Articles liés à Artificial Intelligence in Emergency Medicine: Clinical...

Artificial Intelligence in Emergency Medicine: Clinical Decision Support, Triage, Predictive Analytics, Large Language Models, Workflow Integration, Safety, Governance & Implementation - Couverture souple

Rebell, Sophie

 
9798161143964: Artificial Intelligence in Emergency Medicine: Clinical Decision Support, Triage, Predictive Analytics, Large Language Models, Workflow Integration, Safety, Governance & Implementation

Synopsis

Stop guessing about AI in your emergency department.


AI tools are arriving faster than most departments can evaluate them. Vendors promise earlier sepsis detection, faster triage and the end of charting. Clinicians are already using chatbots on their own. The evidence is uneven, the regulations are changing and the cost of a wrong decision falls on patients.

AI in Emergency Medicine is a practical, evidence-aware guide for emergency physicians, nurses, informatics leads, administrators and quality and safety teams. It does not assume you can code. It teaches you to ask the right questions, read the evidence critically and put sensible governance around the tools you adopt.

What you'll learn

How machine learning and generative AI work, and why they fail in different ways
How to read a validation study, including sensitivity, PPV, calibration and why a high AUROC can still produce unmanageable alerts
Where AI is being used today: triage, deterioration prediction, patient flow, imaging, ECG, ultrasound, ambient documentation, patient communication and agents
How bias, dataset shift, automation bias and alert fatigue cause harm, and how to defend against them
What FDA clearance does and does not mean, and how the EU and other frameworks are changing
How to protect patient privacy when using language models and vendor tools

What makes it a working tool

A ten-question framework for evaluating any AI tool before you buy it
A ten-category scorecard with scoring anchors and veto rules
A ten-phase implementation roadmap, from problem definition to retirement
Governance checklists, a sample acceptable-use policy, a vendor questionnaire, pilot and silent-validation templates, an incident report form and a monitoring dashboard specification
Case studies drawn from published studies, and hypothetical scenarios that show the framework in action
About 30 original diagrams and a glossary of 150 terms

The book is candid about limits. Where evidence is thin it says so, where a claim could not be verified it is flagged, and it separates what AI can do from what has been shown to help patients.

If you have to decide whether an AI tool belongs in your department, this book gives you a method for making that decision.

Les informations fournies dans la section « Synopsis » peuvent faire référence à une autre édition de ce titre.

Vos droits au titre de la garantie légale : Chaque consommateur de l'UE bénéficie d'au moins deux ans de protection au titre de la garantie légale pour tout bien de consommation vendu dans l'UE. Voir plus