Search AMIA.org

Deviation from evidence-based guidelines is common and associated with worse patient outcomes, especially in hectic emergency departments (EDs). Clinical decision support (CDS) systems can improve outcomes by promoting guideline adherence while allowing patient-specific adaptation. Implementation of a CDS system for pneumonia ("ePneumonia") in the ED has been associated with improved guideline adherence and reduced 30-day mortality. However, adoption of ePneumonia has been hindered by a suboptimal user interface (UI).This study [...]

Ambient listening tools utilize generative artificial intelligence (AI) to create clinical notes from real-time conversations between clinicians and patients during an encounter. One of the potential benefits of ambient listening tools is an improvement in reported patient experience.This study aimed to compare the patient experience of an outpatient visit during which an ambient listening tool is used with a standard visit and to quantify any perceived improvements in care.Patients completed [...]

Learning about critically ill children's outcomes after transfer to the pediatric intensive care unit (PICU) can help emergency department (ED) physicians improve future performance. However, there are no standard processes in place to systematically provide this information; thus, most ED physicians obtain inconsistent feedback.We aimed to determine the effect of delivering patient outcome feedback through the electronic health record (EHR) on the frequency of ED physicians' re-access of patients' EHRs [...]

This study aimed to compare two ambient AI documentation tools, Dragon Ambient eXperience (DAX) and Abridge in the emergency department (ED), assessing perceived effects on work burden, usability, documentation quality, satisfaction, and overall preference.We conducted a single-site, prospective crossover study in an ED over 6 weeks, from April to June 2025. Out of 20 faculty enrolled, 18 completed both phases. Participants used both ambient AI scribe tools in alternating 3-week [...]

Real-world data from vendor-aggregated health information exchanges represent a powerful resource for retrospective clinical research. Epic's Cosmos platform, a large-scale centralized data warehouse, enables querying deidentified electronic health record data.We describe the iterative development of a generalizable clinical informatics workflow for retrospective studies using Cosmos.We applied the Plan-Do-Study-Act cycle to iteratively refine a collaborative research process using Cosmo's SlicerDicer interface, focusing on setting, team, technical preparation, and workflow.We identified multiple [...]

To develop a framework that models the impact of electronic health record (EHR) systems on healthcare professionals' well-being and their relationships with patients, using interdisciplinary insights to guide machine learning in identifying value patterns important to healthcare professionals in EHR systems.

This study introduces Smart Imitator (SI), a 2-phase reinforcement learning (RL) solution enhancing personalized treatment policies in healthcare, addressing challenges from imperfect clinician data and complex environments.

We aim to use large language models (LLMs) to detect mentions of nuanced psychotherapeutic outcomes and impacts than previously considered in transcripts of interviews with adolescent depression. Our clinical authors previously created a novel coding framework containing fine-grained therapy outcomes beyond the binary classification (eg, depression vs control) based on qualitative analysis embedded within a clinical study of depression. Moreover, we seek to demonstrate that embeddings from LLMs are informative [...]

To pioneer the first artificial intelligence system integrating radiological and objective clinical data, simulating the clinical reasoning process, for the early prediction of high-risk influenza patients.

Machine learning applications for longitudinal electronic health records often forecast the risk of events at fixed time points, whereas survival analysis achieves dynamic risk prediction by estimating time-to-event distributions. Here, we propose a novel conditional variational autoencoder-based method, DySurv, which uses a combination of static and longitudinal measurements from electronic health records to estimate the individual risk of death dynamically.