Mattias Ohlsson
Professor
Design of an AI Support for Diagnosis of Dyspneic Adults at Time of Triage in the Emergency Department
Author
Summary, in English
We created an AI support for diagnosis in dyspneic adults at time of triage in the emergency department.
Complete data from an entire regional health care system was analyzed, to find AI-derived, unknown, important diagnostic predictors. Most important were prior diagnoses of heart failure or COPD, daily smoking, atrial fibrillation/flutter, life difficulties and maternal care.
Sensitivity for AHF, eCOPD and pneumonia was 75%, 93%, and 54%, respectively, with a specificity above 75%.
Each patient visit received an individual graph with the AI´s underlying decision basis.
Complete data from an entire regional health care system was analyzed, to find AI-derived, unknown, important diagnostic predictors. Most important were prior diagnoses of heart failure or COPD, daily smoking, atrial fibrillation/flutter, life difficulties and maternal care.
Sensitivity for AHF, eCOPD and pneumonia was 75%, 93%, and 54%, respectively, with a specificity above 75%.
Each patient visit received an individual graph with the AI´s underlying decision basis.
Department/s
- Emergency medicine
- EpiHealth: Epidemiology for Health
- LU Profile Area: Natural and Artificial Cognition
- eSSENCE: The e-Science Collaboration
- Artificial Intelligence in CardioThoracic Sciences (AICTS)
- EPI@LUND
- Surgery and public health
- Cardiovascular Research - Hypertension
Publishing year
2023-09
Language
English
Full text
Document type
Poster
Topic
- Cardiac and Cardiovascular Systems
Keywords
- artificial intelligence
- AI
- Dyspnea
- Artificiell intelligens
- AI
- Dyspne
Conference name
European Emergency Medicine Congress 2023
Conference date
2023-09-17 - 2023-09-20
Conference place
Barcelona, Spain
Status
Published
Project
- Resource Management in the Emergency Department by using Machine Learning
- AIR Lund - Artificially Intelligent use of Registers
Research group
- Emergency medicine
- Artificial Intelligence in CardioThoracic Sciences (AICTS)
- EPI@LUND
- Surgery and public health
- Cardiovascular Research - Hypertension