Objective: The aim of this research is to demonstrate how the use of hierarchical cluster analysis on 366 municipalities and other minor entities (parishes) of Venezuela, could be useful to consider regional differences and similarities between territorial entities when designing national public policies of Water, Sanitation and Hygiene (WASH) based on evidence. Methods and results: Consider data from various sources to characterize the population of Venezuela through their territorial entities. Select variables at the level of the territorial entities to cover demographic characteristics, mortality and nutrition, coverage of reliable water and sanitation services, access to education, and access to information and communication technologies. Classify the territorial entities into a limited number of mutually exclusive groups using hierarchical clustering techniques and based on proximity in the multi-dimensional space. Adjust of assignments, reallocating some entities into a different group based on the specialists' opinion about its hierarchy in the cities regional system and its geographic location. Define an indicator to verify the consistency of the groups built. Conduct a statistical analysis to confirm separation of the groups. Demonstrate the utility of the results with some examples of common analysis when building a sanitary public policy, using seven distinct groups of recommendations depending on each cluster. Conclusions: Cluster analysis can be a useful method to analyse relevant differences between territorial entities when designing national public policies based on evidence.

翻译：摘要：目的：本研究旨在展示如何通过对委内瑞拉366个市及其他次级实体（教区）采用层次聚类分析，在基于证据设计国家水、环境卫生与个人卫生（WASH）公共政策时，有效考量区域差异及领土实体间的相似性。方法与结果：整合多源数据以表征委内瑞拉各领土实体的人口特征。选取涵盖人口统计特征、死亡率与营养状况、可靠饮用水及卫生设施覆盖率、教育可及性、信息与通信技术可获得性等层面的变量。运用层次聚类技术，基于多维空间中的邻近性将领土实体划分为有限数量的互斥组群。结合专家对城市区域体系等级及地理位置的判断，调整部分实体的组群归属。构建指标验证组群一致性，并通过统计分析确认组群间差异。以构建卫生公共政策时的典型分析为例，依据七个不同的聚类组别提出针对性建议，证明该方法的实用性。结论：聚类分析可为基于证据设计国家公共政策时，解析领土实体间的关键差异提供有效方法论。