This page provides methods and definitions information for the Problem drug use statistics, which form part of the EMCDDA's Statistical Bulletin 2019.
Generic case definition
We define high-risk drug use as recurrent drug use that is causing actual harms (negative consequences, including dependence, but also other health, psychological or social problems) to the person, or is placing the person at a high probability/risk of suffering such harms.
Substance-specific case definitions
The case definitions for substance-specific prevalence estimates are detailed in the document ‘PDU revision summary’. revision summary emphasizes two aspects. First, it highlights the need for substance-specific prevalence estimates (compared to the former PDU prevalence estimate that included opioids, cocaine and/or amphetamines in a single estimate). Second, it broadens the generic case definition to include people at high riskof suffering harms. first point is useful to obtain separate estimates for high-risk opioid use (HROU) (for which OST is the recommended treatment) and people who inject drugs (PWID) (for which specific harm-reduction strategies have proved effective). The second point is useful for the inclusion of the high-risk cannabis use estimate that can be obtained under certain conditions using scales in general population surveys (GPS) and to allow certain flexibility should new substances appear on the market in the future.
The development of the PDU indicator has been intrinsically linked to the development of prevalence estimation methods and their application across Europe. Due to the stigma and marginalisation often associated with harmful use, general population surveys are known to suffer underreporting when estimating the prevalence of high-risk drug use. The recommended methods to estimate the population of high-risk drug users are indirect methods. They rely on accessible health or police registries where a proportion of the total population of users is captured in order to estimate the hidden proportion that is not captured by these registries. The multiplier method (MM), the capture-recapture method (CRC) and the single source truncated Poisson method (TP) are commonly used to estimate prevalence of hidden populations at the national and local levels. These indirect methods rely on strong assumptions – such as absence of direct referrals between data sources, homogeneity, and closed populations – that need careful consideration. Some methods (e.g. CRC, TP) require unique identifiers in order to identify overlaps between data sources or to count the number of visits per individual in a given period. Other methods (e.g. treatment multiplier) require information on the proportion of high-risk drug users who accessed a treatment service in a given period (usually obtained through a cross-sectional study conducted among a representative sample of the source population).
Data sources commonly used for the estimates include: drug treatment databases, cause-specific mortality registries, drug law offences databases, and observational studies among high-risk drug users. The PDU indicator therefore widely relies on data covered by other key epidemiological indicators and should as much as possible be interpreted in relation to these indicators. The applicability of the recommended case definitions (see PDU revision summary)is often limited by the nature of the available data sources: where frequency of use or diagnosis are not available, contacts with the treatment system or arrests for specific substances can be used as a proxy of harms. Different countries have adopted different methods depending on the available data sources, which in turn depends on the system of social and health care, and on the routine monitoring in place. Comparison between countries should therefore be made carefully, taking into account these differences. When possible, it is recommended to validate the obtained prevalence estimates with external evidence and to critically review them with a group of stakeholders.
Guidelines for the use of the various indirect methods have been developed over a series of EMCDDA projects since 1997 and are available in the references below. The EMCDDA organises training workshops and encourages exchanges between national experts to increase the quality and coverage of prevalence estimates. Incidence estimation guidelines have also been produced.
Further advancements of the high-risk drug use area will be linked to development of new methods. The most recent developments include use of Bayesian statistics and Respondent Driven Sampling techniques to estimate the size of local hidden populations.
Aceijas, C. (2006), ‘Estimates of injecting drug users at the national and local level in developing and transitional countries, and gender and age distribution’, Sexually Transmitted Infections82(suppl_3), pp. iii10-iii17.
European Monitoring Centre for Drugs and Drug Addiction (1997), Estimating the prevalence of problem drug use in Europe, Office for Official Publications of the European Communities, Luxembourg (available at http://www.emcdda.europa.eu/html.cfm/index34027EN.html).
European Monitoring Centre for Drugs and Drug Addiction (1999), ‘Guidelines for the prevalence of problem drug use (PDU) key indicator at local level’, (available at http://www.emcdda.europa.eu/html.cfm/index58064EN.html).
European Monitoring Centre for Drugs and Drug Addiction (2004), ‘Guidelines for the prevalence of problem drug use (PDU) key indicator at national level’, (available at http://www.emcdda.europa.eu/html.cfm/index65519EN.html).
Frischer, M., Hickman, M., Kraus, L., Mariani, F. and Wiessing, L. (2001), ‘A comparison of different methods for estimating the prevalence of problematic drug misuse in Great Britain’, Addiction96(10), pp. 1465–76.
Hay, G., Gannon, M., MacDougall, J., Eastwood, C., Williams, K. and Millar, T. (2009), ‘Capture—recapture and anchored prevalence estimation of injecting drug users in England: national and regional estimates’, Statistical Methods in Medical Research18(4), pp. 323–39.
Hay, G. and Richardson, C. (2016), ‘Estimating the Prevalence of Drug Use Using Mark-Recapture Methods’, Statistical Science31(2), pp. 191–204.
Janssen, E. (2017), ‘Estimating the number of people who inject drugs: a proposal to provide figures nationwide and its application to France’, Journal of Public Healthpp. 1–9.
Jones, H. E., Welton, N. J., Ades, A. E., Pierce, M., Davies, W., Coleman, B., Millar, T., et al. (2016), ‘Problem drug use prevalence estimation revisited: heterogeneity in capture-recapture and the role of external evidence’, Addiction (Abingdon, England)111(3), pp. 438–47.
King, R., Bird, S. M., Overstall, A. M., Hay, G. and Hutchinson, S. J. (2014), ‘Estimating prevalence of injecting drug users and associated heroin-related death rates in England by using regional data and incorporating prior information’, Journal of the Royal Statistical Society: Series A (Statistics in Society)177(1), pp. 209–36.
Kraus, L., Augustin, R., Frischer, M., Kümmler, P., Uhl, A. and Wiessing, L. (2003), ‘Estimating prevalence of problem drug use at national level in countries of the European Union and Norway’, Addiction (Abingdon, England)98(4), pp. 471–85.
Larney, S., Hickman, M., Guy, R., Grebely, J., Dore, G. J., Gray, R. T., Day, C. A., et al. (2017), ‘Estimating the number of people who inject drugs in Australia’, BMC Public Health17(1), doi:10.1186/s12889-017-4785-7. (available at http://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-017-4785-7).
Smit, F., Laar, M. van and Wiessing, L. (2006), ‘Estimating problem drug use prevalence at national level: Comparison of three methods’, Drugs: Education, Prevention and Policy13(2), pp. 109–20.
UNAIDS/WHO Working Group and on Global HIV/AIDS and STI Surveillance (2010), ‘Guidelines on Estimating the Size of Populations Most at Risk to HIV’, (available at http://data.unaids.org/pub/manual/2010/guidelines_popnestimationsize_en.pdf).