Chapter 8
Improving the monitoring of problem and multiple drug consumption

Drug addiction, drug dependence, drug abuse, harmful use, problem use: there are a variety of concepts associated with the EMCDDA problem drug use indicator, each carrying its own subtle distinctions in medical or social dimensions. The EMCDDA indicator of problem drug use (PDU) monitors ‘injecting drug use or long-duration/regular use of heroin, cocaine and/or amphetamines’. Included in the definition, by convention, is the use of other opioids such as methadone.

This definition of PDU is a purely behavioural one based on drug consumption patterns and does not explicitly measure problems in any sense. Nonetheless, it is linked to the various concepts of addiction by the understanding that someone behaving in this way is very likely to fall within the more general concept of a ‘problem user’. It is important to note in this respect that the PDU indicator estimates only an important subgroup of those who can be thought of as having a drug problem of some form. Nevertheless, the approach does have value: as a behaviourally determined concept, its virtues are that:

Figure 14 Distribution of 'new' and 'all' clients requesting treatment according to their primary drug

Figure 14


Based on data from the Czech Republic, Denmark, Germany, Greece, Italy, Cyprus, Hungary, Malta, the Netherlands, Slovenia, Slovakia, Finland, Sweden, the United Kingdom, Bulgaria, Romania and Turkey.


Reitox national focal points.


  • It has allowed monitoring to proceed without being tied to definitions of addiction, dependence, harm and problem itself.

  • It is relatively easy to operationalise in research studies.

  • It groups together different types of drugs and modes of administration as alternatives, without specifically differentiating between them.

Historically speaking, the EMCDDA monitoring indicator was a child of its time – during the 1980s, and to a large extent the 1990s, heroin use and injecting drug use were seen as key components of the drug problem that required estimation. Furthermore, these forms of drug use could not be measured convincingly by survey techniques. The addition of amphetamines made the definition appropriate for some of the Nordic countries where injecting amphetamine use was important; and, although cocaine was included, in practice it was rarely a significant component in any estimates. While the PDU indicator still gives us a useful window on an important element of the drug problem, it is increasingly apparent that it needs further development to meet the requirements of monitoring today’s more heterogeneous European drug situation. Increasingly, we are seeing a more complex picture with respect to chronic drug problems in Europe. To keep its relevance to a changing world of illicit drugs, the monitoring task has to move forward and meet the challenges of covering a broader spectrum of drugs, and covering their use in finer detail than it has to date.

The enlargement of the EU has embraced a greater variety of social behaviours, with illicit drug use no exception. Developments within the drug culture, the rise of synthetic drugs and illicitly used medicines, the shift towards cocaine and the general high prevalence of cannabis use all have to be acknowledged in the understanding of what constitutes the needs of those with problems associated with their use of drugs. It can be noted from earlier sections of this report that, although heroin users still clearly predominate in demands to the drug treatment services, the picture is changing with respect to those who are entering treatment for the first time, among whom reported cannabis and stimulant problems have been increasing (Figure 14).

This changing position must be seen against the background of greatly expanded treatment provision for those with opioid-related problems as well as increased reporting coverage of treatment services. Opioid substitution treatment in particular, locking people into long-term continuous care (note that this not shown in these diagrams, which show only those entering treatment in the current year), emphasises the role of opioids in current treatment burden as compared with new treatment entrants. Nonetheless, in many countries it appears that those developing a drug problem today in Europe are likely to be, in terms of the substances they are using at least, more heterogeneous than has historically been the case.

The PDU indicator and counting hidden populations of drug users

The current PDU approach has proved extremely valuable in getting better estimates of the important group of drug users who represent the major consumers of drug treatment services in Europe. This approach has served to stimulate the development of a range of methods and statistical procedures to estimate the full size of this largely hidden population. In all these approaches measurement is grounded in the idea that a percentage of problem drug users are ‘administratively visible’ as they are in contact with a range of treatment, legal and emergency, and social services, and from this visible minority the size of the population can be estimated if the percentage is known. These kinds of indirect statistical methods complement population survey work, which, for a range of methodological and practical reasons, is less suitable for estimating the prevalence of drug use that is of low prevalence, stigmatised and largely hidden.

In terms of the current implementation of the PDU indicator across the EU, countries have adapted the definition to cover the practicalities of their local situation and the position is therefore heterogeneous. Nine countries essentially follow the EMCDDA definition as it stands, 11 countries estimate only the number of opioid (or heroin) users and a further four countries do not exclude problem cannabis users although cannabis users in general form only a very small part of their estimates (criteria for including cannabis users being quite strict in terms of counting only dependent or very intensive use).

The more widespread use of cocaine, crack cocaine and stimulants in general, along with the overlap of drug problems with problems associated with alcohol and prescribed medicines, implies that even within a consistent implementation of the indicator the estimates would now cover an increased variety of drug-taking repertoires than before, with the possibility of more varied consequences and problems. In addition to monitoring the overall extent of problem drug use, there is therefore a need to monitor separately the different behaviours that make up the PDU indicator, i.e. injecting and each drug type within the PDU definition. This may be particularly important in the light of evidence from some countries of increasing use of cocaine and the patterns of amphetamine use, as well as allowing detailed tracking of trends in opioid use. If all these behaviours are reported only in total, there is clearly a potential for masking important developments and a lost opportunity for a better understanding of trends.

Earlier in this report a separate estimate of heroin use and injecting in Europe was provided for the first time. In addition, we note the increase in treatment availability for opioid problems, with estimates of over half a million opioid substitution treatments in Europe. This suggests that the proportion of heroin users and injectors having or having had contact with treatment services may be quite large in many countries. The EMCDDA is currently exploring with its national technical groups what value can be added by bringing together information on treatment demand, treatment availability, and estimates of heroin use and injecting drug use.

Moving beyond PDU indicator

A further step in developing our ability to understand better the European drug problem is to explore the extent to which intensive drug use can be incorporated into the monitoring exercise, beyond PDU monitoring. The extent to which intensive drug use, however it is defined, is associated with dependence and levels and types of problems requires further elaboration, and, for example, Kandel and Davis (1992) estimated that in the USA around one-third of daily cannabis users could be considered to be dependent. The way forward in this area requires the formalising of the concept of frequent, intensive use of cannabis and other illicit drugs as a specific target for monitoring. Frequent or intensive use can be measured in survey data to complement its estimation through indirect statistical methods. Currently, survey data provide a useful window on different patterns of cannabis use, but these are largely restricted behavioural and frequency of use measures. Survey data are likely to be important if we are to develop robust estimates of the number of users of drugs such as cannabis who could be described as dependent or harmful users, at least by self-report.

This move to defining frequent, intensive use more formally for a range of drugs would also assist the development of research tools to assess levels of problems and dependence related to different levels and patterns of cannabis consumption. A number of European countries are working on developing methodological tools for measuring both intensive use and levels of dependence and problems, and the EMCDDA is promoting collaboration in this area of work.

A complication of identifying intensive use as an indicator of those who are likely to be most at risk of becoming dependent or experiencing problems is that the notion of what constitutes intensive use is to some extent drug specific. Although problem opioid use is strongly characterised by daily patterns of use, this is often not the case for stimulant drugs. With these drugs, binge consumption is often more common, use escalating for short periods and then decreasing, often with the users switching to other drugs or alcohol to alleviate the adverse consequences of abstaining. Both pharmacological and contextual factors may be important in influencing patterns of intensive drug use, but it is clear that behavioural measures of intensive use will need to be sensitive to different patterns of drug use associated with different types of drug.

The challenge of reporting polydrug use problems

For good practical and methodological reasons most reporting on drug use describes each substance separately. This provides the conceptual clarity necessary to facilitate reporting based on the behavioural measures available, but it ignores the fact that individual drug users will often have consumed or be consuming a range of both illicit and licit substances and these users may also be experiencing problems with more than one drug. They may substitute one drug type for another or may change their drug of choice over time or may use them complementarily. This kind of complexity is extremely challenging to a monitoring system even if analysis is restricted to simple behavioural measures of drug consumption in different time periods. If concepts of problematic and dependent use are included, these complexities increase further and very few robust data exist at a European level to permit informed analysis. Nonetheless, it is likely that some countries have a sizable population of chronic problem drug users, who are difficult to classify by primary substance and who may be experiencing problems due to their use of both licit and illicit substances. Addressing this problem requires developing a better understanding of the patterns of polydrug use and applying this to improve national and European-level reporting.

Figure 15 Use of drugs in the last year by different user groups in the general population aged 15–34


Data for Spain 1999. See Figure GPS-34 in the 2006 statistical bulletin for more information.


EMCDDA (2005b).



Within the general concept of multiple drug use, several specific meanings of the term must be considered. At one extreme, there is the use of several substances in an intensive and chaotic way, simultaneously or consecutively, in many cases each drug substituting for another according to availability. For instance, this is the case with problem users who use different opioids, as well as pharmaceuticals, cocaine, amphetamines and alcohol.

This pattern of use seems to exist among some chronic users, possibly among marginalised groups, possibly among people with psychiatric conditions. In many recording systems in Europe, these cases will be attributed to opioids.

As such, these individuals will fall within the definition and monitoring of problem drug users. However, there remains a further question of whether this intensive, chaotic polydrug use is enough of an entity in itself – a drug non-specific addiction therefore requiring targeted epidemiological measurements and treatment, support or harm reduction in a particularly difficult situation. Under these circumstances, any monitoring of the component parts of problem drug use would need to include polydrug use as one of these separately measured components.

Drug combinations

There is a second group of users who take several substances in a systematic way simultaneously, wanting the effects of the pharmacological combination, for instance speedballing – the simultaneous use of heroin and cocaine by injection.

A second drug with a functional or pharmacological rationale is not confined to simultaneous use but may also be used consecutively, as a replacement or as a complementary drug. For example, benzodiazepine can be used to reduce withdrawal symptoms when opioids are not available. In other cases, a second drug may be used for its compensating pharmacological effect: this is the case when the narcotic effect of opioids is modified by using cocaine, or when the anxiety effect of cocaine or amphetamines is placated by the use of opioids or other depressants.

Increased risks through polydrug use

The potentiating effect of one drug on another is sometimes considerable, and here the licit drugs and medicines – such as alcohol, nicotine and antidepressants – have to be considered in conjunction with the controlled psychoactive substances. The risk level will depend on the dosage level of both substances. Concerns exist about a number of pharmacological pairings: alcohol and cocaine increase cardiovascular toxicity; alcohol or depressant drugs, when taken with opioids, lead to an increased risk of overdose; and opioids or cocaine taken with ecstasy or amphetamines also result in additional acute toxicity.

Limits of the data available on polydrug use

The absence of available data limits reporting on many aspects of poly-drug use. Data are available from toxicological reports from drug overdoses and self-reports from those attending treatment services. While these sources allow some insights into poly-drug use, the information available is often limited and the representativeness of the data needs to be considered.

Survey results do show considerable use of more than two drugs by individuals in same time period. However, survey data often poorly report on some forms of drug use and, even where there are data, work needs to be done to develop comparable reporting standards. An example of the extent of poly-drug use data available from population surveys can be seen in a recent technical report of the EMCDDA (2005b). Taking the example of data from Spain, Figure 15 shows that use of one drug results in an increased tendency, as compared to that of the general population, to have used another and that this varies according to the drug considered. For example, among heroin users, the use of cocaine is relatively common, but use of other drugs is less apparent for cocaine users.

Developing an operational definition of poly-drug use: a question of timing?

Consideration needs to be given to what time periods are selected for assessing and reporting on polydrug use. Lifetime prevalence rates are in general not likely to be very useful or relevant to public health issues compared with measures of more recent use.

Polydrug use could be usefully defined operationally as the frequent use of more than one substance over a minimum specified time period, for example 1 month. This does not distinguish the various types of use described above, but gives an overall picture of what might be a high-risk group. The exception to discarding lifetime use in characterising polydrug behaviour is when dealing with the very young – pupils or students – in which case lifetime multiple use may more strongly reflect current use. The evidence, for example from ESPAD surveys, suggests that the more deviant/low-prevalence patterns of drug use among students (ecstasy, amphetamine, hallucinogens, cocaine, heroin) cluster among a few individuals.

Improving monitoring of the drug problem in Europe and improving the sensitivity to polydrug consumption

Achieving a better understanding of the nature and scale of the European drug problem is one of the key tasks of the EMCDDA. Drug use is, however, a complicated issue encompassing a range of behaviours that are variably associated with several important public health and social problems. Drug users may be experiencing or be at risk of problems. Patterns of drug use vary from the experiential, episodic and occasional to the regular, intensive and uncontrolled. Drug users may be classified under clinical definitions of having a drug problem or being dependent, and in research terms both these categories can be elaborated into continuums. To add to this complexity, drug users often consume multiple substances and change their consumption patterns over time. No single reporting instrument can adequately encompass this complexity. In practice, the multi-indicator approach adopted by the EMCDDA is intended to illuminate these different aspects of the drug phenomenon.

The PDU indicator, by focusing on a particular set of behaviours, provides a valuable window on some of the most detrimental and costly forms of drug use. As such, it is an important component in understanding the European drug problem as a whole. However, a clear need now exists to complement the PDU indicator’s overall estimates with substance-specific component estimates in order to address Europe’s increasingly heterogeneous drug problem.

Given that many of the data sources available are based on behavioural reports of drug consumption, the concept of frequent or intensive use needs to be developed on this basis. This will widen the perspective for monitoring drug problems beyond that currently found in the PDU indicator. Locating the PDU information alongside this broader information set will also represent a step forward in the EMCDDA’s efforts to improve its overall understanding of both the scale and the nature of drug problems in Europe. In parallel, work needs to advance on developing reporting standards that enable patterns of polydrug use to be better described at the European level. The first steps in this direction include developing a more sophisticated conceptual framework for looking at different types of multiple drug consumption – including the adoption of appropriate temporal frameworks – and identifying appropriate data sources.