Figure 12 Estimates of the prevalence of problem drug use, 1999–2003 (rate per 1 000 population aged 15–64)

Figure 12


CR, capture–recapture; TM, treatment multiplier; PM, police multiplier; MI, multivariate indicator; TP, truncated Poisson; CM, combined methods. For more details see Tables PDU-1, PDU-2 and PDU-3 in the 2005 statistical bulletin. The symbol indicates a point estimate and a bar indicates an uncertainty interval, which can be either a 95 % confidence interval or an interval based on sensitivity analysis (see Table PDU-3). Target groups may vary slightly owing to different methods and data sources; therefore comparisons should be made with caution. The Spanish estimate does not include problem cocaine use; a higher estimate is available in Tables PDU-2 and PDU-3, which takes this group into account but which may not be as reliable.

Sources: National focal points. See also EMCDDA (2003).


Estimates of the prevalence of problem drug use at national level over the period 1999–2003 range between two and 10 cases per 1 000 population aged 15–64 (based on midpoints of estimates) or up to 1 % of the adult population (105). Prevalence appears to differ greatly between countries, although when different methods have been used within one country the results are largely consistent. Higher estimates are reported by Denmark, Ireland, Italy, Luxembourg, Austria, Portugal, Spain and the United Kingdom (6–10 cases per 1 000 inhabitants aged 15–64 years), and lower rates are reported by Germany, Greece, the Netherlands and Poland (fewer than four cases per 1 000 inhabitants aged 15–64 years) (Figure 12). Among the new countries of the EU and the candidate countries, well-documented estimates are available from only the Czech Republic, Poland and Slovenia, where figures are in the low to middle range at 3.6, 1.9 and 5.3 per 1 000 population aged 15–64 respectively. The weighted average rate of PDU in the EU is probably between four and seven cases per 1 000 population aged 15–64 years, which works out at 1.2–2.1 million problem drug users in the EU, of whom some 850 000 to 1.3 million are active injectors. However, these estimates are far from robust and will need to be refined as more data become available from the new Member States.

Local and regional estimates suggest that the prevalence of PDU can vary greatly between cities and regions. The highest local prevalence estimates in the period 1999–2003 are reported from Ireland, Portugal and the United Kingdom, with rates per 1 000 reaching 16 (Dublin), 17 (Beja), 24 (Aveiro) and 25 (parts of London), up to 29 (Dundee) and 34 (Glasgow) (Figure 13). Geographic variability is, however, marked at the local level, with prevalence in another part of London being estimated at 6 per 1 000. This suggests the need to increase the availability of reliable local prevalence estimates in many other countries, where particularly high (or low) local or regional prevalence rates may exist but are not being measured.

Figure 13: National and local estimates of the prevalence of problem drug use, 1999–2003 (rate per 1 000 population aged 15–64)

Although there have been considerable improvements in estimation techniques, a lack of reliable and consistent historical data complicates the assessment of trends in problem drug use. Reports from some countries of changes in estimates that are supported by other indicator data suggest that there has been an increase in PDU since the mid-1990s (EMCDDA, 2004a); however, in some countries, this appears to have stabilised in more recent years. Repeated estimates are available for 1999 to 2003 from 16 countries (counting Denmark and Sweden, both with data from 1998–2001, and the United Kingdom, where data are for 1996–2001). Of these 16 countries, six reported an increase in PDU estimates (Denmark, Finland, Norway, Austria, Sweden and the United Kingdom), five reported a stable prevalence or a decrease (Czech Republic, Ireland, Germany, Greece and Slovenia) and five (Spain, Italy, Luxembourg, the Netherlands and Portugal) reported different trends depending on the estimation method used, which may partly refer to different target groups among problem drug users (106).

(105) See Table PDU-1 in the 2005 statistical bulletin.

(106) See Figure PDU-4 in the 2005 statistical bulletin.