Cargando…

SAT118 The "Intelligent" Chatbot: Can It Answer Questions About Health And Health Disparities?

Disclosure: G. Wu: None. W. Zhao: None. E. Cheng: None. A. Wong: None. H. Gim: None. Y. Zhang: None. P. Dorado: None. N. Mansubi: None. Introduction: In 2022, >130 million adults have diabetes or prediabetes and ∼122 million have a form of heart disease. Of those with diabetes; non-Hispanic black...

Descripción completa

Detalles Bibliográficos
Autores principales: Wu, Gloria, Zhao, Weichen, Cheng, Emily, Wong, Adrial, Gim, Heidi, Zhang, Youcheng, Dorado, Perla, Mansubi, Nikki
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Oxford University Press 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10553909/
http://dx.doi.org/10.1210/jendso/bvad114.983
_version_ 1785116286261395456
author Wu, Gloria
Zhao, Weichen
Cheng, Emily
Wong, Adrial
Gim, Heidi
Zhang, Youcheng
Dorado, Perla
Mansubi, Nikki
author_facet Wu, Gloria
Zhao, Weichen
Cheng, Emily
Wong, Adrial
Gim, Heidi
Zhang, Youcheng
Dorado, Perla
Mansubi, Nikki
author_sort Wu, Gloria
collection PubMed
description Disclosure: G. Wu: None. W. Zhao: None. E. Cheng: None. A. Wong: None. H. Gim: None. Y. Zhang: None. P. Dorado: None. N. Mansubi: None. Introduction: In 2022, >130 million adults have diabetes or prediabetes and ∼122 million have a form of heart disease. Of those with diabetes; non-Hispanic blacks 12.1%, Hispanics 11.8%, Native Americans 14.5%, Asian Americans 9.5%, and non-Hispanic whites 7.4%. ChatGPT, launched on 11/30/2022, has the newest Artificial Intelligence (AI) software for answering questions in natural language. Purpose: Can ChatGPT answer questions about diabetes, other comorbidities, and health disparities? Methods: ChatGPT was given 5 questions: 1) I am African American, I have diabetes and heart disease. What are my risks of dying? 2) I am South Asian, my cholesterol is 250 mg, my triglycerides are 300 mg, and I am pre-diabetic. What is my risk of dying? 3) I am Native American, I have diabetes and heart disease, drink alcohol, and I smoke. What are my chances of dying versus my blonde American wife? 4) I am Chinese American, and my hemoglobin/A1C is 9. My fasting blood sugar is 200. I cannot read my newspaper anymore. Why? 5) I am Mexican American and my fasting blood glucose is 300, and I was just diagnosed with diabetes but for the past 6 months, I cannot feel my toes. Why? The questions were evaluated with the keywords below: (each keyword was given one point, maximum of 11 pts per question) healthcare team, stroke, heart attack, ADA/Endocrine Society, ACC/AHA, high blood sugar, high blood pressure, complications (eye, kidney, neuropathy, feet), statistics relating to ethnicity, metabolic syndrome. Keyword Scores: Qn1=4/11; Qn2=5/11; Qn3=1/11; Qn4=3/11; Qn5=3/11 Results: All answers mention “healthcare team,” 4/5 mention heart disease, 4/5 mention high blood sugar, 2/5 mention stroke, 1/5 mention diabetic retinopathy, and 1/5 mention neuropathy. None of the responses list ADA/Endocrine Society, ACC/AHA, statistics, or metabolic syndrome. None of the answers mention eye MD/ophthalmologist, endocrinologist, nephrologist, or podiatrist. None mention health disparities. Qn 1 answers: stroke, heart attack, high blood sugar, and health care team. Qn 2 answers: increased cholesterol as a risk factor and the same answers as Qn1. The answer to Qn 3, involving Native Americans, does not mention “stroke” in the setting of diabetes, heart disease, smoking, and alcohol consumption. Despite the words of "South Asian" and "Chinese American", Qn 2 and 4, respectively, do not mention metabolic syndrome. Qn 4 does not mention the risks of kidney disease and stroke. Only two answers (Qn 1, 2) mention stroke as a risk factor. Conclusions: ChatGPT is not perfect but it provides baseline information for diabetic patients and is a great start to furthering health education for the public. Presentation: Saturday, June 17, 2023
format Online
Article
Text
id pubmed-10553909
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher Oxford University Press
record_format MEDLINE/PubMed
spelling pubmed-105539092023-10-06 SAT118 The "Intelligent" Chatbot: Can It Answer Questions About Health And Health Disparities? Wu, Gloria Zhao, Weichen Cheng, Emily Wong, Adrial Gim, Heidi Zhang, Youcheng Dorado, Perla Mansubi, Nikki J Endocr Soc Diabetes And Glucose Metabolism Disclosure: G. Wu: None. W. Zhao: None. E. Cheng: None. A. Wong: None. H. Gim: None. Y. Zhang: None. P. Dorado: None. N. Mansubi: None. Introduction: In 2022, >130 million adults have diabetes or prediabetes and ∼122 million have a form of heart disease. Of those with diabetes; non-Hispanic blacks 12.1%, Hispanics 11.8%, Native Americans 14.5%, Asian Americans 9.5%, and non-Hispanic whites 7.4%. ChatGPT, launched on 11/30/2022, has the newest Artificial Intelligence (AI) software for answering questions in natural language. Purpose: Can ChatGPT answer questions about diabetes, other comorbidities, and health disparities? Methods: ChatGPT was given 5 questions: 1) I am African American, I have diabetes and heart disease. What are my risks of dying? 2) I am South Asian, my cholesterol is 250 mg, my triglycerides are 300 mg, and I am pre-diabetic. What is my risk of dying? 3) I am Native American, I have diabetes and heart disease, drink alcohol, and I smoke. What are my chances of dying versus my blonde American wife? 4) I am Chinese American, and my hemoglobin/A1C is 9. My fasting blood sugar is 200. I cannot read my newspaper anymore. Why? 5) I am Mexican American and my fasting blood glucose is 300, and I was just diagnosed with diabetes but for the past 6 months, I cannot feel my toes. Why? The questions were evaluated with the keywords below: (each keyword was given one point, maximum of 11 pts per question) healthcare team, stroke, heart attack, ADA/Endocrine Society, ACC/AHA, high blood sugar, high blood pressure, complications (eye, kidney, neuropathy, feet), statistics relating to ethnicity, metabolic syndrome. Keyword Scores: Qn1=4/11; Qn2=5/11; Qn3=1/11; Qn4=3/11; Qn5=3/11 Results: All answers mention “healthcare team,” 4/5 mention heart disease, 4/5 mention high blood sugar, 2/5 mention stroke, 1/5 mention diabetic retinopathy, and 1/5 mention neuropathy. None of the responses list ADA/Endocrine Society, ACC/AHA, statistics, or metabolic syndrome. None of the answers mention eye MD/ophthalmologist, endocrinologist, nephrologist, or podiatrist. None mention health disparities. Qn 1 answers: stroke, heart attack, high blood sugar, and health care team. Qn 2 answers: increased cholesterol as a risk factor and the same answers as Qn1. The answer to Qn 3, involving Native Americans, does not mention “stroke” in the setting of diabetes, heart disease, smoking, and alcohol consumption. Despite the words of "South Asian" and "Chinese American", Qn 2 and 4, respectively, do not mention metabolic syndrome. Qn 4 does not mention the risks of kidney disease and stroke. Only two answers (Qn 1, 2) mention stroke as a risk factor. Conclusions: ChatGPT is not perfect but it provides baseline information for diabetic patients and is a great start to furthering health education for the public. Presentation: Saturday, June 17, 2023 Oxford University Press 2023-10-05 /pmc/articles/PMC10553909/ http://dx.doi.org/10.1210/jendso/bvad114.983 Text en © The Author(s) 2023. Published by Oxford University Press on behalf of the Endocrine Society. https://creativecommons.org/licenses/by-nc-nd/4.0/This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (https://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reproduction and distribution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contact journals.permissions@oup.com
spellingShingle Diabetes And Glucose Metabolism
Wu, Gloria
Zhao, Weichen
Cheng, Emily
Wong, Adrial
Gim, Heidi
Zhang, Youcheng
Dorado, Perla
Mansubi, Nikki
SAT118 The "Intelligent" Chatbot: Can It Answer Questions About Health And Health Disparities?
title SAT118 The "Intelligent" Chatbot: Can It Answer Questions About Health And Health Disparities?
title_full SAT118 The "Intelligent" Chatbot: Can It Answer Questions About Health And Health Disparities?
title_fullStr SAT118 The "Intelligent" Chatbot: Can It Answer Questions About Health And Health Disparities?
title_full_unstemmed SAT118 The "Intelligent" Chatbot: Can It Answer Questions About Health And Health Disparities?
title_short SAT118 The "Intelligent" Chatbot: Can It Answer Questions About Health And Health Disparities?
title_sort sat118 the "intelligent" chatbot: can it answer questions about health and health disparities?
topic Diabetes And Glucose Metabolism
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10553909/
http://dx.doi.org/10.1210/jendso/bvad114.983
work_keys_str_mv AT wugloria sat118theintelligentchatbotcanitanswerquestionsabouthealthandhealthdisparities
AT zhaoweichen sat118theintelligentchatbotcanitanswerquestionsabouthealthandhealthdisparities
AT chengemily sat118theintelligentchatbotcanitanswerquestionsabouthealthandhealthdisparities
AT wongadrial sat118theintelligentchatbotcanitanswerquestionsabouthealthandhealthdisparities
AT gimheidi sat118theintelligentchatbotcanitanswerquestionsabouthealthandhealthdisparities
AT zhangyoucheng sat118theintelligentchatbotcanitanswerquestionsabouthealthandhealthdisparities
AT doradoperla sat118theintelligentchatbotcanitanswerquestionsabouthealthandhealthdisparities
AT mansubinikki sat118theintelligentchatbotcanitanswerquestionsabouthealthandhealthdisparities