This research employs recent analytical advances to further our understanding of the influences and profiles of treated drug misuse in the Republic of Ireland, through latent variable mixture modelling Data: The National Drug Treatment Reporting System (NDTRS) an epidemiological database of treated drug misuse in Ireland, for the year 2000, was utilised for this research (n=6994).

Research Design: Recent methodological and software advances within latent class modelling have greatly reduced many of the statistical challenges of working with this type of data. The analytical components included latent variable mixture modelling, multinomial logistic regression using posterior probabilities ('training') and multilevel modelling.

The Findings: Latent class analysis of drug consumption behaviours uncovered 6 distinct profiles. The drug misuse classes were 'Recreational' (n=977, 14%), 'Experimental' (n=563, 8%). 'Multiple-combination' (n=1018, 15%), 'Methadone' (n=977, 14%), 'Benzodiazepine' (n=231, 3%), and 'Heroin' (n=3292, 47%). The unique characteristics, together with protective and risk factors were explored. The study highlighted the lack of consistency in allocating people within the same profile into the same treatment regimen and then explored factors that influence this treatment - profile mismatch.

Conclusions: This research highlighted the need for drug treatment programs to be tailored to reflect the characteristics of the individual, rather than having inflexible regimen within which all individuals are placed. Heroin dependence dominates the drug treatment landscape in Ireland, and the overwhelming focus of drug treatment resources are designed to tackle heroin misuse. This may result in Recreational, Experimental, Multi-combination drug users being placed within mainstream drug treatment settings without suitably tailored treatment programs