Bridging supply and demand: Relationship between cocaine seizures and prevalence estimates of consumption.

Objectives: Drug supply and demand indicators have mainly been analysed without triangulating information on environmental drug availability and consumption. This study proposes an approach to analyse the prevalence of illicit drug consumption together with indirect indicators of drug supply. Study design: Ecological study correlating cocaine seizure data and consumption prevalence estimates at local level, using Italian provinces as the unit of analysis. Methods: The amount of cocaine seized in proportion to the number of potential consumers was computed at Italian provincial level from the 2005 database of the Italian Central Directorate for Anti-Drug Services. Indicators of consumption prevalence, at provincial level, in 15e16-year-old students from the 2005 European School Survey Project on Alcohol and Other Drugs (ESPAD) and 15e54-year-old respondents from the 2005 Italian Population Survey on Alcohol and Drugs (IPSAD) were correlated with the population-adjusted amount of total seizures and seizures of different sizes. ESPAD data were also regressed on seizure data after controlling for the prediction accrued by socio-economic variables that might cause covariation of seizure and consumption prevalence data. Regressionpredicted prevalence values were computed and correlated with the ESPAD prevalence data for the following year. Results: There was a weak general correlation between the population-adjusted total seizure and IPSAD prevalence indicators, but not ESPAD prevalence indicators. In contrast, ESPAD data were more strongly correlated with small seizures. The regression model yielded significant variance in prevalence data (18% and 23% in small and large provinces, respectively) explained by small seizures, and this remained after removing the percentage of variance explained by socio-economic factors and (particularly) education level. A moderate correlation was found between seizure-based predicted values of consumption prevalence and ESPAD values for the following year. Conclusions: Associating prevalence estimates in adolescents and cocaine seizure data at local level, after ‘purifying’ the data from trafficking- and wholesale-related large seizures,may represent a starting point for a geo-referenced, seizure-based analysis of initiation into cocaine consumption. This approach may support the collaboration between enforcement and health systems, helping to establish empirical regularities to be translated into useful information for local planners of drug prevention policies in areas at greater risk of increasing incidence.