Objective To evaluate the practicality and feasibility of generative large language models (LLMs) in pharmaceutical practice within the reproductive and obstetrics-gynecology field, and to provide a reference for pharmacists in the rational application of artificial intelligence tools.

Methods Stratified sampling was adopted to extract outpatient prescriptions from the Guangxi Zhuang Autonomous Region Reproductive Hospital between 2024 and 2025, as well as medication consultation questions from the Standardized Manual for Medication Consultation in Pregnancy and Lactation. Additionally, discharge medication education records of the above-mentioned hospital were randomly sampled to construct a question bank. DeepSeek, ChatGPT and Grok were separately used to generate question-answering outputs. A dual-review process was performed to systematically evaluate and analyze the responses produced by large language models (LLMs).

Results In prescription review, ChatGPT achieved the highest accuracy (83.3%), followed by Grok (76.7%) and DeepSeek (73.3%). All three models showed relatively low precision (45.5%-57.1%). DeepSeek and ChatGPT both achieved 100.0% recall, whereas Grok achieved only 45%. The F1 score was best for ChatGPT (72.7%). In medication consultation, DeepSeek performed best, with an average score of 2.75 ± 0.45. In medication education, ChatGPT performed best, with an average score of 2.50 ± 0.53.

Conclusion Generative LLMs have demonstrated practical utility in pharmaceutical practice in the reproductive and obstetrics-gynecology domain. However, current models cannot yet replace the professional judgment of pharmacists. A "human-machine collaborative" model is recommended to ensure medication safety for special populations in this field.

1. 孙文希, 胡凌娟. 国内外不孕不育症现状及我国的干预策略探讨[J]. 人口与健康, 2019, (12): 19-23. [SunWX, HuLJ. Current situation of infertility at home and abroad and discussion on intervention strategies in China[J]. Population and Health, 2019, (12): 19-23.https://kns.cnki.net/kcms2/article/abstract?v=tPTW3hFh4gfj2Q91RkdLq0FkKRKyCPW3hzNLwM2-ce4dWgS B_KmMnW4e0j8TmQMwfjSyKhiHYRERqYqFItPVQbfYFEHfkkjQKv_qConDICMSaMYf4xi1LFuIXMxtcOWCgQXbjjiLnExVL04FTDlbvOnxSiKTWSpBVnwZz3kcwOlYlCuduJVKQ==&uniplatform=NZKPT&language=CHS.

2. PaudyalV, OkuyanB, HenmanMC, et al. Scope, content and quality of clinical pharmacy practice guidelines: a systematic review[J]. Int J Clin Pharm, 2024, 46(1): 56-69. DOI: 10.1007/s11096-023-01658-x.

3. ShoolS, AdimiS, Saboori AmleshiR, et al. A systematic review of large language model (LLM) evaluations in clinical medicine[J]. BMC Med Inform Decis Mak, 2025, 25(1): 117. DOI: 10.1186/s12911-025-02954-4.

4. 张微微, 唐婧, 冀召帅, 等. 人工智能在临床药学中的应用进展[J]. 药学学报, 2025, 60(10): 3045-3059.ZhangWW, TangJ, JiZS, et al. Advances in the application of artificial intelligence in clinical pharmacy [J]. Acta Pharmaceutica Sinica, 2025, 60(10): 3045-3059. DOI: 10.16438/j.0513-4870.2025-0565.

5. 费姚, 马晶晶, 钱晨月. 开展生殖-产科序贯药学门诊的SWOT分析及发展策略[J]. 中国临床药学杂志, 2025, 34(2): 138-141.FeiY, MaJJ, QianCY. SWOT analysis and development strategies on carrying out sequential pharmaceutical clinics for reproduction and obstetrics[J]. Chinese Journal of Clinical Pharmacy, 2025, 34(2): 138-141. DOI: 10.19577/j.1007-4406.2025.02.011.

6. 国家卫生健康委员会. 处方管理办法[S/OL]. (2007-02-14)[2026-05-01]. https://www.gov.cn/zhengce/2007-02/14/content_5713779.htm.

7. 冯欣. 妊娠及哺乳期用药咨询标准化手册[M]. 北京: 人民卫生出版社, 2016: 2-70.

8. 张志玲, 李强, 闫盈盈, 等. 大语言模型为药品不良反应报告者生成反馈信息的探究[J]. 临床药物治疗杂志, 2024, 22(10): 38-42.ZhangZL, LiQ, YanYY, et al. Applying large language models to generate feedback for reporters of spontaneous adverse drug reactions [J]. Clinical Medication Journal, 2024, 22(10): 38-42. DOI: 10.3969/j.issn.1672-3384.2024.10.008.

9. 牛露露, 张宏亮. 大型语言模型在医院药学中的应用——探索、总结与思考[J]. 广西医学, 2025, 47(8): 1106-1114.NiuLL, ZhangHL. Application of large language model in hospital pharmacy: a study on exploration, summary and reflection[J]. Guangxi Medical Journal, 2025, 47(8): 1106-1114. DOI: 10.11675/j.issn.0253-4304.2025.08.06.

10. 张浩严, 吕文涛, 余润泽, 等. 大语言模型研究现状[J]. 无线电工程, 2025, 55(1): 163-174.ZhangHY, LyuWT,YuRZ,et al. Research status of large language models[J]. Radio Engineering, 2025, 55(1): 163-174. DOI: 10.3969/j.issn.1003-3106.2025.01.020.

11. WangX, XiongZ, ZouK, et al. Reasoning-driven large language models in medicine: opportunities, challenges, and the road ahead[J].  Lancet Digit Health, 2026, 8(1): 100931. DOI: 10.1016/j.landig.2025.100931.

12. 马红燕, 周晓, 张碧瑶, 等. DeepSeeK大语言模型辅助处方合理性评估的实用性[J]. 医药导报, 2025, 44(12): 2057-2061.MaHY, ZhouX, ZhangBY, et al. Practicability of DeepSeek large language model-assisted prescription rationality evaluation[J]. Herald of Medicine, 2025, 44(12): 2057-2061. DOI: 10.3870/j.issn.1004-0781.2025.12.026.

13. OngJCL, JinL, ElangovanK, et al. Large language model as clinical decision support system augments medication safety in 16 clinical specialties[J]. Cell Rep Med, 2025, 6(10): 102323. DOI: 10.1016/j.xcrm.2025.102323.

14. LiL, DuP, HuangX, et al. Comparative analysis of generative artificial intelligence systems in solving clinical pharmacy problems: mixed methods study[J]. JMIR Med Inform, 2025, 13: e76128-e76128. DOI: 10.2196/76128.

15. 刘静, 孙静, 昂韦, 等. 人工智能大语言模型在用药咨询和用药教育中的多维度评估及与药师服务的比较研究[J]. 临床药物治疗杂志, 2025, 23(6): 38-44, 50.LiuJ, SunJ, AngW, et al. Multi-dimensional evaluation of artificial intelligence large language models in medication consultation and medication education and a comparative study with pharmacist services[J]. Clinical Medication Journal, 2025, 23(6): 38-44,50. DOI: 10.3969/j.issn.1672-3384.2025.06.007.

16. KrichevskyB, EngeliS, Bode-BögerSM, et al. Human vs. artificial intelligence: Physicians outperform ChatGPT in real-world pharmacotherapy counselling[J]. Br J Clin Pharmacol, 2026, 92(3): 891-902. DOI: 10.1002/bcp.70321.

17. AbanmyNO, Al-GhreimilN, AlsabhanJF, et al. Evaluating the accuracy of ChatGPT in delivering patient instructions for medications: an exploratory case study[J]. Front Artif Intell, 2025, 8: 1550591. DOI: 10.3389/frai.2025.1550591.

18. AydinS, KarabacakM, VlachosV, et al. Navigating the potential and pitfalls of large language models in patient-centered medication guidance and self-decision support[J]. Front Med (Lausanne), 2025, 12: 1527864. DOI: 10.3389/fmed.2025.1527864.

19. StanceskiK, ZhongS, ZhangX, et al. The quality and safety of using generative AI to produce patient-centred discharge instructions[J]. NPJ Digit Med, 2024, 7(1): 329. DOI: 10.1038/s41746-024-01336-w.

20. 雷珍珍, 梁艳, 张章, 等. 人工智能赋能我国药师开展药学服务模式探讨[J]. 中国药事, 2026, 40(2): 235-239.LeiZZ, LiangY, ZhangZ, et al. Exploring the application of artificial intelligence in empowering pharmacists to provide pharmaceutical services in China[J]. Chinese Pharmaceutical Affairs, 2026, 40(2): 235-239. DOI: 10.16153/j.1002-7777.2025-08-0038.