Holohean E J, Banks S M, Maddy B A
New York State Office of Mental Health, Albany 12229, USA.
J Ment Health Adm. 1995 Spring;22(2):177-88. doi: 10.1007/BF02518757.
The identification and categorization of similar objects or individuals into groups is a universal theme that permeates science and provides a conceptual framework to facilitate comprehension. For mental health administrators, a sound methodology to group individuals in a meaningful way would prove useful in the areas of general system understanding, as well as staffing, program planning and evaluation, and service system research, to name a few. Cluster analysis is a set of techniques that approach this grouping process empirically. Within the context of a large psychiatric hospital system, both the methodological processes involved in the application of cluster analysis and the resulting utility of such an analysis are discussed. Issues fundamental to the understanding of such a system are addressed. Special emphasis is placed on methodological issues regarding the application of cluster analytic techniques, which have left such techniques open for criticism. The value of such analyses, when used appropriately, is illustrated by the development of a stable, five-group typology of psychiatric hospital residents whose group characteristics are particularly germane to service system understanding. The implications of such a model for administration, planning, and research in a psychiatric hospital system are also addressed.
将相似的物体或个体识别并分类成组是一个普遍的主题，它贯穿于科学领域，并提供一个概念框架以促进理解。对于心理健康管理人员而言，一种能以有意义的方式对个体进行分组的合理方法，在诸如一般系统理解、人员配备、项目规划与评估以及服务系统研究等领域将被证明是有用的。聚类分析是一组以实证方式处理这种分组过程的技术。在一个大型精神病医院系统的背景下，本文将讨论聚类分析应用中所涉及的方法过程以及这种分析所产生的效用。还将探讨理解这样一个系统的基本问题。特别强调了与聚类分析技术应用相关的方法问题，这些问题使得此类技术容易受到批评。通过开发一种稳定的、将精神病院住院患者分为五组的类型学，说明了此类分析在恰当使用时的价值，其组特征与服务系统理解特别相关。本文还探讨了这种模型对精神病医院系统管理、规划和研究的影响。
