Comparative Performance Analysis of Large Language Models: ChatGPT, Google Gemini, and DeepSeek in Approach to Resistant Hypertension

Background: Large Language Models (LLMs) Show Promise for Enhancing Medical Decision-Making, but Their Performance in Specialized Domains Like Hypertension Management Remains Understudied. We Evaluated Three LLMs - ChatGPT-۴, Google Gemini, and DeepSeek - Using Guideline-Based Clinical Questions to Assess Their Accuracy and Reliability. Methods: Thirty Clinical Questions (۱۵ Simple, ۱۰ Moderate, ۵ Hard Difficulty) Were Extracted From the ۲۰۲۴ ESC Guidelines for the Management of Elevated Blood Pressure and Hypertension, With Emphasis on Resistant Hypertension. A Senior Cardiologist Scored Responses Using a ۴-Point Scale: (۱) Completely Incorrect Answer, (۲) Correct Answer Containing Incorrect Information, (۳) Correct but Inadequate Answer, and (۴) Correct and Adequate Answer. We Summarized Non-Normally Distributed Data As Median With Interquartile Range (IQR) and Performed Nonparametric Analyses Using Stata ۱۸ (StataCorp LLC), Including Friedman Tests With Post Hoc Wilcoxon Signed-Rank Tests (Bonferroni-Corrected) to Compare Model Performance Across Difficulty Levels. Results: ChatGPT-۴ Achieved the Highest Median Scores for All Question Types (Hard: ۴ [IQR ۴-۴]; Moderate: ۴ [۳-۴]; Simple: ۴ [۳.۵-۴]), With Statistically Superior Performance Over Gemini and DeepSeek for Moderate Questions (p=۰.۰۳). Differences Were Nonsignificant for Hard (p=۰.۳) and Simple Questions (p=۰.۶۲), Though ChatGPT Showed Marginally Better Overall Performance (p=۰.۰۵۹). Gemini and DeepSeek Performed Comparably but Scored Lower Than ChatGPT (Median Range: ۳-۳.۵). Response Consistency (IQRs) Was Similar Across Models. Conclusion: While ChatGPT-۴ Outperformed Peers in Moderate-Difficulty Questions, the Absence of Significant Differences in Hard Questions - Coupled With the Small Sample of Hard Queries - Highlights the Need for Larger, More Balanced Datasets to Robustly Evaluate LLMs in Complex Clinical Scenarios. These Findings Underscore the Importance of Context-Specific Validation When Deploying LLMs in Medicine.

کلیدواژه ها:

LLMs ، Hypertension ، AI Validation