Monitoring emerging drug trends: Psychometrics and validity in earlier warning systems

Within the drug trend monitoring arena, there is an increasing emphasis on the need for more timely identification and reporting of new patterns in drug consumption for policy-making purposes. However, there are documented problems with establishing the reliability and validity of results from emerging drug trend monitoring systems (EDTMS), which tend to be multi-indicator and use mixed methods. The aim of this article is to present a standardised and sequential approach to EDTMS development and refinement, i.e. grounded in the key elements of psychometrics, and illustrate its application using an established city level EDTMS. A five-step process is presented and exemplified, incorporating: (1) theoretical conceptualisation of the construct to be measured; (2) score construction; (3) weighting of indicators; (4) exploration of the prospect of categories (subscales); and (5) checking for external validity. The practical application of these validity enhancing stages are demonstrated using the Bergen Earlier Warning System. For non-traditional systems operating in a fast changing environment, an iterative review and refinement process (rather than a standardised system or instrument) has clear benefits, and can be adopted to enhance validity in existing EDTMS, or be used in the development of new models.