Brief Communication: More Chat about Chatbots Assisting with Routine Clinical Biochemistry Queries

Original Research Articles Brief Communication: More Chat about Chatbots Assisting with Routine Clinical Biochemistry Queries Rucita Severaj, Nareshni Moodley and Verena Gounden* Department of Chemical Pathology, Inkosi Albert Luthuli Central Hospital, University of KwaZulu Natal and National Health Laboratory Service, 800 Vusi Mzimela Road, Cator Manor, Durban 4091, Kwa-Zulu Natal, South Africa * Correspondence: [email protected]; Tel.: +353-9182800 Received: 9 March 2025; Accepted: 25 March 2025; Published: 9 April 2025 Abstract: Background: Artificial intelligence (AI) technologies have the potential to transform many aspects of patient care and is playing an increasing role in health care in diagnostics and patient management. Several studies have already demonstrated that AI can perform as well as or better than humans at key healthcare tasks. The aim of this study was to evaluate and compare the utility of the AI chatbots ChatGPT and Deepseek to answer everyday chemical pathology queries as handled by the registrars in the Chemical Pathology Department at Inkosi Albert Luthuli Central Hospital in Durban, South Africa. Method: All queries received by the registrars over a month period were documented and answers from traditional sources such as laboratory procedures and sample handbooks were documented. The queries were later asked to ChatGPT and Deepseek and further evaluated and the answers evaluated by two blinded senior pathologists in terms of suitability and accuracy of responses and key word recognition. Results: A total of 37 queries were asked to the chatbots. Based on average scoring of the two reviewers 97% (n = 36/37) of queries for ChatGPT responses and 73% (n = 27) for Deepseek were ranked 3 and above for suitability and accuracy. The poorly scoring responses from both Deepseek and ChatGPT related to questions that were very specific to the local laboratory or testing in the laboratory. Conclusion: A longer-term evaluation and verification of ChatGPT as a resource to assist with lab related queries is required. It may be a useful resource not only to the trainee chemical pathology registrar but to under-resourced health care settings where pathologist support may not be present.

Authors

Institutions

Publication Details

Journal
International Journal of Clinical and Translational Medicine
Published
2025-04-08
DOI
https://doi.org/10.53941/ijctm.2025.1000011
Primary Topic
Artificial Intelligence in Healthcare and Education
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Brief Communication: More Chat about Chatbots Assisting with Routine Clinical Biochemistry Queries

Nareshni Moodley, Verena Gounden, Rucita Severaj
International Journal of Clinical and Translational Medicine
Artificial Intelligence in Healthcare and Education
article

Brief Communication: More Chat about Chatbots Assisting with Routine Clinical Biochemistry Queries

Nareshni Moodley, Verena Gounden, Rucita Severaj
article en

Abstract

Original Research Articles Brief Communication: More Chat about Chatbots Assisting with Routine Clinical Biochemistry Queries Rucita Severaj, Nareshni Moodley and Verena Gounden* Department of Chemical Pathology, Inkosi Albert Luthuli Central Hospital, University of KwaZulu Natal and National Health Laboratory Service, 800 Vusi Mzimela Road, Cator Manor, Durban 4091, Kwa-Zulu Natal, South Africa * Correspondence: [email protected]; Tel.: +353-9182800 Received: 9 March 2025; Accepted: 25 March 2025; Published: 9 April 2025 Abstract: Background: Artificial intelligence (AI) technologies have the potential to transform many aspects of patient care and is playing an increasing role in health care in diagnostics and patient management. Several studies have already demonstrated that AI can perform as well as or better than humans at key healthcare tasks. The aim of this study was to evaluate and compare the utility of the AI chatbots ChatGPT and Deepseek to answer everyday chemical pathology queries as handled by the registrars in the Chemical Pathology Department at Inkosi Albert Luthuli Central Hospital in Durban, South Africa. Method: All queries received by the registrars over a month period were documented and answers from traditional sources such as laboratory procedures and sample handbooks were documented. The queries were later asked to ChatGPT and Deepseek and further evaluated and the answers evaluated by two blinded senior pathologists in terms of suitability and accuracy of responses and key word recognition. Results: A total of 37 queries were asked to the chatbots. Based on average scoring of the two reviewers 97% (n = 36/37) of queries for ChatGPT responses and 73% (n = 27) for Deepseek were ranked 3 and above for suitability and accuracy. The poorly scoring responses from both Deepseek and ChatGPT related to questions that were very specific to the local laboratory or testing in the laboratory. Conclusion: A longer-term evaluation and verification of ChatGPT as a resource to assist with lab related queries is required. It may be a useful resource not only to the trainee chemical pathology registrar but to under-resourced health care settings where pathologist support may not be present.

International Journal of Clinical and Translational Medicine
University of KwaZulu-Natal (ZA)
Openalex Percentile: Top 96%
Artificial Intelligence in Healthcare and Education
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.