A5005374943- Artificial Intelligence in Healthcare and Education
- Cutaneous Melanoma Detection and Management
- Machine Learning in Healthcare
- AI in cancer detection
- Topic Modeling
- COVID-19 epidemiological studies
- COVID-19 Pandemic Impacts
- Biomedical Text Mining and Ontologies
- Interpreting and Communication in Healthcare
- Colorectal Cancer Screening and Detection
- Emergency and Acute Care Studies
- Ethics in Clinical Research
- Diversity and Career in Medicine
- Genital Health and Disease
- Peripheral Artery Disease Management
- Organ Donation and Transplantation
- Surgical Simulation and Training
- Cancer Diagnosis and Treatment
- Health Systems, Economic Evaluations, Quality of Life
- Global Cancer Incidence and Screening
- COVID-19 Clinical Research Studies
- Cerebrospinal fluid and hydrocephalus
- Sepsis Diagnosis and Treatment
- Immune responses and vaccinations
- Nonmelanoma Skin Cancer Studies
Stanford University
2021-2025
DePaul University
2024
Hospital Israelita Albert Einstein
2024
Universidad Nacional Autónoma de México
2024
University of Lübeck
2024
Royal Free London NHS Foundation Trust
2024
Chase Farm Hospital
2024
Barnet and Chase Farm NHS Hospitals Trust
2024
Stanford Medicine
2023-2024
University of Alberta
2024
Abstract Large language models (LLMs) are being integrated into healthcare systems; but these may recapitulate harmful, race-based medicine. The objective of this study is to assess whether four commercially available large propagate inaccurate, content when responding eight different scenarios that check for medicine or widespread misconceptions around race. Questions were derived from discussions among physician experts and prior work on medical believed by trainees. We assessed with nine...
We have previously shown that polygenic risk scores (PRS) can improve stratification of peripheral artery disease (PAD) in a large, retrospective cohort. Here, we evaluate the potential PRS improving detection PAD and prediction major adverse cardiovascular cerebrovascular events (MACCE) (AE) an institutional patient created cohort 278 patients (52 cases 226 controls) fit PAD-specific based on weighted sum alleles. built traditional clinical models machine learning (ML) using genetic...
0. Abstract Background The integration of large language models (LLMs) in healthcare offers immense opportunity to streamline tasks, but also carries risks such as response accuracy and bias perpetration. To address this, we conducted a red-teaming exercise assess LLMs developed dataset clinically relevant scenarios for future teams use. Methods We convened 80 multi-disciplinary experts evaluate the performance popular across multiple medical scenarios. Teams composed clinicians, engineering...
Abstract Recently emerging large multimodal models (LMMs) utilize various types of data modalities, including text and visual inputs to generate outputs. The incorporation LMMs into clinical medicine presents unique challenges, accuracy, reliability, relevance. Here, we explore applications GPT-4V, an LMM that has been proposed for use in medicine, gastroenterology, radiology, dermatology, United States Medical Licensing Examination (USMLE) test questions. We used standardized robust...
Red teaming, the practice of adversarially exposing unexpected or undesired model behaviors, is critical towards improving equity and accuracy large language models, but non-model creator-affiliated red teaming scant in healthcare. We convened teams clinicians, medical engineering students, technical professionals (80 participants total) to stress-test models with real-world clinical cases categorize inappropriate responses along axes safety, privacy, hallucinations/accuracy, bias. Six...
Cancer incidence and mortality is increasing worldwide. In 2018, there were an estimated 18.1 million new cancer cases 9.6 deaths. Nigeria, it that 100,000 occur annually, with a high case fatality ratio. The burden of in Nigeria significant, as the country still grapples infectious diseases has limited data on epidemiology. Our study descriptive using from hospital-based registry.This retrospective assesses characteristics patients presented to private center Lagos, Nigeria. We aimed update...
Importance Large language models (LLMs) are being integrated into healthcare systems; but these recapitulate harmful, race-based medicine. Objective The objective of this study is to assess whether four commercially available large propagate inaccurate, content when responding eight different scenarios that historically included medicine or widespread misconceptions around race. Evidence Review Questions were derived from discussion among 4 physician experts and prior work on medical...
Abstract With an estimated 3 billion people globally lacking access to dermatological care, technological solutions leveraging artificial intelligence (AI) have been proposed improve 1 . Diagnostic AI algorithms, however, require high-quality datasets allow development and testing, particularly those that enable evaluation of both unimodal multimodal approaches. Currently, the majority dermatology algorithms are built tested on proprietary, siloed data, often from a single site with only...
Verifying factual claims is critical for using large language models (LLMs) in healthcare. Recent work has proposed fact decomposition, which uses LLMs to rewrite source text into concise sentences conveying a single piece of information, as an approach fine-grained verification. Clinical documentation poses unique challenges decomposition due dense terminology and diverse note types. To explore these challenges, we present FactEHR, dataset consisting full document decompositions 2,168...
In this study, we investigate the performance of computer vision AI algorithms in predicting patient disposition from emergency department (ED) using short video clips. Clinicians often use "eye-balling" or clinical gestalt to aid triage, based on brief observations. We hypothesize that can similarly appearance for prediction. Data were collected adult patients at an academic ED, with mobile phone videos capturing performing simple tasks. Our algorithm, alone, showed better hospital...