Optimizing ChatGPT's performance in hypertension care: Correspondence

Abstract

Dear Editor,

We would like to discuss “Enhancing clinical decision-making: Optimizing ChatGPT's performance in hypertension care.¹” Artificial intelligence, particularly conversational models such as OpenAI's ChatGPT, has profoundly impacted several industries, including the healthcare sector. It is a useful tool in medical research and treatment because of its capacity to analyze large volumes of data and mimic human speech. With its ability to provide recommendations and individualized health monitoring, ChatGPT holds great potential to transform patient care. For best usage in healthcare settings, there are still several areas where it falls short, such as the use of dated data and the absence of clinical judgment and individualized treatment suggestions.

One new highlight is the potential enhancements and optimizations that ChatGPT could bring to hypertension management. By summarizing guidelines, updating information, and providing decision support tools, ChatGPT can improve diagnostic accuracy, tailor treatments, and ultimately enhance patient outcomes. Additionally, as an education tool, ChatGPT can simplify complex medical topics for both patients and healthcare professionals, fostering ongoing learning and improving clinical reasoning. Research and evidence synthesis capabilities of ChatGPT can help healthcare providers make informed clinical decisions through concise overviews of the latest studies and treatments in hypertension management. The fact that ChatGPT may produce incoherent and unhelpful results is a prevalent concern. Temsah et al. stated that because of their unreliability, the present forms of ChatGPT and other Chatbots should not be employed for diagnostic or treatment purposes without human expert oversight.²

Future directions for ChatGPT in hypertension care include increasing its performance by selecting advanced models, customizing user profiles, and integrating clinical guidelines. Staying updated with research findings, creating a feedback loop for continuous improvement, and complementing professional judgment are essential steps for maximizing the utility of ChatGPT in clinical decision-making. Ethical considerations and limitations, such as privacy and security concerns, should also be addressed when using AI tools in healthcare settings. Collaborative efforts among technology developers, healthcare professionals, and patients are crucial for tailoring ChatGPT to meet the diverse needs of all stakeholders and optimizing patient care in the future.

Another obstacle to integration is the potential for bias in AI algorithms. If the data used to train the LLMs is not representative of all patient populations, it can lead to inaccurate or discriminatory outcomes. To address this issue, efforts must be made to ensure diverse and inclusive datasets are used in training AI algorithms. Regular audits and monitoring of AI systems can also help identify and correct bias in real-time.³

Hinpetch Daungsupawong 50% ideas, writing, analyzing, and approval. Viroj Wiwanitkit 50% ideas, supervision, and approval.

The authors declare no conflicts of interest.