This research examined (a) whether individuals can be grouped into distinct profiles based on their drug-use behaviours, (b) how these classes differ in terms of the key predictive factor ‘age’, and (c) whether the individual’s designated drug-treatment regimen/intervention matches their drug misuse profile (profile-treatment matching).

This analysis utilised data from the National Drug Treatment Reporting System (NDTRS) database, from 6,994 individuals who received treatment for problem drug use during 2000. Latent class/profile analysis explains unobserved underlying patterns from observed categorical data; the presence/absence of drug consumption behaviour. Bayesian and Akaike Information Criteria (BIC & AIC), in association with substantive theory, were used to establish the optimal number of latent classes required to describe the patterns of reported drug taking. Multinomial logistic regression was then used to illustrate how these profiles change as a function of age. Each individual’s drug-use profile and corresponding intervention regimen (treatment) was then examined and the corresponding association between them was explored. In a series of unconditioned latent class models, a 6-class solution yielded the model of best-fit, in terms of theory and statistics.

The six distinct drug misuse profiles were, ‘Recreational’, ‘Experimental’, ‘Multiple-combination Users’, ‘Methadone Misusers’, ‘Benzodiazepine Misuser’, and ‘Heroin’ classes. Membership of a specific class was found to change dramatically as a function of age. Further, there was a poor match between an individual’s drug misuse profile and the type of treatment received. This disparity between the type of drug misuse and the treatment interventions is discussed.