目的：为规范药品抽样工作，解决药品抽检中抽样点位缺乏统筹规划管理的问题，提出可量化的抽样评价指标，提升监管效能，促进药品监管信息化建设。方法：通过GIS对北京市海淀区2019年至2022 年药品抽样信息采集分析。利用平均最近邻指数分析法、核密度估计法、空间缓冲区分析法、地理联系率等分析海淀区药品抽样点位的空间分布情况、覆盖情况及影响因素。结果：海淀区药品抽样点位分布呈聚集特点。与海淀区总体可抽样点位相比，克服了部分聚集性，但近年来聚集程度仍有所增加。抽样点位分布表现出沿商业中心、交通要道、综合居民用地分布的特征，在总可抽样点位“三核”“多点” 连带、东南聚集、西北稀疏的地理格局上，逐年变化，呈现出“多核”“南移”“两极”的变化态势。 以覆盖率为抽样评价指标可见，海淀区抽样点位数量覆盖率4年达35.74%，去除重复点位后为21.28%， 年新增覆盖率在3.58%~4.28%，每年逐步增长但增速较为缓慢。各街镇点位数量选取较为平衡，仅个别街道表现出新增覆盖率不足。抽样点位服务面积覆盖率总体趋势表现为与街道行政区域面积呈负相关， 80%以上的街道可实现4年监管周期内对点位服务范围全覆盖，40%以上街道的全覆盖周期可缩短至2 年。但不同街道面积覆盖率差异较大。与抽样点位分布联系密切的因素包括人口密度、住宅区、路网密度及停车区位因素。结论：提出应区分不同分布密度特点区域采取不同抽样原则。点位密集区动态选择重点监管方向，点位分散区加大抽样点位选取密度。利用抽样点位覆盖率作为监管抽样的评价指标。通过GIS对点位的统筹管理规划，平衡抽样实施性与科学性，提升监管效能及数字信息化水平。

Objective: To standardize the drug sampling work, address the problem of lack of overall planning and management of sampling sites in drug inspection, propose quantifiable sampling evaluation indicators, improve regulatory efficiency, and promote the information construction of drug regulatory. Methods: Drug sampling information from 2019 to 2022 in Haidian District at Beijing was collected and analyzed through GIS. The spatial distribution, coverage and influencing factors of drug sampling points in Haidian District were analyzed by means of average nearest neighbor index analysis, kernel density estimation method, spatial buff er zone analysis and geographic linkage rate. Results: The distribution of drug sampling points in Haidian District was characterized by aggregation. Compared with the overall sampling points in Haidian District, some aggregation was overcome, but the degree of aggregation still increased in recent years. The distribution of sampling points showed the characteristics of distribution along commercial centers, major traffic roads and comprehensive residential land, and the geographical pattern of "three cores" and "multiple points", gathering in the southeast and sparse in the northwest. The total sampling points has changed year by year, showing a changing trend of "multiple cores", "southward shift" and "two poles". As seen by the coverage rate as the sampling evaluation index, the coverage rate of sampled points in Haidian District reached 35.74% in four years, 21.28% after removing duplicate points, and the annual increase coverage rate was from 3.58% to 4.28%, which gradually increased every year, but the growth rate was relatively slow. The number of points in each street and town was selected in a balanced manner, and only individual streets showed a lack of new coverage. The overall trend of sampling point service area coverage was negatively correlated with the administrative street area, more than 80% of the street could achieve full coverage of the point service area within four years of the regulatory cycle, and more than 40% of the street full coverage cycle could be shortened to two years. However, the area coverage rate of diff erent streets varies greatly. Factors closely associated with the distribution of sampled points include population density, residential areas, road network density and parking area factors. Conclusion: It is proposed that different sampling principles should be adopted to distinguish different distribution density characteristics regions. The dense point area dynamically selects the key supervision direction, and the dispersed point area increases the sampling point selection density. The sampling point coverage rate is used as the evaluation index of regulatory sampling. Through the GIS point of the overall management planning, the sampling implementation and scientifi c is balanced, and the eff ectiveness of supervision and digital information level is improved.