Full
Outline
To evaluate the efficacy and application potential of different large language models in health education consultation for stroke patients.
Using 21 common clinical questions of stroke and 2 sets of typical clinical scenario simulation cases,three rounds of tests were conducted on DeepSeek R1,ERNIE Bot X1,ChatGPT-4o,and Claude 3.7 Sonnet large language models.Three neurology department chief physicians were selected to evaluate the models from five dimensions:accuracy,comprehensiveness,understandability,humanistic care,and case analysis ability.
The overall performance of all models was good.DeepSeek R1 had the highest total score, followed by ChatGPT-4o,Claude 3.7 Sonnet, and ERNIE Bot X1.DeepSeek R1 and ChatGPT-4o performed outstandingly in complex clinical reasoning,evidence-based medical knowledge transmission,humanistic care,and case analysis.ERNIE Bot X1 had an advantage in the adaptability to the Chinese context.
Large language models showed positive auxiliary application value in the field of stroke health education.However,the clinical review mechanism needs to be combined to enhance the reliability of the information.The clinical application of these models could be promoted by optimizing the transparency of the algorithms and the adaptability to the local context.
33
14
| 科 Family | 属数 Number of genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) | 属 Genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) |
|---|---|---|---|---|---|---|
| 鹅膏菌科Amanitaceae | 2 | 11 | 5.26 | 鹅膏菌属 Amanita | 10 | 4.78 |
| 小菇科 Mycenaceae | 2 | 12 | 5.74 | 丝盖伞属 Inocybe | 5 | 2.39 |
| 多孔菌科 Polyporaceae | 8 | 14 | 6.70 | 蜡蘑属 Laccaria | 5 | 2.39 |
| 红菇科 Russulaceae | 3 | 23 | 11.00 | 小皮伞属 Marasmius | 6 | 2.87 |
| 小菇属 Mycena | 11 | 5.26 | ||||
| 光柄菇属 Pluteus | 5 | 2.39 | ||||
| 红菇属 Russula | 17 | 8.13 | ||||
| 栓菌属 Trametes | 5 | 2.39 |
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
Copyright © 2025 China Association for Science and Technology. All rights reserved. For all open access content, the relevant licensing terms apply.
Sponsored by the Office of the Leading Group for Cybersecurity and Informatization of CAST, and supported by Science and Technology Review Publishing House
