Evaluating the Potential and Accuracy of ChatGPT-3.5 and 4.0 in Medical Licensing and In-Training Examinations: Systematic Review and Meta-Analysis
Bibliographic record
Abstract
BACKGROUND: Artificial intelligence (AI) has significantly impacted health care, medicine, and radiology, offering personalized treatment plans, simplified workflows, and informed clinical decisions. ChatGPT (OpenAI), a conversational AI model, has revolutionized health care and medical education by simulating clinical scenarios and improving communication skills. However, inconsistent performance across medical licensing examinations and variability between countries and specialties highlight the need for further research on contextual factors influencing AI accuracy and exploring its potential to enhance technical proficiency and soft skills, making AI a reliable tool in patient care and medical education. OBJECTIVE: This systematic review aims to evaluate and compare the accuracy and potential of ChatGPT-3.5 and 4.0 in medical licensing and in-training residency examinations across various countries and specialties. METHODS: A systematic review and meta-analysis were conducted, adhering to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Data were collected from multiple reputable databases (Scopus, PubMed, JMIR Publications, Elsevier, BMJ, and Wiley Online Library), focusing on studies published from January 2023 to July 2024. Analysis specifically targeted research assessing ChatGPT's efficacy in medical licensing exams, excluding studies not related to this focus or published in languages other than English. Ultimately, 53 studies were included, providing a robust dataset for comparing the accuracy rates of ChatGPT-3.5 and 4.0. RESULTS: ChatGPT-4 outperformed ChatGPT-3.5 in medical licensing exams, achieving a pooled accuracy of 81.8%, compared to ChatGPT-3.5's 60.8%. In in-training residency exams, ChatGPT-4 achieved an accuracy rate of 72.2%, compared to 57.7% for ChatGPT-3.5. The forest plot presented a risk ratio of 1.36 (95% CI 1.30-1.43), demonstrating that ChatGPT-4 was 36% more likely to provide correct answers than ChatGPT-3.5 across both medical licensing and residency exams. These results indicate that ChatGPT-4 significantly outperforms ChatGPT-3.5, but the performance advantage varies depending on the exam type. This highlights the importance of targeted improvements and further research to optimize ChatGPT-4's performance in specific educational and clinical settings. CONCLUSIONS: ChatGPT-4.0 and 3.5 show promising results in enhancing medical education and supporting clinical decision-making, but they cannot replace the comprehensive skill set required for effective medical practice. Future research should focus on improving AI's capabilities in interpreting complex clinical data and enhancing its reliability as an educational resource.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.015 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".