Skip Page Header

Home > Use of hierarchical cluster analysis to classify prisons in Ireland into mutually exclusive drug-use risk categories.

Codd, Mary and Mehegan, J and Kelleher, Cecily and Drummond, Anne (2016) Use of hierarchical cluster analysis to classify prisons in Ireland into mutually exclusive drug-use risk categories. Drugs: Education Prevention and Policy , 23 , (2) , pp. 93-98.


Aims: The aim of this paper is to propose a methodology for classifying prisons into distinct drug-use risk categories based on prisoners’ use of specified prescription and illegal drugs. The categories of prisons thus defined have particular relevance for the planning of drug-related interventions, education and services in Irish prisons. The methodology may be transferable to, and of value in, prison systems elsewhere.

Methods: From the 14 prisons in Ireland, 824 randomly selected prisoners completed a self-administered questionnaire on lifetime, last year and last month use of specified prescription and non-prescription drugs. Questions were derived from the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) and other sources. Oral fluid samples were collected for verification of recent drug use. The EMCDDA definition of “recent problematic drug use” was used in post hoc hierarchical cluster analysis to define categories of prisons with similar drug-use patterns.

Findings: Four categories of prisons, designated “low”, “medium”, “high” and “very high” drug use, were identified. While the prisons comprising each category were known, drug use patterns in individual prisons were not disclosed because of concerns about prisoner confidentiality in a small jurisdiction. The clusters identified were used to contextualise subsequent analyses of drug use, the availability and use of harm-reduction services and differential service requirements by prison category.

Conclusions: The methodology described allows for identification of prison categories by drug-use patterns. With restrictions on disclosure of information by individual prison, it provides more valuable information than would derive from data on all prisons combined.

Click here to request a copy of this literature (must be logged in)

Repository Staff Only: item control page