In Patients Evaluated In The Emergency Room With Suspected Community-Acquired Pneumonia, ChatGPT 3.5 May Help Physicians With Assessments And Plans.

Introduction The effectiveness of ChatGPT in generating an appropriate assessment and plan for patients arriving in the emergency department (ED) with a syndrome compatible with community-acquired pneumonia (CAP) is not well defined. The objective of this study is to perform a clinical evaluation of the quality and appropriateness of synthetic assessments and management plans generated by ChatGPT 3.5 for adult patients presenting at the ED with suspected CAP. Methods This was a retrospective observational study of 10 cases of adult patients arriving to the ED with respiratory complaints. Patients’ history, physical examination, laboratory and radiology reports were provided to ChatGPT version 3.5. We prompted ChatGPT to provide assessment, diagnostic impression, management, and disposition of these patients, with further prompts asking for clarifying questions as needed. The responses were evaluated and graded by a committee of four physicians (one infectious diseases, two pulmonologists, and one critical care) according to a 10-point scale. Responses were evaluated by each member of the committee, followed by a discussion of each individual patient to reach agreement on the final evaluation. Results ChatGPT’s recommendations were rated between 7 and 9 (of 10) in all cases, with a mean score of 7.8. (SD=0.6), consistent with appropriate evaluation at the level of a post-graduate trainee. In three cases, ChatGPT omitted important information such as positive test results for influenza or RSV, which required further questioning. No assessment or plan was found to be inappropriate or harmful to the patient. Conclusion This study found that ChatGPT 3.5 provided highly appropriate or appropriate recommendations for the assessment and management plan for patients presenting at an ED with a clinical syndrome suspecting CAP. ChatGPT may be a useful tool to support physicians’ decision-making during the initial evaluation of patients presenting to the ED with respiratory complaints.

Authors

Institutions

Publication Details

Journal
Norton Healthcare Medical Journal
Published
2024-12-31
DOI
https://doi.org/10.59541/001c.127927
Citations
1
Primary Topic
Emergency and Acute Care Studies
Type
article
Field-Weighted Citation Impact
0.69
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

In Patients Evaluated In The Emergency Room With Suspected Community-Acquired Pneumonia, ChatGPT 3.5 May Help Physicians With Assessments And Plans.

William A. Mattingly, Timothy L. Wiemken, Matthew Song, Thomas Chandler et al.
1 citations
Norton Healthcare Medical Journal
Emergency and Acute Care Studies
0.69
article

In Patients Evaluated In The Emergency Room With Suspected Community-Acquired Pneumonia, ChatGPT 3.5 May Help Physicians With Assessments And Plans.

William A. Mattingly, Timothy L. Wiemken, Matthew Song, Thomas Chandler, Alex Glynn, Rodrigo Cavallazzi, Julio A. Ramírez, Ruth Carrico, Stephen Furmanek, Alejandro Chirino, Guillermo Cabral, Josue Prado, Kirsten Trudeau
article en
1 citations

Abstract

Introduction The effectiveness of ChatGPT in generating an appropriate assessment and plan for patients arriving in the emergency department (ED) with a syndrome compatible with community-acquired pneumonia (CAP) is not well defined. The objective of this study is to perform a clinical evaluation of the quality and appropriateness of synthetic assessments and management plans generated by ChatGPT 3.5 for adult patients presenting at the ED with suspected CAP. Methods This was a retrospective observational study of 10 cases of adult patients arriving to the ED with respiratory complaints. Patients’ history, physical examination, laboratory and radiology reports were provided to ChatGPT version 3.5. We prompted ChatGPT to provide assessment, diagnostic impression, management, and disposition of these patients, with further prompts asking for clarifying questions as needed. The responses were evaluated and graded by a committee of four physicians (one infectious diseases, two pulmonologists, and one critical care) according to a 10-point scale. Responses were evaluated by each member of the committee, followed by a discussion of each individual patient to reach agreement on the final evaluation. Results ChatGPT’s recommendations were rated between 7 and 9 (of 10) in all cases, with a mean score of 7.8. (SD=0.6), consistent with appropriate evaluation at the level of a post-graduate trainee. In three cases, ChatGPT omitted important information such as positive test results for influenza or RSV, which required further questioning. No assessment or plan was found to be inappropriate or harmful to the patient. Conclusion This study found that ChatGPT 3.5 provided highly appropriate or appropriate recommendations for the assessment and management plan for patients presenting at an ED with a clinical syndrome suspecting CAP. ChatGPT may be a useful tool to support physicians’ decision-making during the initial evaluation of patients presenting to the ED with respiratory complaints.

Norton Healthcare Medical JournalVol. 1(2)
University of Louisville (US), Northern Kentucky University (US), Norton Healthcare (US), Hospital Italiano La Plata (AR), UCLouvain Saint-Louis Brussels (BE), University of Louisville Hospital (US), University of Louisville Physicians (US)
Peace, Justice and strong institutions
Openalex Percentile: Top 23%
Emergency and Acute Care Studies
0.69
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.