Predicting the Size and Characteristics of the Drug Treatment Population – Technical Methods

Aim and data specification

This analysis was based on monthly data from the National Drug Treatment Monitoring System (NDTMS) from December 2005 to November 2016. NDTMS is taken to be a comprehensive description of drug treatment provision for this period.

Looking forward for four years based on patterns observed in the preceding decade, the aim of this analysis is to estimate the size and characteristics of opiate and non-opiate treatment populations (adults only for the non-opiate population) to the end of 2020, with four specifications:

• the projected number in treatment (main models, herein)
• breakdown by age
• breakdown by duration of use (using career, herein, and opiate population only)
• breakdown by previous treatment (opiate population only)

With anticipated uncertainty in each projection (which could be increased with cross-referencing), the statistical models were computed independently. There is logical covariation between components (for example, between age and using career) but this has not been applied directly in the projections.

For the main treatment population models, a decision was taken to truncate the retrospective data to January 2011 (see ‘Reasons for truncation of the time period’ for rationale). Additional modelling was required for the non-opiate models to take into account the additional criteria to limit to adults only, as described below. All analysis was done in SPSS (version 21) using latest available data.