Innovation and Efficacy Testing of Long-Term Care Insurance Service Supervision Tools&#x02014;&#x02014;A Study Based on Intelligent Regulatory Agent

doi:10.19546/j.issn.1674-3830.2026.6.002

China Health Insurance ›› 2026, Vol. 0 ›› Issue (6) : 16-25. DOI: 10.19546/j.issn.1674-3830.2026.6.002

Innovation and Efficacy Testing of Long-Term Care Insurance Service Supervision Tools——A Study Based on Intelligent Regulatory Agent

Author information +

History +

Abstract

As the long-term care insurance (LTCI) system transitions from pilot programs to nationwide implementation, frequent violations such as “fake services, fake assessments, and fake disabilities” in care services have become prominent. Traditional supervision models face challenges including insufficient coverage, low efficiency, and poor interpretability, threatening the security of LTCI funds and the sustainable operation of the system. In response to national LTCI regulatory policy requirements, this study develops an end-cloud collaborative, real-time interpretable supervision agent for LTCI care behaviors. Through deep coupling of device-side intelligence and cloud-side large models, the agent enables a transformation in LTCI supervision from a service “ black box” to a state of “knowing both what and why”. To objectively evaluate the system’s operational capabilities, this study further constructs a benchmark for LTCI care behavior supervision, which includes a video-based benchmark test dataset, a five-dimensional system evaluation index system, and an agent-based comparison baseline, providing a basis for horizontal comparison and iterative optimization of system capabilities. The results indicate that this approach effectively addresses the issues in LTCI supervision and offers an effective technical solution and efficacy testing for intelligent LTCI supervision.

Key words

long-term care insurance / intelligent supervision / agent architecture / benchmark testing

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

Innovation and Efficacy Testing of Long-Term Care Insurance Service Supervision Tools——A Study Based on Intelligent Regulatory Agent[J]. China Health Insurance. 2026, 0(6): 16-25 https://doi.org/10.19546/j.issn.1674-3830.2026.6.002

References

[1] 新华社.国务院常务会议解读|国务院研究推进银发经济和养老服务发展有关工作[EB/OL].(2026-02-24)[2026-04-14].https://www.gov.cn/zhengce/202602/content_7059238.htm.
[2] 何奎.从“扩大试点”到“加快建立”长护险织牢民生保障网[N].上海证券报,2025-04-01(002).
[3] 陈诚诚. 长期护理保险“保基本”的理论解析与实践路径[J].社会保障评论,2026,10(02):78-91.
[4] 姚雪青,窦瀚洋,李蕊,等.守护失能群体,让长护险服务更有力[N].人民日报,2025-12-03(012).
[5] 刘晓婷,石连忠,刘嘉欣.长护险智能监管困境及优化路径研究——以浙江省杭州市为例[J].中国医疗保险,2025(12):78-85.
[6] 杨蓓蕾,梅雨前,褚泽坤.上海市智慧监管长期护理保险的探索[J].中国国情国力,2022(11):17-21.
[7] 陈佳琳.长护险全国推行,上海贡献了哪些经验[N].新民晚报,2026-03-27(004).
[8] 刘昶荣.迈向“十五五”,长护险将全面推开[N].中国青年报,2026-01-04(001).
[9] 杨芳,项晨溪.整体性治理视域下长期护理保险试点“碎片化”难题的破解路径[J].辽宁行政学院学报,2025(04):10-16.
[10] 林小芬,龚衍花,冯文,等.国内外长期护理保险发展进程及其对海南自贸港的启示[J].现代商贸工业,2024,45(08):51-54.
[11] 朱成燕,张苗,姚岚.多源流视角下我国长期护理保险制度的构建逻辑[J].中国卫生政策研究,2026,19(01):2-8.
[12] 吴新建. 长护险经办机构服务绩效评价体系构建研究[J].销售与管理,2025(32):42-44.
[13] 林奕. 长期护理保险数智化治理:理论逻辑、现实困境与制度创新[J].就业与保障,2025(11):28-30.
[14] 汪梦,陈晓丽,王鹏飞,等.社会第三方参与长期护理保险监管的必要性和可行性探讨[J].中国初级卫生保健,2024,38(03):1-4.
[15] 纪斌. 长期护理保险发展回顾与展望——以山东省青岛市为例[J].中国社会保障,2025(11):82-83.
[16] 彭天娇,张韦,朱成燕,等.央地视角下长期护理保险政策比较研究[J].中国卫生政策研究,2026,19(01):9-16.
[17] 何梓彤,陆杭同,张慧.长期护理保险全流程监管指标体系构建研究[J].中国卫生事业管理,2025,42(11):1241-1247.
[18] 李科仲,赖晓薇.筑牢长期护理保险基金安全防线[J].中国医疗保险,2023(12):6-7.
[19] 王广智. 商业保险在政策性长护险中的角色重塑[J].上海保险,2025(12):19-23.
[20] 陈晓丽,李星辉,郑嘉毅,等.上海市长期护理保险服务信息化监管存在的问题及对策探讨[J].中国初级卫生保健,2022,36(9):20-22.
[21] 于保荣,孔维政,廖森,等.长期护理保险监管机制的国内外实践与启示[J].卫生经济研究,2023,40(12):24-27+32.
[22] 朱铭来. 全面实施长期护理保险的挑战与对策[J].中国保险,2025(12):22-25.
[23] 吴息凤. 建设符合中国国情的长期护理保险制度[J].国家治理,2026(04):55-60.
[24] 张灿,朱骏侠,齐焕东,等.基于远程视频图像AI识别与定位的围填海智能监管方法及应用研究[J].海洋通报,1-7.
[25] 刘子恒,焦良葆,魏小玉,等.基于人脸关键点的交互式活体检测[J].计算机与数字工程,2025,53(05):1415-1420+1426.
[26] 张弛岳,云静,乌熙靓,等.自动驾驶中3D目标检测的多模态感知与融合方法综述[J].计算机工程与应用,1-29.
[27] 陈锋,吴欣然.药品监管领域大语言模型(LLM)一体化建设策略研究[J].中国医药导刊,2025,27(01):1-6.
[28] 陈帆,孙凯.自然语言大模型在税收监管中的应用——以非居民税收合同审核为例[J].税收经济研究,2024,29(03):38-46.
[29] SRIVASTAVA A, RASTOGI A, RAO A, et al. Beyond the imitation game: quantifying and extrapolating the capabilities of language models[EB/OL].(2023-05-12)[2026-04-21]. https://openreview.net/forum?id=uyTL5Bvosj.
[30] DOU S, ZHANG M, YIN Z, et al. CL-bench: a benchmark for context learning[EB/OL].(2026-02-03)[2026-04-14]. https://arxiv.org/abs/2602.03587.
[31] DING J, LU L, DING C, et al. MedBenchv4: a robust and scalable benchmark for evaluating Chinese medical language models, multimodal models,intelligent agents[EB/OL].(2025-11-19)[2026-04-14]. https://arxiv.org/abs/2511.14439.
[32] 曹露,许林,张宇洁,等.大语言模型在中医领域的标准化评估[J].南京中医药大学学报,2024,40(12):1383-1392.
[33] RAZZAQ K, SHAH M.Next-generation machine learning in healthcare fraud detection: current trends, challenges, and future research directions[J]. Information, 2025, 16(9): 730.
[34] 邵延华,张铎,楚红雨,等.基于深度学习的YOLO目标检测综述[J].电子与信息学报,2022,44(10):3697-3708.
[35] WEI J, WANG X, SCHUURMANS D, et al.Chain-of-thought prompting elicits reasoning in large language models[J]. Advances in neural information processing systems, 2022, 35: 24824-24837.