Appropriateness of ChatGPT in answering heart failure related questions

Abstract

Background Heart failure requires complex management with increased patient knowledge shown to improve outcomes. The large language model (LLM), Chat Generative Pre-Trained Transformer (ChatGPT), may be a useful supplemental resource of information for patients with heart failure. Methods Responses produced by GPT-3.5 and GPT-4 to 107 frequently asked heart failure-related questions were graded by two reviewers board-certified in cardiology, with differences resolved by a third reviewer. The reproducibility and accuracy between GPT-3.5 and GPT-4 were compared for questions involving basic knowledge, management, prognosis, procedures, and support. Results GPT-4 displayed a greater proportion of comprehensive knowledge for the categories of basic knowledge and management, while GPT-3.5 performed better in the other category (prognosis, procedures, and support) (94.1% vs 64.7%). There were 2 total responses (1.9%) graded as some correct and incorrect for GPT-3.5, while no GPT-4 responses received a grade of some correct and incorrect or completely incorrect. Both models provided highly reproducible responses, with GPT-3.5 scoring above 94% in every category and GPT-4 with 100% for all answers. Conclusions Both GPT-3.5 and GPT-4 answered the majority of heart failure-related questions accurately and reliably, with GPT-4 displaying superior performance overall. ChatGPT may lead to better outcomes in patients with heart failure by providing health education.

Systematic Review of Large Language Models for Patient Care: Current Applications and Challenges

“…One study (n=1/89, 1.1%) was published in 2022 24 , 84 (n=84/89, 94.4%) in 2023 13,25–107 , and 4 (n=4/89, 4.5%) in 2024 108–111 (all of which were peer-reviewed publications of preprints published in 2023). Most studies were quantitative non-randomized (n=84/89, 94.4%) 13,25–27,29–101,103,104,106,107,109–111 , 4 (n=4/89, 4.5%) 28,102,105,108 had a qualitative study design, and one (n=1/89, 1.1%) 24 was quantitative randomized according to the MMAT 2018 criteria. However, the LLM outputs were often first analyzed quantitatively but followed by a qualitative analysis of certain responses.…”

“…The authors were primarily affiliated with institutions in the United States (n=47 of 122 different countries identified per publication, 38.5%), followed by Germany (n=11/122, 9%), Turkey (n=7/122, 5.7%), the United Kingdom (n=6/122, 4.9%), China/Australia/Italy (n=5/122, 4.1%, respectively), and 24 (n=36/122, 29.5%) other countries. Most studies examined one or more applications based on the GPT-3.5 architecture (n=66 of 124 different LLMs examined per study, 53.2%) 13,26–29,31–34,36–40,42–49,52–54,56–61,63,65–67,71,72,74,75,77,78,81–89,91,92,94,95,97–100,102–104,106–109,111 , followed by GPT-4 (n=33/124, 26.6%) 13,25,27,29,30,34–36,41,43,50,51,54,55,58,61,64,68–70,74,76,79–81,83,87,89,90,93,96,98,99,101,105 , Bard (n=10/124, 8.1%; now known as Gemini) 33,48,49,55,73,74,80,87,94,99 , Bing Chat (n=7/124, 5.7%; now Microsoft Copilot) 49,51,55,73,94,99,110 , and other applications based on Bidirectional Encoder Representations from Transformers (BERT; n=4/124, 3...…”

“…Most studies examined one or more applications based on the GPT-3.5 architecture (n=66 of 124 different LLMs examined per study, 53.2%) 13,26–29,31–34,36–40,42–49,52–54,56–61,63,65–67,71,72,74,75,77,78,81–89,91,92,94,95,97–100,102–104,106–109,111 , followed by GPT-4 (n=33/124, 26.6%) 13,25,27,29,30,34–36,41,43,50,51,54,55,58,61,64,68–70,74,76,79–81,83,87,89,90,93,96,98,99,101,105 , Bard (n=10/124, 8.1%; now known as Gemini) 33,48,49,55,73,74,80,87,94,99 , Bing Chat (n=7/124, 5.7%; now Microsoft Copilot) 49,51,55,73,94,99,110 , and other applications based on Bidirectional Encoder Representations from Transformers (BERT; n=4/124, 3.2%) 13,83,84 , Large Language Model Meta-AI (LLaMA; n=3/124, 2.4%) 55 , or Claude by Anthropic (n=1/124, 0.8%) 55 . The majority of applications were p...…”

“…In terms of design limitations, many authors noted the limitation that LLMs are not optimized for medical use (n=46/89, 51.7%) 13,26,28,34,35,37–39,46,49,50,54–59,61,62,65,66,68,70,71,79–81,83–85,88,91,93–98,100–107,109 , including implicit knowledge/lack of clinical context (n=13/89, 14.6%) 28,39,46,66,71,79,81,83–85,98,103 , limitations in clinical reasoning (n=7/89, 7.9%) 55,84,95,102–105 , limitations in medical image processing/production (n=5/89, 5.6%) 37,55,91,106,107 , and misunderstanding of medical information and terms by the model (n=7/89, 7.9%) 28,38,39,59,62,65,97 . In addition, data-related limitations were identified, including limited access to data on the internet (n=22/89, 24.7%) 38,39,41,43,54–57,59,60,64,76,79,82–84,88,91,94,96,104,109 , the undisclosed origin of training data (n=36/89, 40.5%) 25,26,29,30,32,34,36,37,40,46,…”

“…The evaluation of limitations in output data yielded 7 second-order codes concerning the non-reproducibility (n=38/89, 42.7%) 28,29,34,38,39,41,43,45,46,49,54–61,64,65,71–73,76,80,82,83,85,90,91,94,96,98,99,101,103–105 , non-comprehensiveness (n=78/89, 87.6%) 13,25,26,28–30,32–44,46,48–62,64,65,67–79,81–98,100,102–107,109–111 , incorrectness (n=78/89, 87.6%) 13,25–44,46,49–52,54–62,64–66,69–79,81–85,87–107,109–111 , (un-)safeness (n=39/89, 43.8%) 28,30,35,37,39,40,42–44,46,50,51,57–60,62,64,65,69,70,73,74,76,78–80,82,84,85,91…”