered a healthcare event focusing on generative artificial intelligence (GenAI), which generates text in response to detailed prompts vast potential for improving the efficiency of organizations. Our event, New York University Langone Health (NYULH) Prompt-a-thon, primarily sought inspire build AI fluency within our diverse...
Accelerating demand for patient messaging has impacted the practice of many providers. Messages are not recommended urgent medical issues, but some do require rapid attention. This presents an opportunity artificial intelligence (AI) methods to prioritize review messages. Our study aimed highlight portal messages prioritized using a custom AI system integrated into electronic health record (EHR).
Abstract Objectives The study aimed to assess the usage and impact of a private secure instance generative artificial intelligence (GenAI) application in large academic health center. goal was understand how employees interact with this technology influence on their perception skill work performance. Materials Methods New York University Langone Health (NYULH) established secure, private, managed Azure OpenAI service (GenAI Studio) granted widespread access employees. Usage monitored users...
SummaryPatients with serious, life-limiting disease benefit from end-of-life conversations, goal setting, and palliative care. Hospitalized patients at high risk of near-term death are likely to such interventions. As NYU Langone Health expanded its institutional initiatives promoting patient-centered care, leaders developed an artificial intelligence–based system that identifies dying within 2 months. Upon opening a high-risk patient's chart, attending physicians receive interruptive alert...
Abstract Background Automated systems that use machine learning to estimate a patient’s risk of death are being developed influence care. There remains sparse transparent reporting model generalizability in different subpopulations especially for implemented systems. Methods A prognostic study included adult admissions at multi-site, academic medical center between 2015 and 2017. predictive all-cause mortality (including initiation hospice care) within 60 days admission was developed. Model...
Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI
Asynchronous Events (AEs) during mechanical ventilation (MV) result in increased work of breathing and potential poor patient outcomes. Thus, it is important to automate AE detection. In this study, an detection method, Automated Logging Inspiratory Expiratory Non-synchronized (ALIEN) was developed compared between standard manual 11 MV patients. A total 5701 breaths were analyzed (median [IQR]: 500 [469-573] per patient). The Asynchrony Index (AI) 51% [28-78]%. yielded sensitivity 90.3%...
Abstract Objective One primary consideration when developing predictive models is downstream effects on future model performance. We conduct experiments to quantify the of experimental design choices, namely cohort selection and internal validation methods, (estimated) real-world Materials Methods Four years hospitalizations are used develop a 1-year mortality prediction (composite death or initiation hospice care). Two common methods select appropriate patient visits from their encounter...
Purpose: Patient-specific respiratory mechanics can be used to guide mechanical ventilation therapy. However, even in controlled modes, underlying masked by spontaneous breathing efforts. The aim of this study is accurately assess for cycles affected these Methods: A pressure reconstruction eliminating the demand effect respiration (PREDATOR) method reconstruct profiles (breath specific elastance and resistance). tested on both simulated clinical data comprising n=264 breaths. Results: Using...
Abstract Pathology reports are considered the gold standard in medical research due to their comprehensive and accurate diagnostic information. Natural language processing (NLP) techniques have been developed automate information extraction from pathology reports. However, existing studies suffer two significant limitations. First, they typically frame tasks as report classification, which restricts granularity of extracted Second, often fail generalize unseen variations language, negation,...