Estimating the extent of the heroin problem in Slovenia: application of the key informant approach and the nomination technique where there are no other reliable sources of information


A. Sources and methods
B. Estimates based on secondary statistical data
C. Estimate based on the key informant approach
D. Estimates based on the nomination technique
E. Discussion


Author: D. NOLIMAL
Pages: 121 to 134
Creation Date: 1996/01/01

Estimating the extent of the heroin problem in Slovenia: application of the key informant approach and the nomination technique where there are no other reliable sources of information

D. NOLIMAL * Specialist in Social Medicine, Institute of Public Health of Slovenia, Ljubljana


Current evidence reveals a continuing upward trend in the misuse of illicit drugs in Slovenia. However, the science of estimating the prevalence of drug abuse and related problems is still undeveloped. Because of current data gathering practices, the data that are available are often of poor quality. In this paper the author describes two methods for estimating die prevalence of heroin abuse, the key informant approach and the nomination technique, which were used because there were no other reliable sources of information.

These methods produced estimates and brought to light a number of problems that researchers would have to solve in their pursuit of more reliable, relevant and useful data. However, speculating about die extent of illicit drug use in the country is still problematic.

Basic data collection and analyses at the national level must be improved. It is of vital importance to develop strategies and methods for obtaining estimates and thus more adequate information on which to base demand reduction strategies, to increase the number of epidemiologists and to establish a central information unit in the country.


In Slovenia reliable data on the prevalence of illicit drug abuse and related problems are scarce. Before 1989 the problem seemed limited and the country had very little experience in responding to drug problems. It has become to be regarded as more serious, however, in the last five years, judging by the great attention paid by the mass media and the high level of public concern. A considerable number of governmental and non-governmental bodies, professional and other meetings, and articles and publications devoted to the problem have emerged. In particular the injecting of heroin has become an important political issue, not least because of its relationship to transmission of the human immunodeficiency virus (HIV). The fear of a drug-related epidemic of acquired immunodeficiency syndrome (AIDS) has induced government and social agencies to develop some basic preventive measures and treatment. In 1991, the first specialized treatment services for heroin abusers became available.

The preparation of this report was facilitated by many individuals and agencies, including staff of the Ministry of the Interior and the regional institutes of public health and, in particular, Dr. Milan Krek of the Regional Institute of the Public Health of Koper, Nino Rode of School for Social Work, Vladimira Rejc and Viktor Nolimal.

No reliable epidemiological data were available for the Government to plan treatment services and programmes for the increased number of heroin abusers in the country at the beginning of 1990. Individuals who have specialized knowledge of such matters or dealt with drug problems had estimated the number of heroin abusers at between 2,000 and 10,000 [ 1, 2, 3, 4] , although it was not clear how they arrived at those estimates. The numbers seemed to be lower than in western Europe but not negligible compared to the small size of the country and its population of 2 million. It seems they had already reached a level that could facilitate the spread of HIV. If HIV were to enter the network of injecting drug users, seroprevalence among these populations might quickly reach high levels. Also, the need for more accurate information on the level of heroin abuse was given urgency by increasing drug-related mortality among injecting drug users.

A. Sources and methods

The absence of reliable information on drug problems in the country motivated the Institute of Public Health to raise some basic questions to improve the country's response to the problem: What kind of data were available on the number of heroin abusers? How reliable were those data? Would recent rates of increase (or decrease or stagnation) continue? How many of those who had abused heroin in the past would continue to do so? What was the proportion of abusers experiencing social and health consequences (e.g. criminality and illnesses resulting from drug abuse)? What was the proportion of those seeking treatment?

If these questions are to be answered, a number of practical problems (many having to do with resources) and theoretical problems (methodological and conceptual) need to be solved. For this reason the paper reviews the available documentation on the extent of heroin abuse and describes in more detail three sources of information for estimating the prevalence of such use in Slovenia:

  1. Routine secondary statistical data;

  2. Key informants;

  3. Field study based on the nomination technique, used in the snowball sampling method.

It should be emphasized at the beginning that the primary goal of all three approaches was to identify high-risk groups that could be targeted for intervention rather than to measure addiction prevalence. Thus both quantitative and qualitative data and assumptions were used, and there are many potential sources of error.

B. Estimates based on secondary statistical data

1. Results

First attempts to more systematically collect data on illicit drugs started in 1989 at the Institute of Public Health. Some useful information on illicit drug abuse and drug abusers was already available from a number of sources including health and social services, pharmacies, enforcement agencies, schools and prisons. Most of these agencies kept records of some sort. For example, individual doctors and social workers had information on the types of persons who sought help for drug-related problems. Some information on the abuse of drugs such as heroin was available from psychiatric hospitals and outpatient health centres. These included admissions for drug dependence and abuse or drug psychosis.

The prescription statistics on methadone and some other psychoactive substances provided a relative and indirect measure of drug abuse in the country. Indeed, the increasing number of prescriptions for methadone was one of the first warning signals of the growing problem of heroin abuse (figure I).

Similarly, the police had information on persons who had been involved in drug- related crime. Also, information on the quantities of the drugs seized and the numbers of persons arrested for possession of drugs was available from the Ministry of Internal Affairs. The police recorded increased numbers of heroin seizures, and the main feature of the 1989-1996 period as compared to earlier periods was the significant increase in the quantity of heroin seized (figure II).

Also, the courts had some information on people who had been charged with offences related to the abuse of drugs. Voluntary service groups also had some information, and teachers had some information on schoolchildren abusing drugs.

Finally, there had been epidemiological monitoring of AIDS and HIV infection and related data collection by the Institute of Public Health. The number of drug- injection-related AIDS/HIV cases was low at the beginning of the 1990s and still is, compared with some other countries having a considerable heroin problem.


To make productive contact with key authorities in the agencies it was important to inform them of the objectives of the project. The Institute of Public Health started by explaining how information gathered by the agency could help to estimate addiction prevalence. Having gained access to, the drug-related information, it was then important to assess its quality. A relatively large amount of information was available, but many of the records were incomplete and there were problems with validity and reliability.

Figure I. Number of methadone prescriptions by sex, 1988-1995

Full size image: 11 kB, Figure I. Number of methadone prescriptions by sex, 1988-1995

Figure II. Heroin and cocaine seized, 1985

Full size image: 14 kB, Figure II. Heroin and cocaine seized, 1985

The usefulness of the data, irrespective of the approach used, depended to a large extent on the way they were reported. Moreover, the institutes involved in drug-related data collection had never produced a joint report. Some assumptions were made about the number of detected cases. This information was often accepted at face value without looking into the appropriateness of the methodology for collecting it. Other useful sources of information (e.g. treatment demand, surveys) were not available before 1991.

The combined data from a variety of sources allowed estimating the size of the problem which was recognized as considerable and as having increased rapidly since 1989-1990.

C. Estimate based on the key informant approach

Having identified police and public health reports as sources of information, a complementary step was to contact a few knowledgeable individuals, such as doctors, pharmacists, social workers, teachers, community workers and policemen, to obtain their informal qualitative impressions of drug problems and issues.

The next step was to use this qualitative and quantitative information to formulate specific questions. In 1993 staff of the Institute of Public Health designed a standard questionnaire on injection drug abuse. It was sent to the Criminal Investigation Directorate of the Ministry of Internal Affairs and to the nine regional public health institutes. The Criminal Investigation Directorate administered the questionnaire to the 11 regional criminal investigation bureaux, each of which assigned a professional to respond to the questionnaire. In this way, a comprehensive multidisciplinary network of key informants was set up to collect data on drug abuse at the local level. The key informants in turn developed their own networks of local informants with specialized knowledge of different aspects of the drug scene. In each region as many informants as possible were recruited.

All informants were asked to make informed estimates about the number of registered injecting drug abusers at the regional or local level. The Institutes's concerns were reflected in questions such as the following: What was the extent of illicit drug use in the region? How was drug abuse perceived in the community? Approximately how many people known to the informant were considered to be heroin abusers? How many of these individuals had already difficulties with the police? How many of them had already been treated? How many of them injected drugs or had ever done so? How many abusers were HIV-positive?

1. Results

The regional and local estimates were combined to arrive at national aggregate estimates: the police estimate was 1,026-1,226 heroin injectors and the public health estimate was 743-1,045 heroin injectors (table 1, figure III).

Table 1. Number of injection heroin addicts in Slovenia as reported by public health authorities and the police 1992


Public health data

Police data

64 300-400 47 400
Novo Mesto
9 14
28 100
93 400-600 224 410-610
17 40 3 27
M. Sobota
27 35
Nova Corica
Ravne na Kor.
194 743-1045 338 1026-1 226

Source:National and regional institutes of public health and Ministry of Internal Affairs, 1992.

Figure III. Injection drug users, 1992

Full size image: 14 kB, Figure III. Injection drug users, 1992


The key informant approach provided useful information on visible heroin addicts. The registered and estimated numbers of injecting drug abusers were reported by regional police and public health authorities. Some figures were derived from official registers but most were based on best estimates.

Both the police and public health reports showed that a high proportion of injecting drug abusers lived in the two largest Slovene cities, Ljubljana and Koper. Some injecting drug abusers were in contact with the police and/or with health treatment institutions.

The approach generated information relatively easily and at minimal cost. Of course, it should be remembered it relies entirely on estimates by people who themselves could best be described as "local observers". Clearly their reports were limited by incomplete knowledge, personal bias and inaccurate estimations. Two important factors influenced the selection of key informants. their willingness to participate and the lack of resources. Also, there has been a general failure to recognize the extent to which attitudes, values and opinions should be elicited and taken into account when making such estimations. Law enforcement estimates seemed to be more refined and based on more careful analysis of the facts. Public health estimates seemed to have been based on a priori feelings and assumptions rather than on specific knowledge of the problem. There was a lot of non-recognition and underreporting of drug use by some regional public health authorities. The validity and reliability of the regional estimates produced by the key informant approach were limited by other factors as well: e.g. different terminology and different sources of data and ethical and legal considerations.

The study showed that heroin was the drug that caused the most acute health, social and other problems for individuals and communities. It was relatively well known that heroin injectors had a high risk of becoming infected by and spreading the AIDS virus because they exchanged syringes and needles. The study also showed that many problem drug abusers had first asked for treatment when a methadone programme was introduced in the country.

D. Estimates based on the nomination technique

This part of the paper describes in more detail the field study conducted in November and December 1993 by the Institute of Public Health in Koper and the Institute of Public Health in Llubljana [ 5] . The existence of a methadone maintenance treatment programme was particularly valuable for contacting hard-to-reach illicit drug abusers. Because of the lack of expertise and experience in prevalence estimation, the study did not, at the time of its completion, generate a prevalence estimate. Later, in 1996, the data derived from this study and available treatment data were used to retrospectively generate a heroin abuse prevalence estimate for Slovenia.

1. Description of the method

The starting point of the study was. 26 patients selected from the methadone maintenance treatment centre in Koper, where a total of 263 heroin abusers from different parts of Slovenia were being treated in 1991-1993. The study relied on gaining the trust of methadone patients who were knowledgeable about heroin abusers and who were respected by their peers. In addition to contributing information about themselves, these selected patients were able to quickly and relatively smoothly gain access to heroin abusers that would otherwise have taken considerably more resources.

From the initial sample of heroin abusers firsthand information on the local drug scene was collected. The term "snowball sampling" refers to a method of recruiting as many injection drug abusers as possible, seeking out hidden heroin abusers and including them in the study. In effect, the methadone maintenance clinic patients served as field research assistants.

An important criterion in choosing from among the methadone patients was where the patients came from. It was desirable to have as many geographical regions of Slovenia as possible represented; other factors-demographic, occupational and recreational-were less important.

Because there was only one other methadone treatment centre available in the whole country in 1993, a rather unusual situation existed in both centres: the patients came from all around the country, making it possible to achieve geographical representativeness.

The field research assistants were asked to nominate individuals they personally knew to be regular heroin abusers. A questionnaire on risky behaviours was developed and administered to the nominated individuals. The topics covered included past treatment experiences. To avoid double-counting the same person by different field research assistants in a given time period, a code was used that consisted of the initials of first and last names, gender and year of birth.

Also, staff of the Institute reviewed drug treatment records in the country. Most of the information was available at the two methadone maintenance treatment centres and at the Institute of Public Health, which had records on hospitalizations for heroin abuse but not abusers as a whole. These records provided benchmark information on the number of treated heroin abusers in the country.

2. Results

More than 890 heroin abusers were nominated and then contacted by 26 field research assistants, and all of them answered the questionnaire. It needs to be stressed that owing to the lack of scientific rigour, it was not clear at the time when the data were collected how the snowball groups had been formed. Retrospectively, staff of the Institute tried to understand the situation from anecdotal information Eventually, the study consisted of at least 21 snowball groups. The most active field research assistant came back with information on 139 heroin abusers (19 per cent of all the nominees), but this success rate was an exception. Most of the rest were able to nominate between 15 and 20 heroin abusers. For six fieldworkers, it is not clear how many heroin abusers they had contacted.

There were 616 (69.2 per cent) male and 241 (27.1 per cent) female heroin abusers from different parts of Slovenia who answered the questionnaire. The abusers came from most part of the country. Not surprisingly, however, most of the subjects resided in Koper (table 2). Of these abusers, 362 (40.7 per cent) had been treated at least once in their life. This percentage was used as a multiplier to calculate prevalence.

Table 2. Nominated heroin abusers in Slovenia by place of residence, 1993



Distribution (%)

135 15.1
449 50.4
24 2.7
Nova Gorica
8 0.9
5 0.6
227 25.5
Novo Mesto
21 2.4
Ravne na Kor.
6 0.7
M. Sobota
0 0
3 0.3
12 1.4
890 100

Since no specialized treatment services were available for heroin before 1991, it is assumed that most of those who were treated in the past sought treatment after 1991, when the first treatment centres were established. The total number of heroin abusers treated nationwide with methadone maintenance was approximately 360. The total number of hospitalized heroin abusers in the same period was 160. These figures were used as the benchmark to calculate prevelance:

E=(n 1+n 2)x 100/ m

where E = prevalence estimate for the period 1991-1993, n 1= total number of treated methadone patients for that period, n 2= total number of hospitalized heroin patients for that period, m (multiplier) = the proportion of the heroin abusers who were in treatment in that period (362/890 = 0.40) and n 1 +n 2 (benchmark) = 360 + 160 = 520. Thus,

E = 520 x 100/40 = 1,300

Of course, the confidence limits of this estimate of the heroin-abusing population in Slovenia from 1991 to 1993 (1,300 persons) are extremely wide.

3. Quality and validity of the collected information

It needs to be stressed that this exercise is a second look at data that were originally gathered for purposes other than to generate an estimate. There was almost no structured data collection on heroin abuse and related problems in Slovenia in the studied period, so the available data are useful only for allowing an initial assessment of the extent of heroin abuse in Slovenia. Many issues could be addressed only superficially. The Institute does not know enough about the quality of the information gathered: this includes the job performance of the field research assistants and the reliability of self-reporting by the informants. It had to retrospectively arrive at an understanding of the field research from anecdotal information and incomplete data. A number of simplifying assumptions were adopted to facilitate subsequent analysis of the data. A full statistical analysis of the snowball sample and the benchmark group would require more sophisticated methods and more accurate and complete data.


The recruitment of drug abusers from the treatment environment to act as field research assistants was crucial in gaining access to hidden heroin abusers. The assistants helped to draw a preliminary geographic map of the areas and cities where heroin abuse was most prevalent. They showed credibility with other drug abusers and were able to establish trusting, respectful relationships with them. It is important to emphasize the wealth of information and ideas provided by the individual drug abusers who were part of the treatment programme, and one of the important recommendations of this study is to include drug abusers and seek their help where possible.

One of the basic concerns of an approach such as this is how to achieve a representative sample of a hidden population. Snowball sampling was used to investigate patterns of drug abuse, which is greatly stigmatized, and to learn about heroin abuse and, in particular, injection of that drug. The sampling started with 26 individuals, all methadone patients, who were known to meet two criteria: the ability to gain access to and the trust of hidden heroin abusers and broad geographical distribution of primary residence. Sensitive issues of identifying and gaining access to a hidden population for the purpose of research were carefully discussed with these individuals, who became, in effect field research assistants.

The main question that must be answered Ls how representative the intial sample was. Another problem is the lack of information on how nominees were selected. It is not clear how many successive stages of snowball were produced by different individuals. Also, it has not been possible to learn retrospectively if some of the same drug abusers were nominated more than once.

Information on non-participating nominees (i.e. those who refused to answer the questionnaire) was lacking. There were, as well, errors in reporting nominees' treatment status. The amount of effort that was devoted to reducing the number of non- respondents was unknown. The fact that heroin abuse is a highly stigmatized, illegal activity probably increased the unreliability of the self-reports. The effects of drug use itself on the accuracy of the self-reports were unknown (all reports were prone to falsehoods) in the absence of biochemical tests to verify drug abuse.

The other methodological problem in this approach concerned the registration of the treated heroin abusers. The medical files were incomplete because not all new data had been entered, double entries of the patients were not excluded and there were considerable variations in the way treatment centres in different parts of the country recorded the data. Moreover, some documentation on treated abusers did not distinguish among different drugs of abuse or between those who had simply tried a drug and those who were addicted. It should be emphasized that with the exception of psychiatric hospitals, which treated 20-50 heroin abusers per year before 1991, there was no specialized treatment infrastructure available before that time in Slovenia. The introduction of methadone maintenance treatment programmes in 1991 brought many more heroin abusers into the treatment programmes.

Ethical issues related to field research and the use of methadone patients as field research assistants should be considered. All persons who volunteered personal data were adequately informed about the nature of the project and its objectives. The collected data were used only for research purposes. No action directly affecting the drug abusers was taken. Respect for the privacy of participating patients and other drug abusers was guaranteed. Research was under-taken with anonymous data. To secure anonymity, each individual was given a code. The methadone patients who were asked to collect data were under no obligation to do so. They were told that they were at liberty to withhold their cooperation. Also, they had a right to withdraw from further cooperation without giving a reason. Their participation in the project had no influence on their participation in methadone maintenance treatment.

E. Discussion

At the beginning of the 1990s, when a large number of young people in Slovenia got involved in heroin abuse, reliable and comparable epidemiological data on drug abuse in Slovenia were lacking. Because there were no treatment facilities, no prevention research infrastructure and not enough specialized knowledge and experience to address drug problems, the estimation approaches lacked methodological and conceptual clarity.

While there was little scientific information on the phenomenon, the growing number of police seizures, reports on rising numbers of treatment requests and anecdotal information caused serious concern on the part of many individuals and organizations. One of these, the Institute of Public Health, started a drug epidemiology programme to help to develop an adequate response.

It seems that the need for more reliable data, including addiction estimates, was intensified when injecting drug abusers became an increasingly significant group in the AIDS pandemic. The Pompidou Group assistance programmes in 1993-1996 were the first to introduce basic concepts of drug use epidemiology [ 6] . The standardized methods and indicators developed by the Pompidou Group raised awareness of the threat to public health and the need to collect information [ 7] .

One of the results of that assistance programme is this paper, which is the first attempt to use a more scientific approach to estimating the extent of illicit drug abuse in Slovenia. Two current articles estimate the total number of heroin abusers in the country at between 2,000 and 4,000 [8, 9]. The Institute's nomination technique has generated a considerably lower number, 1,300 heroin addicts in the period 1991-1993, but the confidence limits for this estimate are extremely wide. The key informant approach has generated an estimate for 1992 of between 1,045 and 1,226 persons. Thus, two different methods arrived at approximately the same results. Since both used a treatment population as the source group, it can be speculated that the treatment population is overrepresented in the samples. Little is known, for example, about abusers who never get in contact with the legal or treatment systems.

The lack of conceptual clarity, including the case definition, became evident when reviewing efforts to estimate the extent of drug abuse in the country. The absence of a "common language", including an agreed nomenclature, definitions and standardized instruments for the assessment of drug use, was also clear. In the literature, authors seem to equate drug abuse with drug dependence. Terms like heroin user, misuser, abuser, addict, person dependent on heroin, and the like were often used with unclear reference to the real addiction status of the observed individuals. Also, there was no real agreement about concepts such as frequency and harmfulness. Presumably the term addiction was meant to indicate a problem of greater magnitude for the individuals. Also, the term drug use might have been an all-inclusive one that covered drug users with lighter and heavier patterns of drug use and with or without consequences. Also, addiction sometimes involves physical dependence and craving and sometimes not. It is important for future research to begin to use the same terms in the same ways.

1. Recommendations for improving the estimation approaches

The lack of specialized epidemiological knowledge in addressing the prevalence of drug abuse and related problems has been preventing the development of better estimates. Since the drug problem is a relatively new one, it will be necessary to train personnel and to guarantee the resources needed for their work.

The described estimations based on the key informant approach and the nomination technique are more reliable and have allowed recognizing a number of problems which researchers have to deal with in their pursuit of more reliable, relevant and useful data. Improvement of basic data collection and analyses and especially of estimates is needed at the national level and in each of the regions, taking due account of local characteristics.

There still is a shortage of resources to improve the epidemiological responses to drug problems in Slovenia. Because of great public concern and stiff competition for public funds, anectodal estimates only inflate the magnitude of the problem. It is of vital importance for the development of drug information systems to increase the number of drug abuse epidemiology experts in the country and to establish a central unit that would collect, process, analyse and interpret information on the abuse of drugs and the problems they cause in Slovenia. The unit would monitor trends over time and would be responsible for improving the relevance of information to drug policy.

Finally, the scope and methods of drug abuse epidemiology with its growing potential for collection and use of data has led to concern that individual privacy might be violated. Ethical dilemmas and controversies have always been a part of field research, but in the context of this work, the ethical issues are often not clear-cut and a field researcher can face major moral dilemmas. As should generally be the case where human beings are the subject of investigations, great care must be taken to protect the rights of the drug-abusing individuals.

2. Conclusion

All too often the planning and evaluation of drug policies and services are handicapped by the lack of relevant, reliable estimates. Of course, if knowledge of methodological and conceptual issues is limited and if the resources are scant, it cannot be expected that the response to the drug problem will be adequate.

Slovenia's socioeconomic and political situation is conducive to a further increase in drug abuse. Growing heroin availability, limited economic prospects and the loss of traditional values have contributed to spreading heroin abuse. Of course, speculating about the current and future extent of the problem when there are no reliable sources of information is not an easy task. If the response of the Institute is to succeed, and in the face of the AIDS epidemic it had better succeed, basic epidemiological questions need to be answered.

The preliminary character of this report must be emphasized. The two different estimation methods presented arrived at approximately the same estimate: between 1,000 and 1,300 heroin abusers in the country in the period under study. However, given the developmental state of the research, speculating about the extent of heroin abuse in Slovenia and about the adequacy of existing services is still a challenge. If drug policy is to be reasonably well informed and not merely the product of wishful thinking, the quality of the information being collected on the hidden drug-abusing population must be improved.



D. Nohmal and M. Premik, "Some social-medical aspects of drug abuse", Zdrav Vestnik, vol. 61 (1992), pp. 133-136.


V. Flaker and others, "Project Stigma: The interim report to the WHO Global Programme on AIDS (The Help, Self-Help, Information and Advice on Drugs and AIDS Society, Ljubljana, Slovenia)", World Health Organization, Geneva, 1992.


D. Nolimal and others, "Descriptive epidemiology of the group of street injection drug users in the regions of Koper and Ljubljana in the year 1991 ", Zdrav Var, vol. 32 (1993), pp. 161-164.


D. Nolimal and S. Onusic, Overview of drug Misuse in Slovenia: Epidemiology and Research (Ljubljana, Institute of Public Health of Slovenia, 1993).


M. Krek, J. Krek-Misigoj and D. Nolimal, "Difficulties with field research of the illicit drug use", in Anthology of the Papers Presented at the Information Systems and Applied Epidemiology of drug Misuse Follow-up Seminar, D. Nolimal and M. Belec, eds. (Ljubljana-Piran, Institute of Public Health, 1995), pp. 60-64.


Council of Europe, Report Information Systems and Applied Epidemiology of Drug Misuse Follow-up Seminar, P-PG(94)20 (Strasbourg, Council of Europe, 1994).


R. Hartnoll, "Recent trends in drug consumption, policy and research 1994. synthesis of national reports", P-PG/Epid(95)18 (Strasbourg, Council of Europe, 1995).


D. NolimaL "Self help and methadone in Slovenia", Euro-Methwork Newsletter (Slovenia), vol. 7, 1996, p. 15.


A. Kastelic and R- T. Kostnapfel "Treatment of addicts of prohibited drugs in Slovenia" , Euro-Methwork Newsletter (Slovenia), vol. 8, 1996, pp. 11-14.