目的：构建儿童合理用药智能辅助决策系统，为儿童用药安全提供有力参考。方法：构建知识库、规则库，根据药品说明书、药物属性、循证医学资料等设置可视化、自定义的儿童用药规则，通过对系统应用前后的效果进行对比，评估其效能。结果：构建了 400 个儿童常用药品和 8 000 余条儿童合理用药自定义规则，系统应用后，共审核了 564 287 张儿科处方，系统对所有类型的处方错误警示率均高于人工审查，RR 为 11.26（95% CI：10.73 ～ 11.82）。与人工审核相比，系统在药物相互作用、用法用量和联合用药不适宜方面的警示率显著提高，具体 RR 值分别为 35.49（95% CI：28.33 ～ 44.45）、36.55（95% CI：31.61 ～ 41.49）和 18.89（95%CI：11.30 ～ 31.58）。在临床医师接受率方面，人工审核与系统审核之间并未表现出显著差异。结论：儿童合理用药智能辅助决策系统能够显著提升药师的工作效率，优化合理用药，进而推动医疗服务质量的整体改善。

Objective To construct an intelligent assistive decision-making system for children's rational medication, and to provide a strong reference for the safety of children's medication. Methods By building a knowledge base and rule base, visualized and customized child-specific medication regulations were set up according to drug instructions, drug attributes and evidence-based medical data. The effectiveness of the system was evaluated by comparing the effects before and after its application. Results Nearly 400 commonly used drugs for children and more than 8,000 customized rules for children's rational medication were constructed. After the system was applied, a total of 564,287 pediatric prescriptions were reviewed. The system had a higher error alert rate for all types of prescriptions than manual review, with an RR of 11.26 (95% CI: 10.73 to 11.82). Compared with manual prescription review by pharmacists, the alert rates for drug interactions, dosage, and inappropriate co-prescribing were significantly improved,  with specific RR values of 35.49 (95% CI: 28.33 to 44.45), 36.55 (95% CI: 31.61 to 41.49), and 18.89 (95% CI: 11.30 to 31.58), respectively. However, there was no significant difference in clinician acceptance rate between manual and systematic reviews by pharmacists. Conclusion The intelligent assistive decision-making system for pediatric rational medication can significantly improve the work efficiency of pharmacists, optimize rational medication practices, and then promote the overall improvement of medical service quality.