Comparison of emergency medicine specialist, cardiologist, and chat-GPT in electrocardiography assessment

Artificial intelligence (AI) is a term that first emerged in 1955, defining the utilization of a computer to model intelligent behavior with minimal human intervention, also referred to as machine learning [1]. Since its initial definition, AI has been employed in a variety of different models. ChatGPT, an AI robot developed by OpenAI, was introduced in November 2022. Initially released with the GPT-3 model, it currently operates on the most recent version, GPT-4. GPT-4 has been evaluated in numerous challenging international exams, achieving results with an accuracy rate of 43–99% [2]. Concurrently, a significant number of studies related to the field of medicine are being conducted with ChatGPT, with cardiovascular diseases being one of these areas of research [3].

Electrocardiography (ECG) is an electrophysiological method that has been used for many years in the diagnosis and treatment of cardiac arrhythmias and diseases. Today, AI is employed to assist in ECG evaluations [4]. AI-supported methods can detect arrhythmias with an accuracy rate of up to 99% [5]. In one study, an AI-supported program classifying ECG rhythms demonstrated superiority over internal medicine physicians, emergency medicine physicians, and cardiologists [6]. AI has also been shown to perform similarly to cardiologists in other cardiac diagnoses in another study [7]. However, there is no existing study demonstrating the extent of GPT-4's capabilities related to ECG evaluations. Although GPT-4 is the latest product from OpenAI, its image evaluation feature has not yet been made available to users. Nevertheless, it has been observed to assist in diagnosis when features observed in ECG are defined. There is no available study on the level of GPT-4 in proceeding with diagnosis based on ECG data, and its level of competence in this area is unknown.

The aim of our study is to evaluate the diagnostic assistance accuracy of GPT-4 when provided with ECG data and to compare the diagnostic accuracies of three groups: GPT-4, emergency medicine specialists, and cardiologists.

留言 (0)

沒有登入
gif