In-depth analysis of ChatGPT's performance based on specific signaling words and phrases in the question stem of 2377 USMLE step 1 style questions.

ChatGPT has garnered attention as a multifaceted AI chatbot with potential applications in medicine. Despite intriguing preliminary findings in areas such as clinical management and patient education, there remains a substantial knowledge gap in comprehensively understanding the chances and limitations of ChatGPT's capabilities, especially in medical test-taking and education. A total of n = 2,729 USMLE Step 1 practice questions were extracted from the Amboss question bank. After excluding 352 image-based questions, a total of 2,377 text-based questions were further categorized and entered manually into ChatGPT, and its responses were recorded. ChatGPT's overall performance was analyzed based on question difficulty, category, and content with regards to specific signal words and phrases. ChatGPT achieved an overall accuracy rate of 55.8% in a total number of n = 2,377 USMLE Step 1 preparation questions obtained from the Amboss online question bank. It demonstrated a significant inverse correlation between question difficulty and performance with rs = -0.306; p < 0.001, maintaining comparable accuracy to the human user peer group across different levels of question difficulty. Notably, ChatGPT outperformed in serology-related questions (61.1% vs. 53.8%; p = 0.005) but struggled with ECG-related content (42.9% vs. 55.6%; p = 0.021). ChatGPT achieved statistically significant worse performances in pathophysiology-related question stems. (Signal phrase = "what is the most likely/probable cause"). ChatGPT performed consistent across various question categories and difficulty levels. These findings emphasize the need for further investigations to explore the potential and limitations of ChatGPT in medical examination and education.

Knoedler L ，Knoedler S ，Hoch CC ，Prantl L ，Frank K ，Soiderer L ，Cotofana S ，Dorafshar AH ，Schenck T ，Vollbach F ，Sofo G ，Alfertshofer M ... -  《Scientific Reports》

Pure Wisdom or Potemkin Villages? A Comparison of ChatGPT 3.5 and ChatGPT 4 on USMLE Step 3 Style Questions: Quantitative Analysis.

Knoedler L ，Alfertshofer M ，Knoedler S ，Hoch CC ，Funk PF ，Cotofana S ，Maheta B ，Frank K ，Brébant V ，Prantl L ，Lamby P ... -  《-》

How Does ChatGPT Perform on the United States Medical Licensing Examination (USMLE)? The Implications of Large Language Models for Medical Education and Knowledge Assessment.

Chat Generative Pre-trained Transformer (ChatGPT) is a 175-billion-parameter natural language processing model that can generate conversation-style responses to user input. This study aimed to evaluate the performance of ChatGPT on questions within the scope of the United States Medical Licensing Examination (USMLE) Step 1 and Step 2 exams, as well as to analyze responses for user interpretability. We used 2 sets of multiple-choice questions to evaluate ChatGPT's performance, each with questions pertaining to Step 1 and Step 2. The first set was derived from AMBOSS, a commonly used question bank for medical students, which also provides statistics on question difficulty and the performance on an exam relative to the user base. The second set was the National Board of Medical Examiners (NBME) free 120 questions. ChatGPT's performance was compared to 2 other large language models, GPT-3 and InstructGPT. The text output of each ChatGPT response was evaluated across 3 qualitative metrics: logical justification of the answer selected, presence of information internal to the question, and presence of information external to the question. Of the 4 data sets, AMBOSS-Step1, AMBOSS-Step2, NBME-Free-Step1, and NBME-Free-Step2, ChatGPT achieved accuracies of 44% (44/100), 42% (42/100), 64.4% (56/87), and 57.8% (59/102), respectively. ChatGPT outperformed InstructGPT by 8.15% on average across all data sets, and GPT-3 performed similarly to random chance. The model demonstrated a significant decrease in performance as question difficulty increased (P=.01) within the AMBOSS-Step1 data set. We found that logical justification for ChatGPT's answer selection was present in 100% of outputs of the NBME data sets. Internal information to the question was present in 96.8% (183/189) of all questions. The presence of information external to the question was 44.5% and 27% lower for incorrect answers relative to correct answers on the NBME-Free-Step1 (P<.001) and NBME-Free-Step2 (P=.001) data sets, respectively. ChatGPT marks a significant improvement in natural language processing models on the tasks of medical question answering. By performing at a greater than 60% threshold on the NBME-Free-Step-1 data set, we show that the model achieves the equivalent of a passing score for a third-year medical student. Additionally, we highlight ChatGPT's capacity to provide logic and informational context across the majority of answers. These facts taken together make a compelling case for the potential applications of ChatGPT as an interactive medical education tool to support learning.

Gilson A ，Safranek CW ，Huang T ，Socrates V ，Chi L ，Taylor RA ，Chartash D ... -  《-》

Sailing the Seven Seas: A Multinational Comparison of ChatGPT's Performance on Medical Licensing Examinations.

Alfertshofer M ，Hoch CC ，Funk PF ，Hollmann K ，Wollenberg B ，Knoedler S ，Knoedler L ... -  《-》

Exploring the Performance of ChatGPT Versions 3.5, 4, and 4 With Vision in the Chilean Medical Licensing Examination: Observational Study.

The deployment of OpenAI's ChatGPT-3.5 and its subsequent versions, ChatGPT-4 and ChatGPT-4 With Vision (4V; also known as "GPT-4 Turbo With Vision"), has notably influenced the medical field. Having demonstrated remarkable performance in medical examinations globally, these models show potential for educational applications. However, their effectiveness in non-English contexts, particularly in Chile's medical licensing examinations-a critical step for medical practitioners in Chile-is less explored. This gap highlights the need to evaluate ChatGPT's adaptability to diverse linguistic and cultural contexts. This study aims to evaluate the performance of ChatGPT versions 3.5, 4, and 4V in the EUNACOM (Examen Único Nacional de Conocimientos de Medicina), a major medical examination in Chile. Three official practice drills (540 questions) from the University of Chile, mirroring the EUNACOM's structure and difficulty, were used to test ChatGPT versions 3.5, 4, and 4V. The 3 ChatGPT versions were provided 3 attempts for each drill. Responses to questions during each attempt were systematically categorized and analyzed to assess their accuracy rate. All versions of ChatGPT passed the EUNACOM drills. Specifically, versions 4 and 4V outperformed version 3.5, achieving average accuracy rates of 79.32% and 78.83%, respectively, compared to 57.53% for version 3.5 (P<.001). Version 4V, however, did not outperform version 4 (P=.73), despite the additional visual capabilities. We also evaluated ChatGPT's performance in different medical areas of the EUNACOM and found that versions 4 and 4V consistently outperformed version 3.5. Across the different medical areas, version 3.5 displayed the highest accuracy in psychiatry (69.84%), while versions 4 and 4V achieved the highest accuracy in surgery (90.00% and 86.11%, respectively). Versions 3.5 and 4 had the lowest performance in internal medicine (52.74% and 75.62%, respectively), while version 4V had the lowest performance in public health (74.07%). This study reveals ChatGPT's ability to pass the EUNACOM, with distinct proficiencies across versions 3.5, 4, and 4V. Notably, advancements in artificial intelligence (AI) have not significantly led to enhancements in performance on image-based questions. The variations in proficiency across medical fields suggest the need for more nuanced AI training. Additionally, the study underscores the importance of exploring innovative approaches to using AI to augment human cognition and enhance the learning process. Such advancements have the potential to significantly influence medical education, fostering not only knowledge acquisition but also the development of critical thinking and problem-solving skills among health care professionals.

Rojas M ，Rojas M ，Burgess V ，Toro-Pérez J ，Salehi S ... -  《-》