Supported by the

UNDCP Global Assessment Programme on Drug Abuse

Conducted in partnership with
the Narcotics Control Division, Anti-Narcotics Force
Government of Pakistan

DRUG ABUSE IN
PA K I S TA N
Results from the year 2000 National Assessment
Supported by the
UNDCP Global Assessment
Programme on Drug Abuse

Conducted in partnership with

the Pakistan Anti-Narcotics Force

The Global Assessment Programme on Drug

Abuse (GAP) improves the global information
base on
patterns and trends in drug consumption through
supporting Member States to build the systems
necessary for collecting reliable data to inform
policy and action; encouraging sharing of
experiences and technical developments through
regional partnerships; and encouraging the
adoption of sound methods to collect comparable
data.
UNITED NATIONS OFFICE FOR DRUG CONTROL AND CRIME PREVENTION

D R U G A B U S E I N PA K I S TA N
Results from the year 2000 National Assessment Study

Supported by the
UNDCP Global Assessment Programme on Drug Abuse
Conducted in partnership with
the Narcotics Control Division, Anti-Narcotics Force
Government of Pakistan

UNITED NATIONS
New York, 2002
The designations employed and the presentation of material in this publication do not imply the expres-
sion of any opinion whatsoever on the part of the Secretariat of the United Nations concerning the legal
status of any country, territory, city or area, or of its authorities, or concerning the delimitation of its
frontiers or boundaries.

This publication has not been formally edited.

Acknowledgements

The National Drug Abuse Assessment Study in Pakistan

This study was conducted under the auspices of the Pakistan Anti-Narcotics Force (ANF) and the Pakistan
Regional Office of the United Nations International Drug Control Programme (UNDCP). However, the
study was also reliant on the goodwill and hard efforts of a large number of individuals without whose
participation this exercise would not have been possible. These include not only the research team but
also those who facilitated and participated in the study. These include government functionaries and
NGO representatives who made possible the access to treatment and prison facilities and provided other
expert advice and support. We are also indebted to the large number of citizens who participated in the
research exercise. We are grateful to the support of all those who contributed and in particular would
like to acknowledge the contribution of the research and implementation team listed below.

Government of Pakistan (Anti-Narcotics Force)

Saiyed Mohib Asad, Deputy Director-General.

Mr Mairaj. A. Arif, Deputy Director.

UNDCP (Pakistan)

Mr Bernard Frahi, Representative.

Mr Thomas Zeindl Cronin, Programme Management Officer.
Dr Nadeem-Ur-Rehman, Demand Reduction Consultant/Study Coordinator.

UNDCP (Global Assessment Programme, GAP)

Mr Paul Griffths, Senior Epidemiologist.

Dr Kamran Niaz, Regional Epidemiological Adviser.

Statistical and Research Consultant

Mr Colin Taylor, National Addiction Centre (UK)

Data Processing

Khan Mohammad Kashif, Computer Programmer.

Imran Zafar, Data Entry Operator.
Khalid Siddique, Data Entry Operator.

Field Work Supervisors

Ms Beena Hassan, Sindh Province.

Dr. Shahryar Aftab, Punjab Province.
Mr. Edwin Arthur, Balochistan Province.
Mr. Zakir Shoaib, N.W. Frontier Province.

iii
Interviewers

Naushad Ahmed Malik, Mirza Akhtar Mughal, Maqbool Ahmed Mashori, Zeba Anjum, Sarfaraz Qutab,
Javed Ali, Abdul Rauf, Wajid Ali, Dr. Qasim Brohi, Ms. Mehmooda Aftab, Mrs. Rukhsana Hussain, Bushra
Saeed, Aratus Attarad, Shahid Bilal, Siraj Din Bhatti, Pious Bhatti, Dr. Khalida Bajwa, Muzaffar Hussain
Khan, Mr. Asim Shahzad, Ommi Kalsoom Khalduna, Edward Arthur Iqbal, Nafeesa Sultana, Shaheena
Fakhir, Muhammad Shafiq, Muhammad Hafeez Farooq, Iqbal Bano, Akbar Ali Durrani, Aftab Ali, Andrew
Florence, Siqa Fateh, Abdul Waheed Khattak, Zulfiqar, Hilal Ahmed, Nosheen Sabhzada, Asad Haroon,
Fawad Rizwan, Gull Haider, Babar Ahmed Bahar, Fazle Ghaffar, Hidayatullah Neak Akhtar, Gul Akbar Khan,
Zahid Akhtar.

iv
Contents

Pages
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1. Background ....................................................... 3

Pakistan in context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Drug cultivation and production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
The drug abuse problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Status of convention adherence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Legislation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
National drug control policies, priorities and plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
National institutional framework and capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Socio-economic characteristics of Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2. Rationale, methods and sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3. Mapping the national contours of drug abuse: the key informant interviews . . . . . . . . 11

Key informant interviews sample description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Perceptions of drug use in the locale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Trends in drug abuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Perceptions of problems arising from drug abuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Attitude questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

4. Interviews with drug abusers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Demographics—age distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Prison contact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Prison history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Demographics—education and employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Treatment contact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
First treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Need for treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Drug abuse history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Current drug use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Dependence and problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

5. Drug injecting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Drug injection among the sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Drug injecting amongst women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Age, education and employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Patterns of use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

v
 Pages
Injection risk behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Karachi profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Lahore profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Quetta profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Peshawar Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Drug treatment history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Arrest and imprisonment history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Living and support status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

6. Data on prisons activity in the four cities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Prison data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Overview: drug charges in Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

7. Drug treatment in Pakistan: the national treatment register update . . . . . . . . . . . . . . . 43

Updating the register of drug treatment in Pakistan ...... .. ....... . . . . . . . . . . . . . 43

Overview and history of drug treatment in Pakistan . ...... .. ....... . . . . . . . . . . . . . 43
UNFDAC funded treatment facilities . . . . . . . . . . . ...... .. ....... . . . . . . . . . . . . . 44
Integrated drug demand reduction project . . . . . . . ...... .. ....... . . . . . . . . . . . . . 44
Past assessments of drug treatment programmes . . ...... .. ....... . . . . . . . . . . . . . 45
National survey of drug treatment and rehabilitation services in Pakistan . . . . . . . . . . . . . 45
National treatment registry update 2000 . . . . . . . . ...... .. ....... . . . . . . . . . . . . . 46

8. National drug contour mapping: estimating the prevalence of hard-core heroin use
in Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Data structures of the contour mapping exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Method of estimation: treatment multiplier method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Method of estimation: geographical coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Estimation of prevalence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Improving the initial estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Refinements to initial multipliers for estimating the number of addicts . . . . . . . . . . . . . . . 55
Interpretation of the initial estimates and their extrapolations . . . . . . . . . . . . . . . . . . . . . 56

9. Drug abuse in Pakistan: the implications of the Pakistan national assessment study . . . 59

Comparisons with the earlier national assessment studies . ........ ..... . . . . . . . . . . 59

Drug use by women . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ ..... . . . . . . . . . . 60
Access to and delivery of drug treatment . . . . . . . . . . . . . ........ ..... . . . . . . . . . . 61
Drug injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ ..... . . . . . . . . . . 62
Developing a permanent drug information system to inform policy and action . . . . . . . . . . 62
Other research needs . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ ..... . . . . . . . . . . 63

Annexes

I. Methodological discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
II. Supplementary figure and tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
III. Tables supplemental to prevalence calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Tables

1. Key informants’ occupational groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2. Number of contacts by respondents in the last year with drug abusers . . . . . . . . . . . . . . . 13

vi
 Pages
3. Respondents’ perceptions of the scale of drug use in the locales . . . . . . . . . . . . . . . . ... 14
4. Respondents’ perceptions of the scale of drug use in the locales: overall ratings . . . . . ... 15
5. Respondents’ perceptions of the differences in the drugs most commonly used by men,
women and young people . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
6. Perceptions of long-term trends in drug use in the locales (last 5 years) . . . . . . . . . . . . . . 17
7. Perceptions of long-term trends in drug abuse (last 5 years): overall ratings . . . . . . . . . . . 18
8. Perceptions of short-term trends in drug use in the locales (last year) . . . . . . . . . . . . . . . 19
9. Perceptions of short-term trends in drug abuse (last year): overall ratings . . . . . . . . . . . . . 19
10. Perceptions of problems caused by different drugs in the locales . . . . . . . . . . . . . . . . . . . 20
11. Perceptions of problems caused, overall-rating scores: urban/rural and provincial
comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 20
12. Difference in respondents’ perceptions of differences in the drugs causing the most
problems for men, women and young people in the locale . . . . . . . . . . . . . . . . . . . . . . . 21
13. Usual type of employment (percentage by sample groups) . . . . . . . . . . . . . . . . . . . . . . . . 25
14. Means of financial support in the 12 weeks before interview . . . . . . . . . . . . . . . . . . . . . . 26
15. Sample demographics and first treatment contact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
16. History of drug use (all respondents) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
17. Route of administration (selected drugs only — percentage of whole sample) . . . . . . . . . . 29
18. Summary of injecting practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
19. Sharing of needles and syringes in the last year . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
20. Data of four prisons (all inmates) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
21. Prison data record (inmates who have drug related charges) . . . . . . . . . . . . . . . . . . . . . . 41
22. Comparison of drug treatment services available in 1994/1995 and 2000 . . . . . . . . . . . . . . 47
23. Locales used in the key informant study component . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
24. Locales in the treatment register study and in the four cities study . . . . . . . . . . . . . . . . . 51
25. Basic demographic profiles provided by the census data, using figures for each of the
locales in the study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
26. Reduced overall target population of males in the age band of 15 to 45 years . . . . . . . . . . 54
27. Clinic profiles derived from the treatment register update study . . . . . . . . . . . . . . . . . . . . 55
28. Treatment multipliers derived from key informant and addict interview data . . . . . . . . . . . 56
29. Extrapolated estimates of the numbers of addicts in the country as a whole . . . . . . . . . . . 57
30. Comparisons of heroin abusers in the 1993 and 2000 studies . . . . . . . . . . . . . . . . . . . . . 60

Annex tables

A.II.1. Respondents’ perceptions to the scale of drug use in their locale: rural/urban
comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
A.II.2. Respondents’ perceptions to the scale of drug use in their locale: provincial
comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
A.II.3. Key information perceptions of long-term trends in drug use in their locale
(last 5 years): rural/urban comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
A.II.4. Key informant perceptions of long-term trends in drug use in their locale (last 5 years):
provincial comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
A.II.5. Key informant perceptions of short-term trends in drug use in their locale (last year):
rural/urban comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
A.II.6. Key informant perceptions of short-term trends in drug use in their locale (last year):
provincial comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
A.II.7. Key informant perceptions of problems caused by drug use: urban/rural comparisons . 77
A.II.8. Key informant perceptions of problems caused by drug use: provincial comparisons . . 78
A.II.9. Key informants: attitudes to drug abuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
A.III.1. Basic data provided by the census for each of the locales in the study . . . . . . . . . . . . 83
A.III.2. Reduced overall target population to males in age band 15 to 45 years . . . . . . . . . . . 83
A.III.3. Data from the treatment register update study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
A.III.4. Treatment multipliers derived from key informant and addict interview data . . . . . . . . 84
A.III.5. Extrapolated estimates of the numbers of addicts in the country as a whole . . . . . . . . 85

vii
 Pages
Figures

I. Percentage of key informants that reported drug use as common in their locale . . . ... 15
II. Cumulative long term trend index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 18
III. Cumulative problems arising from drug abuse index, reported by key informants in
each state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
IV. Age distribution of sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
V. Usual route of heroin administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
VI. Severity of dependence: all respondents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
VII. Percentage of sample who inject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
A.II.1. Severity of dependence scores—by sample group . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

viii
Executive summary

The Pakistan National Assessment Study comprises of a set of surveys that each in its own right provides
vital information on drug use in Pakistan; and which, when taken together, also provides a prevalence
estimate of drug abuse. The surveys consist of (a) a national contour mapping exercise—to produce
an overview of patterns and trends, (b) Four provincial city studies of the hard-core heroin using and
injecting population, (c) An audit and update of the information on drug treatment facilities in Pakistan,
(d) An estimation exercise (based on data collected in (a)-(c) for the number of hard-core heroin abusers
and drug injectors in Pakistan.

Key informants
Key informants were asked to report on drug abuse patterns and trends in their local areas. A total of
36 sampling sites (locales) were selected to allow a broadly representative national picture of the drug
abuse. The sample was structured to include 18 matched pairs of rural and urban areas. In total, 283 key
informants were interviewed for this study. Key informants were selected on the basis that they had an
informed understanding of drug abuse patterns in their particular locale.

Cannabis type drugs (hashish and charas) were the drug type most often reported to “be commonly used”
in the locales. Only 5% of respondents reported hashish or charas to be “rarely” used in their locales.
These data indicate that in terms of the number of consumers, cannabis represents the major illicit drug
abused in Pakistan. This is in contrast with the findings of the 1993 survey, which suggested that
heroin was the most commonly used drug. Heroin and alcohol are reported, after cannabis, as the next
two most commonly used substances. Both drugs receive high “commonly used” ratings (46% and 45%),
suggesting that the consumption of both substances is relatively widespread in Pakistan.

The high reported use of cannabis, heroin, and alcohol should not distract from the fact that worryingly
high levels of other types of drug abuse were also reported. In particular, it is a cause for concern that
9% of respondents are reporting the use of “other opiates” as common, that 20% report psychotropics
as “commonly used”, and 12% report drug injection as “common” in their locale.

When hashish and charas are considered, little variation is found across Pakistan either in terms of rural
versus urban comparisons or with respect to a provincial breakdown. Rural areas do report slightly
higher overall figures, as do the provinces of Balochistan, and North-West Frontier Province (NWFP);
lowest provincial figures were from Sindh.

Heroin abuse is reported as a more urban phenomenon (54% of those in urban locales reporting com-
mon use as opposed to 32% of those in rural areas). Distinct provincial differences are also observable,
with heroin abuse most commonly reported in Balochistan, then Punjab. NWFP had the lowest rating
for reported heroin use with only 12% of key informants reporting use that was “common” in their locale.
Drug injection appears least common in NWFP, where no key informants report injection as common in
their locale.

When long-term trends (last 5 years) for heroin use are considered, provincial differences are observable.
Data from Punjab, the province with the highest population and high reported heroin prevalence, sug-
gest a stable or even slightly decreasing trend. This can be contrasted with data from Sindh and

ix
D R U G A B U S E I N PA K I S TA N

Balochistan, which suggest an increase in heroin abuse over this period. In NWFP, little change is reported
and, abuse levels are reported as low. These data would fit with the tentative suggestion that heroin use
is still diffusing in Pakistan to new areas, including more rural areas, whilst at the same time stabilizing,
or even declining, in some of the larger urban centres where use has been longest established.

An upward trend in injecting is reported in all provinces although the increase in NWFP is marginal.
However, Sindh, Balochistan and Punjab all report a strong upward trend. Injecting is reported to be
increasing more strongly in urban rather than rural locales, but for both the overall trends are up.

In all provinces, in both urban and rural locales, and regardless of the relative prevalence rate, heroin
was the drug most associated with causing problems. Eighty-one per cent of respondents reported that
heroin caused “major problems” in their locales. In Punjab province, 88% of respondents reported “major
problems” in their communities resulting from heroin abuse.

After heroin, alcohol received the second highest overall problem score in all provinces with the excep-
tion of Balochistan, where hashish and charas problems received a higher rating. In Punjab, nearly half
(48%) of all key informants reported that alcohol use was causing “major problems” in their community.

Overall, the key informant data suggest that whilst drug abuse remains a more urban phenomena this
difference is not as pronounced as expected. Even in rural areas worrying levels of drug problems were
reported. That said, if hashish and charas are excluded, NWFP, appears to be the province least effect-
ed by drug problems.

This study cannot report comprehensively on the abuse of drugs by women in Pakistan. Key informant
data suggest that, with the exception of psychotropics, for all drug types abuse levels among women are
far lower than those found among men. However, assessing drug abuse among women in Pakistan is a
particularly difficult objective and problems may remain hidden. Many key informants did not feel able
to comment on this question. Even if prevalence levels are much lower the difficulties experienced by
women who abuse are likely to be particularly pronounced and their options for obtaining assistance
limited. The problems that accrue from psychotropic use among women remain unknown. For all these
reasons, a conclusion of this study is that a further research exercise that focuses on substance abuse
among women is required. In any such exercise considerable attention will need to be played to select-
ing the appropriate methods to collect information in this particularly sensitive area.

Addict interviews
Samples of regular heroin users and/or injectors were interviewed in four cities; Karachi, Lahore, Quetta,
and Peshawar. In each city, a sample of drug abusers were recruited in street settings, in treatment facil-
ities and in prisons. One thousand and forty-nine interviews were conducted in total. All but 12 of those
interviewed were men. Subsequent analysis is restricted to the 1,037, male respondents only.

The heroin abusers interviewed were older than in the 1993 exercise. This suggest that the demograph-
ics of the heroin abusing population in Pakistan has changed and is now more similar to age pattern
found elsewhere. The mean age of respondents was between 31 and 33, with 40% aged 25-35 years old.
The age distribution of the sample was similar regardless of the setting in which the interview took place.
Only 5% of respondents were aged between 15 and 20 years old and less than half the sample (41%)
were aged under 30 years old. This contrast with the earlier study, where nearly a quarter of those inter-
viewed were under 20 years old and nearly two thirds, under thirty years old.

Of the three sample groups street addicts were the most socially marginalized; they had lower levels of
education, higher levels of homelessness, lower levels of employment and less family contact. They also
reported more previous prison experience than the treatment sample. Addicts interviewed in treatment
were the most affluent, likely to be in employment and well educated. However, across the samples over-
all, addicts appeared poorer and more socially marginalized than in the 1993 exercise.

x
 E X E C U T I V E S U M M A RY

About half of those interviewed in the streets (52%) and just over a third of those in prison (37%) had
at some stage received treatment for a drug problem. For almost all this was for a heroin problem. The
mean age of first treatment contact for the sample as a whole was 26 years—this did not vary signi-
ficantly between groups. Although, many of the sample had had contact with treatment agencies usu-
ally this contact was brief, for nearly half of subjects 10 days or less. Similar findings emerged from the
interviews with current treatment attendees. This suggest the need to review the provision of community
based services to supplement the inpatient detoxification provision currently provided.

On average respondents had been using heroin for nearly five years (4.6) before entering treatment. Over
half (66%) of subjects reported that at some time they had not been able to access treatment when they
had needed help. The reason for this was usually financial—with 80% of respondents reporting that
they had failed to receive help in the past because of financial constraints. Most (80% of those inter-
viewed on the streets and 72% of those interviewed in prison), reported that they were in current need
of treatment for a drug problem. This study strongly supports the need for enhanced and affordable treat-
ment provision to be made available for those with drug problems in Pakistan.

Overall, patterns of drug taking were remarkable similar across the three groups. The drug users inter-
viewed overwhelmingly report daily and chronic patterns of heroin abuse. Dependence measures sug-
gested a highly dependent population who would be appropriate for entry into a drug treatment pro-
gramme. Poly-drug abuse patterns were common across the samples although heroin was the drug
consumed most frequently, followed by hashish/charas and alcohol. The mean age of first heroin use was
22. This figure was lower for hashish/charas (18) and alcohol (19).

Fourteen per cent of respondents had used synthetic opiate drugs, 11% in the year prior to interview.
The age of first use, 27 was also higher than that found for all other drug types. Although only a minority
of those interviewed were using synthetic opiates, those that did, appeared to be doing so on a regular
basis. Most (80%), were also injecting their drugs. This is a worrying observation and supports anecdotal
reports that a synthetic opiate injecting subculture had developed among a small proportion of the
longer-term users—possibly as a reaction to poor quality street heroin.

Injecting
Nearly three-quarters of respondents reported “smoking” as their usual route of heroin administration
(usually by “chasing the dragon”). However, a worryingly high 15% of respondents reported that they usu-
ally injected the drug. Those using synthetic opiates and morphine commonly reported injection as a
usual mode of administration. This level of prevalence in injecting practice represents a marked increase
from the results found in the 1993 study, where injection was reported by less than 2% of the heroin-
using sample.

On a lifetime measure of injecting 31% respondents reported “ever injecting” and 27% “injecting in the
last year”. Injecting rates varied by city with over half (55%) of those interviewed in Karachi reporting
lifetime injecting prevalence as opposed to only 12% in Peshawar.

On average respondents would inject four times on a typical injecting day. Only 31% reported always
using a new syringe and needle for each injection. Injecting risk behaviour was high, and this was espe-
cially true for the street recruited sample. Over half of all those who had ever injected (51%) reported
using a syringe after others had already used it. For many this was a regular occurrence. More worry-
ingly still, 42% of respondents reportedly regularly passing on a syringe to be used by others. These data
suggest that among the increasing numbers of drug injectors in Pakistan the potential for a dramatic
increase in the infection rate for blood borne viruses, such as HIV, is considerable. Interventions
that target drug injectors should therefore be considered a priority need for Pakistan. It is also highly
desirable to monitor the HIV status of drug injectors and related behavioural information, to better
understand the dynamics of any future epidemic and monitor the impact of any interventions with this
group.

xi
D R U G A B U S E I N PA K I S TA N

Treatment
Seventy-three specialist treatment services for drug abusers were located during the update of the 1994
register. This was considered to represent the majority of current treatment capacity in the country. The
provision offered may have diversified slightly. In the 2000 audit, all services reported providing detoxi-
fication, 59% offered both in and out patient facilities and over half (51%) reported some form of out-
reach provision.

In terms of the number of services identified there was a decline in all sectors (government, private and
NGO), from the number identified in the 1994/95 audit.

The mean number of admissions in the last year before the survey was 264 admissions per facility. The
total number of admissions in this period for all facilities audited was calculated as 17,425. Young peo-
ple and women are less commonly found in treatment facilities. Of the current treatment caseload only
30% of patients were aged 25 or less and only 3% of patients were women. All treatment agencies except
one reported they kept records on their patients that included demographic and drug taking data.
Considerable potential therefore exists for the future development of a treatment-based reporting sys-
tem as a drug use indicator for Pakistan. Such a system would provide a longitudinal data set to allow
consideration of changes in drug abuse patterns over time.

Prevalence estimation
The work to estimate the number of heroin abusers in Pakistan builds on previous national assessment
studies, (1982, 1986, 1993). These earlier surveys have been used to estimate a growth rate for the preva-
lence of drug use in Pakistan by updating the prevalence rate obtained in the immediately previous sur-
vey (using the 1982 data as a base). Over time the estimates formed in this way are likely to drift away
from the true position. In order to avoid this problem the present survey adopts an alternative method
for calculating heroin prevalence. In doing so, the study benefits from some of the recent advances in
statistical methods for addressing the problem of estimating drug prevalence levels.

The prevalence estimation exercise relates only to the chronic and regular use of heroin or drugs by
injection. In this respect the definition is more restricted than that used for other national estimates in
Pakistan, although it reflects common practice for this kind of calculation. Therefore, the estimate could
be revised upwards if occasional or non-chronic users were included.

The best estimates are likely to be for those areas where the most detailed information has been obtained,
that is, in the four cities of Karachi, Lahore, Peshawar and Quetta. Information on other urban areas and
the rural areas that was collected in the other sections of the study in conjunction with national census
data was then used to produce a national estimate.

The upper estimate of approaching 500,000 was calculated as the number of chronic heroin abusers
(including drug injectors) for Pakistan. Given impact of the exclusion criteria in the methods used taking
this upper estimate is appropriate for policy and planning purposes. This figure is lower than previous
estimates but still represents an extremely serious heroin abuse problem in the country. When an analysis
of population levels and the demographics of heroin use in Pakistan is conducted, it is extremely
difficult to see how higher levels than this would be credible. In particular, estimates that are many times
this number would appear extremely unlikely, given the overall population characteristics of the
country.

It should be noted that some less populous geographic areas have been excluded from the prevalence
estimates for a number of technical reasons. For example, Balochistan province outside the Quetta dis-
trict proved too difficult to estimate by the present methods. Prevalence rates quoted in the provisional
figures are for males aged in the target age-band of 15 to 45 years of age, which according to this research
and to previous government figures covers the vast majority of the heroin users in the population.
Heroin use by women is not therefore included in this figure. Future studies that provided information

xii
 E X E C U T I V E S U M M A RY

on these groups could allow them to be included in the estimated total. Similarly, relaxation of the target
population definition to include less intensive patterns of use would result in an increased figure. A
number of factors, among them the possibility of a reporting bias in favour of treatment attendance,
suggest that selecting the higher multiplier estimates was more appropriate than selecting the medium
value.

Within some cities the prevalence rate is estimated to be very high, possibly as great as 4%, compared
with about 0.5% in some rural areas, for males in the target age-band. These figures—up to 1 male in every
25 in the age-band—by most countries’ standards would be considered unrealistically high, but it should
be borne in mind that these are likely maximum figures that are being quoted for the worst affected
Pakistan cities. Even so, previous reports of levels of heroin addiction in Pakistan were considerably
higher again, and as a result fell a long way short of achieving international credibility.

The prevalence of cannabis abuse in Pakistan has not been a primary focus of this study. There were
good methodological resigns for this approach as no single research strategy would be likely to deliver
a robust estimate across these drug types. However, the data do strongly suggest that cannabis preva-
lence is considerable greater than that of heroin and therefore the overall number of consumers of any
type of drug in Pakistan is considerably greater. However, it is not possible to give a credible estimate
of the magnitude of abuse at this time. Further studies would be needed to address cannabis con-
sumption issues in detail and attention is required into the sampling strategies most likely to pay
dividends in this area.

The overall prevalence expressed in terms of the whole population of Pakistan is around one third of
one per cent and is not out of line with other countries expressing prevalence rates of one quarter or
one half of one per cent. There are various special considerations relating to Pakistan in term of the
country’s age structure that make such international comparisons difficult to interpret, but the provisional
figures suggest that, in the international setting, and taking account the population characteristics of the
country, Pakistan has one of the highest rates of heroin abuse documented.

The following should be noted when considering the difference between the estimate provided here and
those derived from previous studies. For all countries prevalence estimation is a challenge and estimates
are refined over time. The estimate given here is no exception and suggestions are given in the techni-
cal annex of how further improvements in the precision of this figure can be made in the future. It
should also be remembered that this study provides an estimate for male regular/chronic heroin abuse
only. Thus, the estimate does not take into account use among women, use in the tribal areas or use
among the refugee population. The previous study used less rigorous criteria for defining heroin use and
it may also be that heroin abuse has fallen, or patterns of use changed, in the intervening years. Certainly
the methods used in the previous studies raise the possibility of considerable upward drift in errors over
time. When comparing prevalence rates between countries it is important to remember that not all the
community are at equal risk. Crude prevalence figures can therefore be misleading. When account is
taken for the demographics of Pakistan (a very young population, the sex ratio and geographical factors)
the new estimate can been viewed as exceedingly high by international standards. The credibility of esti-
mates greatly in excess of this is questionable especially if they are placed in the context of the total
number of males in Pakistan in the most affected age bands.

Part of the purpose of this study was to provide the basis for the further development of an enhanced
capacity to monitor patterns and trends in illicit drug consumption in Pakistan. To this end a number
of suggestions for future activities can be found in chapter 9 of this report. These include, the develop-
ment of a four city drug surveillance network, developing the routine collecting of drug treatment data
to provide an ongoing register, and the incorporation of forensic/drug market data into ongoing surveil-
lance activities.

A number of research topics were also identified as meriting further study. These include solvent use
and other drug problems among vulnerable young people such as street children, patterns of drug abuse
and service needs of women, and the relative prevalence and patterns of use of hashish and charas in
different communities in Pakistan.

xiii
Introduction

Developing effective response to the problem of drug abuse requires a sound understanding
of the nature of the problem. This fundamental conclusion is enshrined in the Declaration of
the Guiding Principles of Drug Demand Reduction, which accompanied the United Nations
General Assembly Special Session (UNGASS). At this historically important meeting it was also
acknowledged that success in the fight against drug abuse requires a balanced approach that
combines both supply and demand reduction measures. Today it is recognized that drug prob-
lems have the potential to affect all countries and therefore national governments have a cor-
responding responsibility to develop sound responses. This assessment study is intended to
help assemble the knowledge base necessary for informing the development and targeting of
drug demand reduction measures in Pakistan.

The need for this enhanced understating was recognized by the government of Pakistan who
through the offices of the Pakistan Anti-Narcotics Force (ANF) took a lead role in instigating
the study and in its implementation. Part of UNDCP’s role is to work in partnership with natio
nal governments to help facilitate the development of responses to drug abuse problems. This
work is guided by the mandates given to UNDCP by the international community and by dec-
larations agreed by member States coming together at the Commission on Narcotic Drugs. It
is therefore particularly appropriate for UNDCP to work in partnership with the government of
Pakistan on this exercise that demonstrates Pakistan’s commitment to both the challenge and
the vision expressed at UNGASS.

This study takes place against a background of considerable political and public concerns about
the problems that drug abuse are causing within Pakistan. The widespread recognition of the
issue of drug abuse in the country is, by international standards, a relatively recent phenom-
enon. For example, heroin abuse was largely unknown in the 1970s but became recognized as
a major social problem during the early and mid 1980s. Similarly, until quite recently there
was no history of injecting drug use in the country. Today it is accepted by all that Pakistan
has a serious drug problem to address. However, the information available to understand, and
therefore to respond to the problem is poor. A number of national assessments have been
undertaken in the past that provided valuable insights. In addition, a number of more
localized research projects have also been conducted. Despite this no good picture exists of
current patterns and trends in drug abuse in the country. The last national assessment dates
from 1993. We know today that drug abuse problems have the ability to develop and change
with alarming speed. There is therefore a clear need for a contemporary understanding of the
current drug abuse problem in Pakistan.

No one-research exercise can answer all questions about the nature, scale and dynamics in
patterns of drug abuse in any country. Nonetheless, the objectives of this study were consid-
erable. Understanding the nature of drug problems is a complex task and the resources avail-
able are always limited. It is also important that the methods used are sound if the resulting

1
D R U G A B U S E I N PA K I S TA N

data are to be accepted as credible. This study does Furthermore, collecting data on drug abuse is not
not therefore attempt to address in detail all the an end in itself. The information is required for
important questions that relate to drug problems in informing the debate on how best to develop and
Pakistan. Rather, a more realistic perspective has target interventions designed to reduce drug abuse
been adopted to focus on those areas of major problems. In the concluding section of this study
importance. The key objective of this study was to those issues that emerge for developing demand
produce a broad overview of current patterns and reduction activities are elaborated.
trends in drug abuse. In addition it has been possi-
ble to look at a smaller number of issues in detail. In summary, the Pakistan National Assessment
In so doing, important questions for further consid- Study is intended to provide an enhanced under-
eration are identified. This research exercise is there- standing of patterns of drug abuse in Pakistan. The
fore configured in line with the UNDCP approach to objectives of this project were ambitious. The
helping countries develop a sustainable capacity to research aims to:
monitor the drug abuse situation, and it is hoped
❏ Provide an overview of patterns of the drug
that this study can form the baseline for future and
abuse;
on-going drug abuse surveillance activities. Pakistan
is one of the priority countries included in the first ❏ Describe perceived abuse problems and
phase of UNDCP’s Global Assessment Programme recent trends in consumption;
(GAP) which assists countries in improving data col- ❏ Improve the understanding of the scale of
lection capacity. The weakness of one-off assess- the problem of chronic drug-abusing;
ments is that they soon become out-dated and are
always partial in their coverage of drug problems. ❏ Make recommendations for the establish-
What is required is the establishment of expert net- ment of an on-going surveillance capacity;
works within countries that can collect and analyse ❏ Identify the key areas and questions for fur-
data on a regular basis and help identify the current ther research activity;
important policy-relevant questions. Policy makers
❏ Review treatment provision in Pakistan;
can then be informed by an on-going debate on how
and
trends in patterns of drug abuse are developing and
on the impact of their interventions. Such networks ❏ Identify knowledge gaps and key develop-
have to be sustainable within the resources avail- ment issues for demand reduction activities.
able and have practical and realistic objectives.
The fact that such a broad agenda could be
One further aim of the current research exercise was addressed is a result of the support and hard work
to identify those areas of data collection that may of a large number of people without whose efforts
prove fruitful for incorporation in an on-going sur- the study would not have been possible.
veillance system for drug abuse in Pakistan. As such Acknowledgements can be found at the end of this
the Pakistan National Assessment Study should be document. This study arose out of concern by the
seen as constituting the initial steps in a longer and Government of Pakistan to ensure that it had the
more ambitious journey. This study is intended not information available to develop effective drug con-
only to provide valuable information about today’s trol policies. As such it represents an important step
drug problem in Pakistan but also to provide a base- forward in the endeavour to address the consider-
line for future assessments and therefore a resource able health, social and other problems that the
for considering ongoing trends. country faces as a result of the abuse of illicit drugs.

2
1

Background

The surveys of drug use in Pakistan preceding the present assessment (1993, 1986) used the
opportunity to estimate a growth rate for the prevalence of drug use, with the prevalence rate
obtained in the immediately previous survey being used as the basis for the next survey’s esti-
mates. The base figures for this procedure are those obtained from the 1982 drug use survey.
Over time the estimates detailed in this way are likely to deviate away from the true position,
partly because any discrepancies that may have occurred in the base figures will be magnified
and partly also because the growth rates themselves are established through key informant
data that can only give a rough idea of the growth that may actually be taking place. In order
not to be held hostage to previous methods and past estimates, the present survey therefore
deliberately seeks to start anew rather than continue to build on the existing figures for drug
use prevalence.

The 1982 survey base figures were essentially obtained by a combination of area sampling to
establish geographical units, and a case-finding technique within these units that was based
on key informant information. Since that time, international research in the area has seen the
development of other techniques beyond simple case-finding for accessing “hidden” popula-
tions that are hard to contact. These newer methods are intended to overcome a recogniza-
ble bias in case-finding techniques that tends to give too low a prevalence rate—although the
performance of this technique when used with key informant identification procedures has not
been documented. The indirect estimation techniques in general centre on types of estima-
tion that in part use other existing data sources as secondary information; this secondary infor-
mation is used in combination with new survey data to give a more reliable picture. These
methods are often termed benchmark/multiplier techniques in this context, and it is one of
these techniques that is implemented in the Pakistan 2000 survey.

The method of benchmark/multiplier estimation that is used therefore attempts a fresh esti-
mate of drug use prevalence in order not to compound further any possible drift that has
occurred in the latest available estimates. It further tries to move beyond the case-finding
methods to attempt to make use of other data sources that are available.

Pakistan in context
In 1998, the population of Pakistan was estimated to be 135 million inhabitants, ranging from
densely populated urban districts to sparsely populated mountain villages. The territory of
Pakistan comprises the provinces of North West Frontier, Balochistan, Sindh and the Punjab
together with the Federally Administered Tribal Areas (FATAs) and the Federal Capital Area of
Islamabad. The main ethnic groupings are the Punjabis, Sindhis, Balochis and Pathans plus a
number of smaller tribes in the remote northern areas. The Mohajirs, or refugees from India

3
D R U G A B U S E I N PA K I S TA N

at the time of independence, are sometimes The drug abuse problem

considered an ethnic group. A recent addition to
the overall population are Afghan refugees, esti- Pakistan is confronted with a significant drug abuse
mated in 1997 to total about 1.75 million, who are problem. Trends in drug use patterns indicate a
mainly concentrated along the tribal areas near the marked increase in heroin consumption that
border. emerged during the 1980s. The 1993 National Survey
on Drug Abuse, whose results remain by no means
undisputed, estimated the number of drug users at
3 million of which approximately 50% were addict-
Drug cultivation and production ed to heroin. Other estimates have suggested that
the total number of drug abusers may have reached
Pakistan has been a producer of opium for export 4 million by 2000.
and traditional domestic consumption since the
time of Muslim rule and the later British Empire. In Recent trends suggest a shift from inhaling and
1979, the Government of Pakistan responded to the smoking heroin to injection of drugs, in particular
problem of increased illicit opium production and pharmaceutical drugs, bearing the high risk of an
trade by the enforcement of the Hadd Ordinance. HIV/AIDS epidemic and the spread of other blood
The Hadd Ordinance brought existing law into line borne diseases. Injecting drug use has been report-
with Islamic injunctions and prohibits trafficking, ed from all major cities and some Afghan refugee
financing or possession of more than 10 grams of camps around Peshawar and Quetta. Results from a
heroin or 1 kg of opium. In 1979, all poppy cultiva- study commissioned by UNDCP/UNAIDS in 1999,
tion (licit and illicit) was banned and all Government indicated that sharing and multiple use of injection
controlled processing plants and retail outlets for needles is common practice. Some evidence exists
licit opium were closed. As a result of the Hadd that “shooting galleries” have become established in
Ordinance and partly because of massive stock pil- Lahore and in Karachi. No cases of HIV/AIDS have
ing of opium following a bumper harvest in 1979, been detected in that study, but the high prevalence
opium cultivation and production sharply declined of Hepatitis-C (180 out of 200 cases) indicates the
in the 1980s. The Government’s commitment to enormous potential of an HIV/AIDS epidemic and
make Pakistan poppy free, increased efforts in law other transmittable diseases among the injecting
enforcement, the impact of alternative development drug users.
assistance from the international community, and a
drop in retail prices for opium gum due to the mas- The above study has prompted the undertaking of
sive increase in production in Afghanistan, are major two joint UNDPC/UNAIDS pilot projects in Lahore
factors that contributed to a further decline in opium (launched in March 2000) and Karachi (estimated
cultivation since the mid 1990s. An analysis of poppy start January 2001) aiming at HIV/AIDS prevention
harvesting trends at the national level reveals a among injecting drug users.
decline in the amount harvested from 9,441 ha in
1992 to less than 284 ha in 1999. Of the three main In order to enhance the knowledge base on women
poppy growing areas, Dir district where UNDCP has and drug abuse in Pakistan, UNDCP commissioned
been active since 1985, accounted for approximate- a study targeted at drug abusing women in 1998. In
ly 60% of the opium harvested in the country. Over total, 98 drug addicted women from Karachi and
this period, UNDCP spent US$ 34 million on alter- Lahore were surveyed. Questionnaire results were
native development projects in Dir District. complemented by interviews and focus group dis-
Alternative development interventions coupled with cussions. Approximately one-third (32) of the
demonstrated Government commitment led to a women who participated in the study regularly took
decrease in opium poppy cultivation in Dir District heroin. Tranquilizers (26) were the second most pre-
from over 3,500 ha in 1992 to near zero in 2000, mak- ferred drug, followed by hashish (16), alcohol (11)
ing the Dir project one of UNDCP’s most successful and opium (5). Heroin is mainly taken in the form
alternative development projects. of prepared cigarettes or by inhaling the fumes. No

4
 BACKGROUND

injection of heroin was reported. The heroin addicts were unable to cope with their addiction and
came from various backgrounds ranging from work- arranged for their imprisonment. Treatment services
ers with post-graduate qualifications to illiterate in prisons, if available, are limited to medical inter-
women. Women were introduced to heroin at an vention to bring relief from acute withdrawal symp-
average age of 22 years mainly through friends, hus- toms. Prison authorities admit that they do not have
bands and relatives. Peer pressure and a stressful the capacity to properly deal with the problem of
life were frequently mentioned as causes for heroin drug addicts.
abuse. Five out of the 32 heroin using women had
been arrested at least once. Women mentioned that Relatively little research has been done on the con-
they found it difficult to receive treatment or coun- sequences of drug abuse in the country. There are
selling on drug abuse as none of the established no official statistics or reliable reports available on
treatment centres provide specialized services for the socio-economic impact of drug abuse, drug-
women. related deaths or drug-related violence. Very recent
reports indicate the availability of high purity
Overall, drug addicts have little access to effective heroin in some of Pakistan’s drug-markets, which
treatment. With a few exceptions the services pro- has caused several deaths among drug users.
vided by government-run drug treatment facilities Generally, however, the purity of heroin available to
are limited to the management of acute withdrawal street addicts in Pakistan is believed to be quite low.
symptoms of 7-10 days duration. Most of the drug However considerable geographical and temporal
treatment facilities in the public sector are situated variation may exist in the composition of heroin
within the departments of psychiatry in teaching or available on the illicit market. As this factor may
district headquarters hospitals. The staff within directly influence the nature of drug abuse problems
these facilities are trained in psychiatry and there- Pakistan is faced with, there is a need to develop
fore largely view addictive behaviour from this per- methods for monitoring the composition of drugs
spective. Specialist training in the management of on the illicit market. This topic is returned to in a
substance misuse problems is rare. NGOs differ in later part of this report.
levels of development and capacity in terms of pro-
viding drug treatment. Generally, NGOs are more
receptive to developing new broader treatment Status of convention adherence
responses and providing a range of services beyond
medical interventions to their clients. However, Pakistan is a signatory to the United Nations Single
again staff often are not specially trained and there Convention on Narcotic Drugs, 1961, the 1971
is a need to develop a better understanding of the Convention on Psychotropic Substances and the
issues in offering interventions such as social and 1988 United Nations Convention against Illicit
vocational rehabilitation. Only very few well estab- Traffic in Narcotic Drugs and Psychotropic Sub-
lished NGO and private clinics offer comprehensive stances. In February 1999, the Government agreed
treatment packages including rehabilitation and to ratify the 1972 Protocol, which amends the 1961
social reintegration services, but these services often Convention.
are too expensive for the average street addict to
access. Whilst no systematic follow-up is done on
the success of treatment, readmission of relapse Legislation
cases in public institutions are reported to be as
high as 90%. The Control of Narcotic Substances (CNS) Act, 1997,
arising from an ordinance bearing the same name
It is estimated that approximately 20% of Pakistan’s and promulgated in 1995, effectively covers all
prison population has been incarcerated because of aspects of Pakistan’s drug control efforts. It deals
drug abuse, possession of drugs and other drug- with cultivation, manufacture, production, trafficking
related offences. Many young drug addicts find and possession offences as well as with treatment
themselves in prison because their family members and rehabilitation of drug addicts (see below). The

5
D R U G A B U S E I N PA K I S TA N

Act also allows the Government to set up Special enforcement, two refer to demand reduction and one
Courts with exclusive jurisdiction in drug matters to supply reduction. The financial requirements for
and to establish a National Fund for the Control of the five-year duration of the plan have been esti-
Drug Abuse to be partially funded from assets for- mated at Rs 2,832 million (approx. US$ 56 million).
feited under the legislation. Provision is also made Federal and Provincial Governments, United Nations
for the mandatory reporting by banks and financial agencies, multi- and bilateral donors have been indi-
institutions of suspicious financial transactions. cated as possible sources of funding.

Chapter VI of the CNS Act 1997 deals with treat- National institutional framework
ment and rehabilitation of addicts:
and capacity
Article 52 stipulates that Provincial Governments
The Narcotics Control Division: Pakistan’s drug control
shall register all drug addicts for the purpose of
policy making and planning is the responsibility of
treatment and rehabilitation while the Federal
the Narcotics Control Division which forms part of
Government is held responsible to bear the cost
the Ministry of Interior and Narcotics Control which
for first-time compulsory detoxification or de-
was created in 1989. The Narcotics Control Division
addiction of an addict.
is headed by a Secretary who is UNDCP’s designat-
ed government counterpart.
Article 53 requests the Provincial Governments to
establish as many treatment centres as neces-
In the Anti Narcotics Force Act 1997, the Anti
sary for detoxification, de-addiction, education,
Narcotics Force (ANF) has been given primary
after-care, rehabilitation, social integration of
responsibility for interdicting the production, smug-
addicts and for supply of such medicines as are
gling, trafficking and abuse of narcotics substances
considered necessary for the detoxification of
and illicit psychotropic substances. Although the
the addicts.
Anti Narcotics Force Act does not make direct refe-
rence to demand reduction activities, ANF continued
In 1998, the CNS Act was extended to the Federal
to implement demand reduction projects which had
and Provincial Administered Tribal Areas, but over-
been previously implemented by the Pakistan
all implementation of the Act has progressed rather
Narcotics Control Boards (PNCB). The Drug Abuse
slowly. Up till the present time, no provincial regis-
Prevention Resource Centre (DAPRC) that was estab-
tration of addicts has taken place and the provincial
lished under the PNCB with significant assistance
treatment centres referred to in Article 53 are yet to
from USAID, has been maintained by the ANF and
be established.
has been quite active as long as assistance from
donors, mainly from the United States and the
National drug control policies, European Commission, was forthcoming. Currently,
there are two departments within ANF that deal with
priorities and plans
demand reduction: DAPRC and the Planning and
Development Department. Both Departments are
In 1995/96, the Government of Pakistan with assis-
headed by a Director who report to the Deputy
tance from UNDCP prepared a comprehensive
Director-General of ANF.
Master Plan for Drug Abuse Control: 1998-2003. In
1998, UNDCP revitalized this plan and advocated at
the highest level for its speedy approval. Following
minor amendments, the Master Plan was approved Socio-economic characteristics
by the Prime Minister’s Cabinet in February 1999. of Pakistan
The Master Plan is structured around six objec- Poor social and economic conditions in Pakistan
tives and related strategies, outputs and activities. have a broad impact on Pakistan’s overall health
Three of the objectives aim at strengthening law situation and increase vulnerability of the general

6
 BACKGROUND

population to drug problems. Recent estimates indi- drugs. A number of socio-economic factors such as
cate that 34% of the population live below the pover- youth, unemployment, large disparities between
ty line, and this percentage is even higher for those income groups, poverty, urbanization are known to
in rural areas. Poverty is inextricably linked to an be linked to drug abuse and crime. But as yet, little
array of difficulties that reduce the life chances of is know about the relationship of these factors to
individuals and overall health and well being of drug abuse problems in Pakistan. Furthermore, no
households and communities. For example, those ongoing surveillance information is available to
who are poor often have the least access to educa- monitor trends over time. For example drug related
tion and social services, and therefore least access overdoses or deaths are also not recorded in the
to the information and tools which might help them country nor is there a centralized register on the
protect themselves from drug abuse and other health behaviour of treatment attendees. The involvement
and social threats. In addition, poverty sometimes of drug addicts in criminal activity is also not meas-
influences people to make choices that make them ured. A high level of poverty and other pressing
more vulnerable to infections, for example, the frus- needs on the public purse makes it difficult to invest
trations related to poverty can drive people to abuse in prevention and treatment activities.

7
2

Rationale, methods and sampling

Rationale of the Pakistan national assessment exercise 2000

As the first national assessment exercise since 1993, the range of possible topics for inclusion
in this study was considerable. Drug abuse impacts on many aspects of society and the infor-
mation needed to inform a policy response is correspondingly diverse. However, successful
studies require sound methods that reflect reasonable objectives. It was therefore important
that the aspiration for a comprehensive and detailed coverage of all aspects of the drug abuse
phenomenon be balanced against a careful consideration of what can be realistically achieved
within the resources available.

Technical decisions on the design of the study were based on the initial set of important pol-
icy questions identified by the Government of Pakistan in their discussions with UNDCP. From
these discussions it was clear that the study design would have to both address general ques-
tions about patterns and trends in abuse across the country and at the same time focus on
a number of more specific research questions. These two requirements placed against the
available budget dictated the research strategy. An initial decision was taken that the study
would consist of three distinct, but interrelated, research exercises. In addition, each of these
exercises would collect information that could be subsequently combined to produce an esti-
mate of the hard-core heroin abusing and injecting population.

The study would therefore comprise of the following elements:

(a) A national contour mapping exercise—to produce an overview of patterns and trends. For
this exercise a national drawn sample of key informants were interviewed.
(b) Provincial city studies of the hard-core heroin using and injecting population. For this aspect of
the study samples of heroin (or other opioid users) and/or drug injectors were inter-
viewed in a major city in each of the four provinces. The cities were Karachi, Lahore,
Quetta and Peshawar. Three samples were drawn in each city. The groups sampled
were drug abusers in treatment, drug abusers in prison and drug abusers contacted
on street settings.
(c) An audit and update of the information on drug treatment facilities in Pakistan.
(d) An estimation exercise (based on data collected in (a)-(c) on the number of hard-core
heroin abusers and drug injectors in Pakistan.

In summary: the method of the national assessment exercise is to carry out a set of surveys
on particular aspects of the drug problem that each in its own right provides vital informa-
tion on drug use in Pakistan; and which when taken together also can provide a prevalence
estimate of drug use. For a full discussion of the methodological issues relating to this study
please refer to annex I.

9
3
Mapping the national contours of drug abuse:
the key informant interviews

Key informant interviews sample description

In this section data are presented from the national mapping exercise. Key informants were
asked to report on drug abuse patterns and trends in their local areas. A total of 36 sampling
sites (locales) were selected to allow a broadly representative national picture of the drug
abuse situation to be produced. The sample was structured to include 18 matched pairs of
rural and urban areas.

The tables referred to in this section are numbered sequentially. Those tables (tables A.II to
A.II.9) where the table number is preceded with the letter “A” can be found in annex II of this
report.

A key informant approach was selected as the most appropriate method for generating a natio
nal picture of patterns and trends in drug abuse. In total, 283 key informants were interviewed
for this study (on average seven respondents per locale). This corresponds to the study design
that aimed for a minimum of five key informant interviews per locale, but allowed extra inter-
views to be conducted where other particularly appropriate individuals could be located. The
sampling strategy was designed to ensure that a range of respondents, who had knowledge of
drug abuse from different perspectives, were included in the sample. By including such diverse
occupational groups as teachers, medical workers and police in the sample, it was hoped that
a fuller picture of the local drug scene would be produced—as each group would have a dif-
ferent perspective on patterns of local drug abuse. Key informants were not simply selected
by occupational group. Interviewers were instructed to identify key informants in each locale
who were most likely to have had contacts with drug abusers and therefore be best placed to
describe the local situation.

The vast majority of respondents were male (94%), with only 17 females being interviewed in
total. Whilst the sampling strategy matched rural and urban locales in practice more suitable
individuals were, perhaps unsurprisingly, located in urban sites. In the final sample 64% (181)
of respondents were classified as reporting from urban sites, and 36% (101) from rural ones.
It was not possible to code one interview on the urban/rural dimension and it was therefore
excluded. Urban-rural comparisons are therefore made here on an achieved sample size of
282 responses. This represents a mean of six interviews per rural site as opposed to 10 per
urban location. In both cases, the minimum target sample of five interviews per locale was
achieved. In the analysis that follows urban versus rural breakdowns are provided where this
variable impacts on the interpretation of the data.

Sampling sites were drawn across the four provinces of Punjab, Sindh, Balochistan, and North-
West Frontier (NWFP). In Punjab, 86 interviews were conducted in total (65, urban and 21, rural),
in Sindh 64 interviews in total (49, urban and 15, rural), in Balochistan 30 interviews in total

11
D R U G A B U S E I N PA K I S TA N

(19 urban and 11 rural) and in NWFP 88 interviews Table 1. Key informants’ occupational
were conducted in total (42 urban, and 46 rural). A groups
small number of interviews (15) were conducted out-
side of these provinces or had coding problems and Local location
List of
are therefore not included in a provincial analysis. occupational group Urban Rural Total
The sub-sample total for all provincial comparisons Medical 37 10 47
is therefore 268. The reader should note that due to 21% 10% 17%
sporadic missing values or non-response to individ-
ual questions, the actual sub-sample numbers used NGO/Gov. 34 7 41
19% 7% 15%
in individual comparisons will vary. The reader
should also note that due to rounding percentages Police 26 11 37
may not always exactly sum to 100. 15% 11% 13%

Community leaders 5 7 12
In summary: the overall number of key informants
3% 7% 4%
interviewed was 283,181 of whom were classified as
reporting from urban locales, and 101, from rural Ex-addict 17 10 27
locales, thereby giving a base of 282, for urban/rural 10% 10% 10%
comparisons. A provincial breakdown was possible
Teacher 10 7 17
for 268, interviews and provincial comparisons are 6% 7% 6%
correspondingly based on this number. Sporadic
missing values, non-responses, and non-applicable Welfare organization 13 5 18
questions result in some variation in individual sub- 7% 5% 7%
sample numbers used for comparisons.
Local business 17 25 42
10% 26% 15%
In table 1 data are presented on the occupational
classification of the key informants interviewed for Religious leader 6 9 15
3% 9% 5%
this study. It was important to interview key inform-
ants with a range of occupational backgrounds. The Other qualified workers 11 4 15
study was successful in this respect. Furthermore, 6% 4% 5%
the mix of occupational groups is broadly compara-
ble between the urban and rural samples. Some Other 2 2 4
1% 2% 1%
minor differences do exist in the samples. For exam-
ple, more medical personnel (21% of urban sample Total 178 97 275
versus 10% of the rural sample) were interviewed in 100% 100% 100%
urban settings, probably reflecting the dispropor-
tionate location of medical facilities in urban areas,
and more religious leaders were interviewed in rural comment with authority on longer-term changes in
areas (9% of rural sample versus 3% of urban sam- local patterns and trends in drug abuse.
ple).
To comment accurately on patterns of drug abuse
As key informants were expected to comment on within the locale, it was also desirable that key
trends over time it was desirable that they had long- informants had come into contact in some way with
term local knowledge. For the majority of respon- drug abusers in their local community. As noted
dents this was the case. On average, key informants above, interviewers were instructed where possible
had been living in their locale for 24 years (28 years to seek out such individuals for inclusion in the
for rural respondents and 21 for urban), with only a sample. Encouragingly, respondents reported con-
small minority of either sample having been resident siderable contact with drug abusers. This question
in the locale for less than 5 years (13%). This allows was asked separately for all “drug abuse” and speci-
for some confidence that the key informants could fically for “heroin abuse and drug injection”. In the

12
 M A P P I N G T H E N AT I O N A L C O N T O U R S O F D R U G A B U S E

Table 2. Number of contacts by However, the most striking difference here relates to
respondents in the last year the provincial analysis. Respondents in NWFP
with drug abusers reported on average approximately the same level of
contact with “drug abusers in general”, but far lower
Mean number Mean number contact with “heroin abusers”. This point is borne
of contacts of contacts
Any drug Heroin or IV only out further by noting that no key informants in
Punjab or Balochistan, and only one respondent in
All 83 44
Sindh, reported “no contact” with heroin users in the
Urban locales 89 50 twelve months before interview, whilst in NWFP this
Rural locales 70 31 figure was 13% of all the key informants interviewed.
Punjab 86 61
Sindh 87 53
Balochistan 75 43
NWFP 77 13
Perceptions of drug use
in the locale

12 months prior to interview, on average key inform- Key informants were asked for each of the index
ants reported contact with over 80 drug abusers of drugs considered by the study how “commonly used”
who about half were heroin users. These data are that drug was in the locale. Respondents had the
given in table 2. Data provided in this table should opportunity to respond across the following scale:
be interpreted with caution as some occupational “commonly used”, “some use” “rarely used” or “don’t
groups (for example the police or medical person- know”. The “don’t know” category was included as an
nel) reported contact with extremely large numbers active category (i.e. it was read out as part of the
of drug users. However, two points are worth noting. scale) so as not to encourage respondents to answer
Contact with drug abusers as a group, whilst vary- arbitrarily.
ing, was high for the sample as a whole. Only four
individuals (1.5% of the sample) reported no contact In table 3 data on the overall perception patterns of
in the preceding year with the users of any drug, and use can be found and in figure I the percentage of
only 11 (4%) individuals did not report any contact respondents reporting that a drug was commonly
with “heroin abusers” in the preceding 12 months. used in their locale is presented. Cannabis type
Contact was rarely limited to just a few individual drugs (hashish and charas) were most often report-
drug users. Ninety-two per cent of key informants ed to be commonly used in the locales. This is in
reported contact with 10 or more “abusers of any accord with the general impressions of the research
drug” in the last year. Slightly less, but still over two- team after the field visits to Pakistan but in conflict
thirds (72%) of the sample, reported contact with 10 with the picture produced by the 1993 research exer-
or more “heroin abusers” in the previous twelve cise. In this study only 5% of respondents reported
months. These data suggest that drug abuse in gen- hashish or charas were rarely used in their locale
eral, and the use of heroin in particular, is both geo- suggesting this drug type represented the major
graphically widespread in Pakistan, and that signifi- illicit drug consumed in Pakistan in respect to the
cant numbers of individuals are involved. The data total numbers of consumers.
also suggest geographical differences exist in respect
to consumption patterns. After cannabis type drugs, heroin and alcohol are
reported as the next two most commonly abused
Whilst the average number of abusers of “any drug” substances. Both drugs receive high “commonly
that the key informants had contact with did not used” rating (46% and 45%) suggesting that the con-
vary significantly between locales, this is not the sumption of both substances is relatively wide-
case when “heroin abusers” are considered. Those spread in Pakistan. It is worrying to note that only
in rural locales reported less contact with “heroin 8% of respondents reported that heroin abuse was
abusers” than key informants in urban areas. rare in their locale.

13
D R U G A B U S E I N PA K I S TA N

Table 3. Respondents’ perceptions of the scale of drug use in the locales

Drug Type Commonly used Some use Rarely used Don’t know

Hashish and charas 194 (70%) 67 (24%) 13 (5%) 2 (>1%)

Heroin 127 (46%) 111 (40%) 21 (8%) 16 (6%)
Opium 28 (10%) 91 (34%) 121 (45%) 31 (11%)
Other opiates 24 (9%) 34 (13%) 75 (29%) 130 (49%)
Cough syrups
(for intoxication) 31 (11%) 24 (10%) 73 (28%) 136 (52%)
Psychotropics 54 (20%) 88 (33%) 80 (30%) 48 (18%)
Drug injection 30 (12%) 73 (28%) 63 (24%) 95 (36%)
Solvents (glue, etc) 10 (4%) 36 (14%) 61 (24%) 147 (58%)
Alcohol 121 (45%) 92 (34%) 54 (20%) 3 (1%)

The high perception of the use of cannabis, heroin, In interpreting the provincial comparisons it is
and alcohol should not distract from the fact that important to bear in mind that these data relate to
high levels of other types of drug abuse are also areas with distinctly different population totals. Thus
reported. In particular, although relatively low, it is the relative implications for the scale of the drug
still a cause for concern that 9% of respondents are abuse problem in terms of the number of individu-
reporting the use of “other opiates” as common, that als affected is likely to correspondingly vary. For
20% report psychotropics as “commonly used”, and example, Balochistan has a population of 6,511,000
12% report drug injection as common in their locale. as compared to 72,585,000 for Punjab. Therefore, the
However, caution should be exercised when inter- reported “common use” of any substance in
preting these figures because the “don’t know” rate Balochistan will affect far fewer individuals than will
increases dramatically for these drug types. For its “common use” in Punjab. The population totals
example, nearly half the sample (49%) cannot com- for NWFP and Sindh are 17,555,000 and 29,991,000
ment on the use of “other opiates” and over-half respectively (all figures based on 1998 census).
(52%) cannot comment on the use of cough syrups
for the purposes of intoxication. To a large extent it In annex table A.II.1, key informant’s perceptions of
may be that these responses indicate that this pat- local consumption patterns are compared by the
tern of use is uncommon in the locale. However, it urban versus rural classification of the locale. In
is also probable that some abuse patterns, such as annex table A.II.2, data is presented separately for
the use of psychotropic substances by women (for each province. As visually comparing tabulated data
example benzodiazepine use), is more hidden than of this type across tables is difficult, in table 4 an
the use of some of the other drug types listed. overall rating score has been calculated that com-
bines all the information from the table. Comparing
It is also important to note that distinct geographi- ratings scores in this table allows the relative
cal differences are often observable in patterns of responses to be compared across drug type and
drug abuse and that drug consumption does not locale classification. “Don’t know” responses have
therefore have to be a national phenomenon to been included in this analysis. The higher the over-
cause acute local problems. This is one reason that all rating the more the index drug was reported as
many drug information systems are configured to used in the locale.
collect data at city or local level, thereby informing
the delivery of local responses, as well as providing When hashish and charas are considered, little
information for national estimates. To explore dif- variation is found across Pakistan either with respect
ferences in abuse patterns in Pakistan, in this study to a comparison of rural versus urban areas or
it is possible to make both rural versus urban and with respect to a provincial breakdown. In urban
provincial comparisons. areas, 71% of respondents report use is “common”

14
 M A P P I N G T H E N AT I O N A L C O N T O U R S O F D R U G A B U S E

Table 4. Respondents’ perceptions of the scale of drug use in the locales:

overall ratings

Drug type All Urban Rural Punjab Sindh Balochistan NWFP

Hashish and charas 16 17 16 16 15 19 17

Heroin 13 15 11 16 15 18 9
Opium 5 5 6 6 4 9 5
Other opiates 3 4 1 7 2 1 1
Cough syrups
(for intoxication) 3 3 3 8 1 1 2
Psychotropics 7 8 5 10 8 3 5
Drug injection 5 6 4 9 5 7 1
Solvents (glue, etc) 2 2 2 6 1 1 0
Alcohol 12 13 11 16 12 13 8

Note: Higher values indicate that use is reported to be more common.

Figure I. Percentage of key informants When heroin is considered, urban and rural differ-
that reported drug abuse as ences are more apparent, as indicated by overall rat-
common in their locale ing scores of 15, and 11, respectively. Heroin abuse
is thus reported as a more urban phenomenon with
100 54% of those in urban locales reporting common use
90 as opposed to 32% of those in rural areas. Distinct
80 provincial differences are also observable. Heroin
70
use is reported to be most common in Balochistan,
60
where 77% of respondents report use as “common”.
50
However, sample sizes are disproportionately lower
from this province reflecting its lower population
40
total (only 30 key informants interviews conducted),
30
so some caution is necessary in making comparisons
20
between Balochistan and other provinces. Rates are
10
also high in Punjab where 58% of key informants
0
report “common use” and this is supported by a
Other opiates

Psycotropics
Cough syrups

Solvents

Alcohol
Opium
Hashish

Heroin

injection
Drug

correspondingly high overall rating figure of 16. It is

important to look at this overall rating figure, in
addition to individual scores, as this measure takes
into account information from all responses to the
question (i.e. “commonly used,” “some use”, “rarely
used” and “don’t know”). NWFP has the lowest rat-
compared to 69% of those in rural locales. Similarly,
ing for reported heroin use with only 12% of key
provincial levels of reported use are also high,
informants reporting use was common in their
although some differences are observable. Use is
locale.
reported to be most common in Balochistan, where
90% of respondents report hashish or charas as
“commonly used” closely followed by NWFP, where Drug injection is the route of administration most
the figure is 75% of responses. Lowest use is report- associated with health and other problems. As such,
ed in Sindh (overall rating of 15), but here still nearly even the existence of a relatively small injecting
two-thirds of key informants report the drug type is population can have considerable impact on the
“commonly used”. overall costs (health, social and criminal justice)

15
D R U G A B U S E I N PA K I S TA N

associated with drug problems. Historically, in Drug abuse patterns are commonly known to vary
Pakistan drug injecting has not been common across demographic dimensions. In this study, to
although, as noted earlier in this report, some evi- explore variations by sex and age, key informants
dence exists to suggest this now may be changing. were asked to separately rate the drug “most com-
At least for some parts of the country the data pre- monly used” by “men”, “women” and “young people”,
sented here support this view. Drug injecting is (the last defined for the purposes of this study as
reported to be common in urban rather than rural less than 25 years of age). These data can be found
locales and in the province of Punjab. Drug injection in table 5.
appears least common in NWFP, where no key in-
formants report injection as common in their locale. Table 5. Respondents’ perceptions of the
differences in the drugs most
Opium has a long history of use in Pakistan and has commonly used by men, women
in the past been widely used by traditional medical and young people
practitioners (Hakim) to treat a range of minor ail-
ments. However, anecdotal accounts suggest that Young
Drug type Men Women people
the use of the drug as an intoxicant has declined in
recent years. In the current research exercise in Hashish and
Balochistan, opium was the drug most often report- charas 140 (51%) 12 (10%) 177 (65%)
ed to be commonly used, although once more the Heroin 74 (27%) 33 (27%) 46 (17%)
small sample sizes suggest this result should be Opium 5 (2%) 16 (13%) —
interpreted with caution. Other opiates — 1 (>1%) —
Cough syrups
There was some tendency to report higher rates of (for
alcohol use in urban as opposed to rural areas intoxication) — — 1 (>1%)
although this difference was not particularly pro- Psychotropics 3 (1%) 45 (37%) 4 1%)
nounced—in urban and rural areas a significant Drug injection 7 (3%) 7 (6%) 7 (3%)
number of key informants reported that alcohol was Alcohol 47 (17%) 8 (7%) 39 (14%)
commonly used. Alcohol use was less often report-
ed as commonly used in NWFP than in other Note: Sample N’s (men 276, women 122, young people 274).

provinces, with only 19% of key informants report-

ing use was common in their locale. This is a marked As would be expected, male drug use in terms of
contrast with Punjab where 71% of key informants most commonly used substance broadly reflects the
report the use of alcohol as common and only 3% overall pattern discussed above. Differences from
that its use was “rare” in their locale. the overall picture are evident when women or young
people are considered. It should be noted that the
Overall, the key informants reports would suggest number of respondents falls to 122 for the question
that drug abuse, with the exception of opium, is on women’s drug use. This is because many key
more a feature of urban rather than rural life in informants did not feel competent to answer this
Pakistan. That being said, this difference is not as question. This is probably due to both the facts that
pronounced as expected, and abuse of other drugs drug abuse among women is a lower frequency activ-
in many rural locales was reported at worryingly high ity than male drug abuse, and that drug use among
levels. Similarly, provincial differences in reporting women is a more hidden behaviour. Psychotropic
rates are clearly observable, with the Punjab, the drugs were the type most commonly reported to be
largest province in population terms, having used by women, followed by heroin. It should be
amongst the highest levels of reported common use. remembered that this question does not relate to
No province appeared to be free of drug abuse prob- the overall numbers of abusers, but rather to the
lems. However, if hashish and charas are excluded, drug most commonly used amongst those who do
NWFP overall reported that use was less common abuse drugs in the locale. It is therefore debatable
than elsewhere in Pakistan. what this result implies about the overall prevalence

16
 M A P P I N G T H E N AT I O N A L C O N T O U R S O F D R U G A B U S E

of heroin use among women in Pakistan. a provincial breakdown is provided. As an aid to con-
Interestingly, unlike the male pattern of use, hashish sidering the data as a whole an overall rating has
or charas is not commonly reported as the main been computed, which can be found in table 7. In
drug type used by women in most locales. the overall measure, negative values have been
placed in brackets and indicate that key informants
Among young people, hashish or charas is reported are reporting a decrease in use.
to be the most common drug used by 65% of key
informants, followed by heroin and alcohol (17% and Some caution is required when reflecting on per-
14% respectively). The abuse of solvents, which is ceived trends over time, especially long time inter-
elsewhere commonly associated with younger ages vals, as memory effects and other factors may influ-
groups, or street children, was not reported as the ence perceived changes. It can be argued that when
main drug used by young people by any informant. social problems such as drug abuse or crime are
considered, respondents are often more inclined to
Trends in drug abuse report negative rather than positive changes. None-
theless, these data do provide a useful overview of
Key informants were asked to reflect on trends in observed trends and have been used in the previous
drug abuse in their locales. Two time periods were national assessment exercises for quantifying annu-
used for these questions. For each index drug, al increases in prevalence rates. In interpreting the
respondents were asked whether use had “decreased data it is also important to look closely at the
a lot”, “decreased a little”, “not changed”, “increased provincial breakdown. Trends must be viewed with
a little”, or “increased a lot” in their locale. The first respect to their population base. For example, a
time period respondents were asked to assess small upwards or downwards trend in an area of high
changes over was “the last five years” (i.e. from prevalence will have far more impact, in terms of the
1995). This is long term measure approximating the number of individuals affected, than a similar trend
period since the last national assessment exercise in an area of low prevalence.
was conducted (1993). “Five years” was selected as
the maximum reasonable long-term recall period. With this in mind data on heroin trends in Punjab
The second set of questions covered the same topic is particular interesting as this province contains
but asked for short-term trends. In this case a stan- some of the large cities where heroin problems are
dard “last year” reporting period was adopted. most apparent and where the highest provincial
population rates are found. The overall rating is 0
Data on long term trends for the whole sample can for this province indicating overall that responses
be found in table 6. In annex table A.II.3. urban suggesting increases in use are balanced with those
versus rural comparisons are given and in table A.II.4 suggesting decreases.

Table 6. Perceptions of long-term trends in drug use in the locales (last 5 years)

Decreased Decreased Increased Increased Don’t

Drug type a lot a little No change a little a lot know

Hashish and charas 7 (3%) 32 (12%) 34 (12%) 97 (35%) 102 (37%) 4 (1%)
Heroin 19 (7%) 57 (21%) 28 (10%) 77 (28%) 78 (29%) 13 (5%)
Opium 15 (6%) 38 (14%) 132 (50%) 36 (14%) 19 (7%) 25 (9%)
Other opiates 7 (3%) 8 (3%) 43 (17%) 41 (17%) 19 (8%) 128 (52%)
Cough syrups 7 (3%) 16 (7%) 27 (11%) 34 (14%) 26 (11%) 134 (55%)
Psychotropics 4 (2%) 12 (5%) 31 (12%) 101 (39%) 68 (26%) 44 (17%)
Drug injection 3 (1%) 11 (4%) 31 (12%) 75 (30%) 37 (15%) 96 (38%)
Solvents 2 (1%) 5 (2%) 30 (13%) 59 (25%) 11 (5%) 130 (55%)
Alcohol 7 (3%) 26 (10%) 71 (26%) 85 (31%) 73 (27%) 9 (3%)

17
D R U G A B U S E I N PA K I S TA N

Table 7. Perceptions of long-term trends in drug abuse (last 5 years):

overall ratings

Drug type All Urban Rural Punjab Sindh Balochistan NWFP

Hashish and charas 9.4 9.0 10.1 8.0 7.3 14.8 11.4
Heroin 5.3 4.9 6.1 0.0 10.2 18.3 1.7
Opium 0.3 (–0.3) 1.3 0.0 (–1.1) 9.2 (–1.3)
Other opiates 4.8 5.6 2.2 8.6 1.8 1.1 1.1
Cough syrups
(for intoxication) 5.1 4.6 6.7 10.2 (–3.0) 5.0 1.4
Psychotropics 10.0 10.0 10.2 10.9 10.4 7.9 9.2
Drug injection 8.4 9.0 6.7 8.8 9.7 12.2 1.4
Solvents (glue, etc) 6.7 6.8 6.7 8.0 5.0 5.3 3.3
Alcohol 7.3 7.8 6.4 8.2 9.7 8.8 5.4

Note: Positive values indicate reports of overall increases in abuse levels and negative values decreases in abuse.

Figure II. Cumulative long term trend Balochistan. This analysis would fit with the tenta-
index tive suggestion that heroin use is still diffusing in
Pakistan to new areas, including more rural areas,
20
whilst at the same time stabilizing, or even possibly
18
slightly declining, in some of the urban centres
16
where use has been longest established. This pic-
14
ture remains preliminary and requires further inves-
12 tigation before any firm conclusion should be drawn.
10 However, the “last-year” overall trend measure indi-
8 cates a small decline in heroin use in the province
6 of Punjab (–0.9) (see table 9).
4

2 If a mixed picture is found for heroin use this is not

0
the case for drug injection where data give cause
for considerable concern. Increased injecting is
Other opiates

Solvents

Alcohol
Hashish

Heroin

Drug injection

reported in all provinces although the increase in

NWFP is marginal. However Sindh, Balochistan, and
Punjab all report strong upward trends. Injecting is
Punjab Sindh Balochistan NWFP reported to be increasing more strongly in urban
rather than rural locales, but for both the trend is
Note: high scores indicate increased reports that drug abuse
has increased in the last five years. upwards.

Key informants’ perception of long term trends in

This compares to ratings of 10 and 18 for Sindh and hashish and charas abuse appear uniformly up with
Balochistan, respectively, and 1.7 in NWFP. This some the highest overall ratings in Balochistan and
point is illustrated in figure II. Overall, the data sug- NWFP. Opium abuse on the other hand was report-
gest that in Punjab heroin prevalence may have ed to be stable or in decline in all provinces with
been stable or even in possible decline over the last the exception of Balochistan. Trends in alcohol use
five years. Elsewhere, it would indicate no great were reported to be up, as was the abuse of solvents
change in NWFP, where reported heroin use is low (glue, etc), with the locales in the province of Punjab
in any respect, and indicate increases in Sindh and reporting the greatest overall increase.

18
 M A P P I N G T H E N AT I O N A L C O N T O U R S O F D R U G A B U S E

Table 8. Perceptions of short-term trends in drug use in the locales (last year)

Decreased Decreased Increased Increased Don’t

Drug type a lot a little No change a little a lot know

Hashish and charas 9 (3%) 24 (9%) 58 (21%) 90 (33%) 91 (33%) 4 (1%)

Heroin 22 (8%) 62 (23%) 50 (18%) 73 (27%) 53 (19%) 12 (4%)
Opium 14 (5%) 33 (13%) 143 (55%) 38 (15%) 8 (3%) 26 (10%)
Other opiates 3 (1%) 9 (4%) 51 (21%) 30 (12%) 20 (8%) 130 (54%)
Cough syrups 5 (2%) 11 (5%) 40 (17%) 32 (13%) 23 (10%) 128 (54%)
Psychotropics 2 (1%) 6 (2%) 42 (16%) 99 (38%) 61 (24%) 49 (19%)
Drug injection 1 (>1%) 10 (4%) 38 (15%) 75 (30%) 27 (11%) 97 (39%)
Solvents 2 (>1%) 3 (1%) 39 (16%) 48 (20%) 17 (7%) 133 (55%)
Alcohol 4 (>1%) 23 (8%) 95 (35%) 78 (29%) 64 (24%) 7 (3%)

Table 9. Perceptions of short-term trends in drug abuse (last year): overall ratings

Drug type All Urban Rural Punjab Sindh Balochistan NWFP

Hashish and charas 8.5 8.2 9.1 6.7 7.8 8.3 11.1
Heroin 2.8 2.5 3.5 (–0.9) 5.6 13.1 0.8
Opium (–0.3) (–0.6) 0.4 0.3 (–0.4) 4.8 (–2.5)
Other opiates 4.9 6.0 0.8 8.5 2.9 0.0 0.8
Cough syrups
(for intoxication) 5.1 5.2 4.8 10.2 (–2.9) 0.0 1.9
Psychotropics 10.0 10.2 9.7 11.7 12.2 3.3 8.4
Drug injection 7.7 8.3 6.1 9.0 8.5 8.6 0.0
Solvents (glue, etc) 6.9 6.8 7.0 9.4 4.3 1.2 5.0
Alcohol 6.6 7.6 4.8 7.8 10.0 4.6 4.5

Note: Positive values indicate reports of overall increases in abuse levels and negative values decreases in abuse.

High overall scores were also found for upward Data from the whole sample can be found in table 8,
trends in the use of psychotropic drugs. Given the urban versus rural comparisons in annex table A.II.5
earlier findings that these drugs are reported as the and provincial breakdowns in annex table A.II.6.
most commonly used drug type by women this may Once more, an overall rating has been calculated to
imply that drug abuse among women in Pakistan is summarize these data and this can be found in
increasing. Data in this report do not explore the annex table A.II.9.
use of psychoactive substance by women in any
great detail. However, data do suggest that this is a The overall picture of perceived short-term trends
topic that merits further attention. broadly corresponds to the data for trends over the
longer time period. As noted above, trends in
Perceptions of trends in drug abuse over the last heroin abuse, with the exception of Balochistan,
year were also explored. This is a topic included in appear flat or even down slightly (Punjab). Drug
the UNDCP annual reports questionnaire (ARQ). injection is regarded as increasing, as is hashish and
Drug information systems often report on yearly charas abuse and again psychotropic drugs receive
trends so these data may have utility for any future a particularly high score on the overall rating for a
on-going drug abuse monitoring activity in Pakistan. rising short-term trend.

19
D R U G A B U S E I N PA K I S TA N

Perceptions of problems arising table 10 for the whole sample, and in table 11, for
the sub-sample comparisons.
from drug abuse
Different patterns and levels of drug abuse are The perceived problems different drugs types were
known to impact differently on society. To gauge the causing were not simply related to their reported
key informant’s perception of the extent to which the prevalence levels. In all provinces, and in both urban
drug abuse patterns described impact on their com- and rural locales, heroin was the drug most associ-
munities, they were asked to assess the amount of ated with problems, regardless of the relative preva-
problems each index drug was causing. A standard lence rate. Eighty-one per cent of respondents
question format was used in which each drug type reported that heroin caused major problems in their
was read out and respondents reported if it caused locales. This figure was slightly higher for urban
“major problems” “some problems” “few problems” rather than rural areas, but in both types of locale
or “no problems” in their locale. substantially high levels of problems resulting from
heroin abuse were reported. In Punjab province, 88%
Data for the whole sample can be found in table 10, of respondents reported major problems in their
a rural versus urban breakdown in annex table A.II.7, communities resulting from heroin. Those locales
and provincial comparisons in table A.II.8. Again, an with injecting populations also reported that these
overall rating has been calculated and is included in were causing major problems locally.

Table 10. Perceptions of problems caused by drugs in the locales

Major Some Few No Don’t Overall

Drug type problems problems problems problems know rating

Hashish and charas 46 (17%) 130 (47%) 73 (26%) 28 (10%) — 17

Heroin 222 (81%) 29 (11%) 6 (2%) 4 (1%) 12 (4%) 27
Opium 22 (8%) 56 (21%) 88 (33%) 82 (31%) 18 (7%) 10
Other opiates 14 (6%) 31 (12%) 45 (18%) 40 (16%) 120 (48%) 6
Cough syrups 22 (9%) 24 (10%) 26 (11%) 40 (16%) 134 (54%) 6
Psychotropics 20 (8%) 40 (15%) 85 (33%) 65 (25%) 50 (19%) 9
Drug injection 58 (23%) 55 (22%) 28 (11%) 25 (10%) 88 (35%) 12
Solvents 20 (8%) 30 (13%) 45 (19%) 27 (11%) 118 (49%) 7
Alcohol 85 (31%) 90 (33%) 71 (26%) 18 (7%) 6 (2%) 19

Table 11. Perceptions of problems caused, overall-rating scores: urban/rural

and provincial comparisons

Drug type Urban Rural Punjab Sindh Balochistan NWFP

Hashish and charas 17 16 19 14 16 18

Heroin 28 24 29 27 30 24
Opium 10 10 11 5 16 11
Other opiates 7 3 11 4 10 3
Cough syrups 7 4 13 2 2 3
Psychotropics 10 6 12 8 5 8
Drug injection 14 9 19 10 20 3
Solvents 8 5 16 3 5 1
Alcohol 19 19 24 18 10 18
Overall average 12 9.6 15.4 9.1 11.4 8.9

20
 M A P P I N G T H E N AT I O N A L C O N T O U R S O F D R U G A B U S E

After heroin, alcohol received the second highest Some provincial differences merit note. In Punjab,
overall problem score in all provinces with the as many as 24% of respondents reported major local
exception of Balochistan, where hashish and charas problems arising from solvent abuse, whereas this
problems received a higher rating. In Punjab nearly was not the case for the other provinces. The use of
half (48%) of all key informants reported that cough syrups for intoxication was also perceived as
alcohol caused major problems in their community resulting in local problems in Punjab. In
and none reported that there were no local Balochistan, opium use was seen as resulting in
problems due to alcohol abuse. In Balochistan major problems in 33% of locales, whereas else-
alcohol problems were not widely recognized—with where opium was not widely regarded as causing
only one key informant reporting major problems major problems.
resulting from the use of alcohol and nearly half
(47%) reporting only few alcohol related problems in
their area.
Table 12. Difference in respondents’
Whilst hashish/charas was the drug reported to be perceptions of differences in
most commonly used in all locales it was not nec- the drugs causing the most
essarily perceived to be causing major problems by problems for men, women and
respondents. Overall ratings suggest that its use was young people in the locale
perceived as causing fewer problems than either
Young
heroin or alcohol. That being said, hashish and cha- Drug type Men Women people
ras consumption was not regarded as non-problem-
atic. Seventeen per cent of all key informants repor- Hashish and
charas 27 (10%) 5 (5%) 113 (41%)
ted that major problems in their locale were caused
by hashish or charas, and just under half (47%) that Heroin 206 (75%) 36 (37%) 112 (41%)
some problems were due to the use of this drug Opium 5 (2%) 14 (14%) 1 (>1%)
type. Other opiates — 1 (1%) —
Psychotropics — 27 (28%) 1 (>1%)
Drug injection 6 (2%) 2 (2%) 4 (2%)
Alcohol 30 (11%) 12 (12%) 41 (15%)
Figure III. Cumulative problems arising
from drug abuse index, Note: Sample numbers: 274, 97, 273 respectively.

reported by key informants

in each state

35

30
Attitude questions
25

To explore some of the issues that emerged during

20
the preparatory work for the study a set of attitude
15
questions was prepared to which key informants
10 were asked to give a standard attitudinal response
5
ranging across a Likert scale from “strongly disagree”
to “strongly agree”. These data can be found in annex
0
table A.II.9. Key informants tended to agree with
Other opiates

Psychotropics
Cough syrups

Solvents

Alcohol
Opium
Hashish

Heroin

Drug injection

questions that suggested that many drug abusers

were too poor to access treatment provision and
were also responsible for considerable amounts of
Punjab Sindh Balochistan NWFP
local crimes. Respondents were more equivocal in
Note: Higher score indicate more reports of problems. reply to a question on whether heroin use had

21
D R U G A B U S E I N PA K I S TA N

declined in their area. Interestingly, those in Punjab suggestion that there has been a diffusion of heroin
and in urban areas were more likely to agree with to new areas but that abuse rates have not risen in
this statement than those elsewhere, supporting the areas with an established problem.

22
4

Interviews with drug abusers

In this section data from the interviews with drug abusers are presented. As previously noted,
in each of the four cities included in the study (Karachi, Lahore, Quetta, and Peshawar), three
samples of interviews with drug abusers were collected. These were: drug-abusers in treatment
(persons receiving treatment for an alcohol problem only were not included in the sample),
drug-abusers in prison and drug-abusers recruited in street settings. For the purposes of bre
vity and convenience we have used the term “addict interviews” to refer to these samples. The
reader should note that the sampling inclusion criterion used was that the respondents should
be regular users of heroin or drug injectors, and in practice the sample did consist largely of
apparently dependent opioid users. A formal measure of dependence itself was not used as a
sampling criterion, however, although such a measure was made in the interview.

Data collection
The analysis presented here will consider the three groups of “addict interviews” in their total-
ity and also make comparisons across the three groups. All interviews were conducted by
trained members of the research team, and were carried out on a voluntary and confidential
basis. As recruiting street samples of drug abusers is methodologically challenging members
of the research team with previous experience of this kind of work or who had life experiences
that allowed them to more easily access drug abusers were used to assist with this part of
the study.

One thousand and fifty-five interviews were conducted for this study. After data cleaning and
checking procedures, a small number of interviews (6) where excluded for reasons of the qual-
ity of the reporting. Thus 1,049 interviews were available for analysis and of these 12 were
female and 1,037 male. This sample cannot be thought of as reflecting patterns of illicit drug
use among Pakistani women. Whilst the abuse of heroin appears lower among women than
for men, the relative levels cannot be estimated from data reported here. For a number of cul-
tural reasons and other considerations, women appear disproportionately unlikely to appear
in any of the samples collected for this study and the topic of drug abuse among women is
taken up in the concluding part of this report. For the purposes of clarity and focus in report-
ing on the study, the following analyses are restricted to the sample of 1,037 male respon-
dents. Slightly more of these interviews were conducted with treatment attendees (415), than
with the street addict sample (321), or the prison addict sample (301).

Demographics—age distribution
The mean age of the three sample groups were remarkably similar, 33 years for respondents
interviewed in both treatment and street samples, and 31 years for those respondents

23
D R U G A B U S E I N PA K I S TA N

interviewed in prison. The age of the sample ranged reported some prison attendance as a result of a
from 12 to 74 years of age, with 40% of subjects drug-related offence. Overall about a third of respon-
being between 25 and 34 years of age. Comparing dents (35%) in the treatment and street samples had
the three samples, young respondents were least spent time in prison for a drug related offence. Those
common among the treatment sample, where only in the street sample more commonly reported drug
2% were under 20 years of age, as compared to 9% related prison attendance than in the treatment
of street addict interviews and 8% of the prison sam- sample (44% as compared to 29%, respectively). This
ple; they comprise 6% of the whole combined sam- again suggests, as reflected in their treatment his-
ple. Similarly with older addicts, only 6% of respon- tory and dependence, that the street addicts were a
dents were over 50 years, comprised of 5% of the particularly disadvantaged group.
treatment and prison samples and 7% of the street
sample. Some caution is merited in making direct Prison history
comparisons with the 1993 survey because of the dif-
ferent sampling strategy used in each study and The mean age of 26 for first prison attendance—
because full data on age characteristics were not amongst those who had been to prison—did not
available for the earlier exercise. Nonetheless, the vary across the sample groups. For those with a
current study does suggest a more mature popula- prison history, 45% had only one experience of incar-
tion of heroin users than might have been previously ceration. However histories of multiple prison atten-
identified, with less evidence of young heroin users. dance were not uncommon and around 10% of sub-
For example, in the 1993 exercise, 24% of heroin jects who had prison experience reported five or
users were found to be (from 15 to) under 20 years more incarcerations. Among those in the prison
old. In the 2000 study only 5% were within this age sample just under half (45%) reported they were in
group. Similarly, 73% of heroin users were found to prison for an offence related to the possession of
be less than 30 years of age in the 1993 exercise, drugs and a quarter (26%) reported they were impris-
whereas in this study the equivalent figure is 41%. oned because of an offence to do with the selling of
An age breakdown is displayed in figure IV. drugs. A further 18% reported that they were in cus-
tody because of an offence associated with being
Figure IV. Age distribution of sample intoxicated with drugs. Virtually all offences (93%)
(percentage in 5 year groups) were reported to be associated with the respondents
involvement with the use of heroin.
25

A similar picture emerges when data on living situ-

20
ation is examined. Perhaps not surprisingly, the
15 street recruited sample is most likely to report hav-
ing no fixed place of abode and living “on the street”
10
(41%) than the treatment (6%) or the prison groups
5 (13%). This figure was 18% for the entire sample, an
increase on the 8% reported in the 1993 research
0
10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50+
exercise. Clinic and prison attendees are by the
same token also more likely to report living with
All Treatment Street Prison
their parents (63% and 54% respectively) than the
street recruited sample (28%). Overall, for the sam-
ple as a whole, half of all respondents report living
with their parents. A minority of respondents (16%),
Prison contact were also living with their wife and this percentage
did not vary greatly across the individual samples,
All respondents in the study were asked about their although once again street addicts reported this less
experience of prison. Seven per cent of the treatment often than members of treatment or the prison sam-
recruited sample and 18% of the street sample ples (12% for street addicts, compared to 16% and

24
 INTERVIEWS WITH DRUG ABUSERS

19% respectively for treatment and prison addicts). more likely to report involvement in “business” and
Street addicts were far more likely to report living less likely to report “no usual employment” or
alone (22%) that those in treatment (3%) or prison “unskilled labouring”. When asked about actual
(9%). employment over the proceeding year (or, where
relevant, the year before entering treatment or
prison), nearly half the respondents (43%) reported
Demographics—education they had been unemployed in this period. Street
and employment addicts were more likely to report this than mem-
bers of the treatment or prison samples (56%, com-
On average, respondents reported having six years pared to 39% and 36% for treatment and prison sam-
of education. However, this varied by sample group ples respectively). Of the whole sample, 26%
with the treatment attendees reporting considerably reported full time employment in this period.
more time on average in education (7.7 years) as Treatment attendees reported this most often at
compared to the street group (4.8 years) and the 38%, compared with 27% of prison attendees and
prison sample (4.5 years). This finding is reflected in 11% of the street sample.
an analysis of those respondents that report no for-
mal education, which represents only 14% of the In the three months before interview (or, where rel-
treatment sample, but 28% of the street sample and evant, entering treatment or prison), 66% of all
40% of the prison sample; the figure for no formal respondents had been unemployed for some part of
education in the sample as a whole is 26%. Overall, this period. Of these who were unemployed at some
the treatment sample appears considerably better point in the three months, the majority (70%) had
educated than the other two groups, with nearly half been unemployed for the whole period. Again the
(46%) reporting 10 years in education compared with street sample most commonly reported unemploy-
similar status reported by 17% of street addicts and ment in this period (83%) compared to 55% of treat-
21% of the prison sample. The prison sample ment attendees and 66% of the prison sample.
appears the least well educated but contains a sig-
nificant minority, about a quarter of respondents,
with a high educational level (in terms of years). The Treatment contact
street sample, although closer to the prison sample
in terms of overall exposure to education, is less As an important part of this study was to look at the
polarized. overlap between the different populations of drug
users studied, all respondents were asked about
In table 13 data on “usual type” of employment are their experience of treatment (and prison) atten-
presented. The treatment group appears overall dance. Previous contact with drug treatment facili-
more affluent that the other two samples and is ties was relatively high among the prison and street

Table 13. Usual type of employment (percentage by sample groups)

Category Treatment Street Prison All

Professional 11 10 1 8
Clerical 6 2 2 4
Business 25 9 12 16
Agriculture 9 3 3 5
Skilled labourer 24 29 34 28
Unskilled labourer 8 23 32 19
Student 2 1 2 3
None 8 18 7 11
Other 7 5 7 6

25
D R U G A B U S E I N PA K I S TA N

Table 14. Means of financial support in the 12 weeks before interview

(percentage)

Category Treatment Street Prison All

Wages/salary 16 7 11 12
Casual work 14 27 27 22
Family 11 23 23 18
Begging 5 36 11 16
Selling drugs 6 14 20 13
Self employment 21 5 9 12
Pick pocketing/theft 7 10 16 11
Other 10 8 5 8

Note: Multi-response question—reflects period before entering treatment or prison for these groups.

samples with around 18% of both reporting contact 21 days (range 1 to 210); but again there was con-
in the 12 month prior to interview (or prior to enter- siderable variation, with two-thirds (66%) of subjects
ing prison in the case of the prison sample). It at the time of interview reporting having spent
should be noted that the cities for this study where 10 days or less in this current episode of treatment.
the interviewing took place have disproportionately
more treatment facilities than other areas of Pakistan. A supplementary multi-response question asked all
As such, treatment contact figures are likely to be respondents for which drug(s) they had ever been
lower elsewhere. Lifetime contact with treatment treated (for a problem with that drug). Of those who
services (of any sort) was even greater. Forty-four had been treated, the vast majority (96%) reported
per-cent of respondents interviewed in a prison or that they had been treated for a heroin problem,
street setting reported receiving treatment for a drug with far lower numbers reporting treatment for the
problem at some stage in their life, being just over abuse of opium (7%), hashish (5%), synthetic opi-
half (52%) of those interviewed on the streets and ates (5%), morphine (3%), cough syrups (3%), and
just over a third (37%) of prison respondents report- tranquillizers (3%). No subject ever reported having
ing lifetime contact with treatment services. been treated for a problem related to the use of
amphetamine or barbiturate problems.
The fact that many respondents had previously had
some contact with treatment facilities does not nec- When respondents in the treatment sample were
essarily imply that they had received intensive or asked for which drug they were currently receiving
sustained therapy. Contact with treatment facilities treatment, again the data are consistent with the
was often relatively brief. Respondents were asked findings from the sample as a whole. On a multi-
about their most recent treatment contact (exclud- response question, 94% of current attendees were
ing the interview index treatment for the treatment receiving help for problems related to their use of
sample). On average the contact lasted for 28 days, heroin, 6% for synthetic opiates, 4% hashish, 4% opium
although there is considerable variation in range of and 1% for morphine, tranquillizers and alcohol. In
the period of contact reported. A few individuals the treatment sample, 20% of subjects had injected
reported extended periods of therapeutic contact a drug in the month before entering treatment.
(maximum nearly one year), but for nearly half (49%)
of the sample, the most recent treatment episode
lasted 10 days or less. This finding was broadly con- First treatment
sistent with the length of time subjects in the treat-
ment sample had been attending the current treat- The data are remarkably constant across the sample
ment episode at the time of interview. The mean in respect of the mean age of first treatment
length of time subjects had been attending was attendance, which is 26 years of age for both the

26
 INTERVIEWS WITH DRUG ABUSERS

Table 15. Sample demographics and first treatment contact

Treatment Street Prison All

Mean age 33 33 31 32
Treatment contact
Percentage last year — 18 19 19
Percentage ever — 52 37 44

Mean age first treatment 26 25 26 26

First treatment
Percentage under 20 years old 18 21 15 18
Percentage 20-29 years old 56 55 55 55
Percentage 30-39 years old 21 20 23 21
Percentage 40 years old and above 6 4 7 6

treatment and prison sample and 25 years of age for samples. In the treatment group, 48% of respondents
the street sample. Summary data on sample demo- reported an experience of not being able to access
graphics and age of first treatment episode can be treatment as compared to 85% and 73%, respect
found in table 15. ively, in the street and prison groups. The reasons
for failing to access treatment were explored in a
The mean age from first use of heroin to first treat- follow-up question. The overwhelming majority
ment admission was calculated at 4.6 years. This is response (80% of all respondents, 70% of treatment
typical of many studies of heroin users conducted group, 84% of the street sample and 85% of the
internationally, that generally find a lag of between prison sample) was that they could not afford the
three and six years after initiation of use before seek- financial cost of entering treatment. Other reasons
ing assistance. However, there is considerable reported were a lack of available places (23%), a lack
individual variation on this measure with some indi- of inpatient facilities in government hospitals, a dis-
viduals rapidly developing problems that had led like of the treatment regime on offer (7%), a dislike
them to seek help and support. Others wait for con- of the treatment agency (7%), and a dislike of treat-
siderable periods of time before they decide, or are ment staff (5%).
able, to seek help.
Street and prison recruited respondents were also
asked to rate their own current need for treatment.
Need for treatment Eighty-one per cent of those interviewed on the
streets and 72% of those interviewed in prison
Respondents were asked whether it was difficult to reported that they currently needed treatment for
get into treatment in their area. Most (64%) replied their drug problem. The finding that drug abusers
that this was the case, with 53% of treatment atten- who are not current in contact with treatment serv-
dees, 62% of the street sample, and 80% of the ices recognize their need for them is not a univer-
prison sample reporting it was difficult for drug sal one. Studies that interview drug users in com-
abusers to enter treatment in their area. munity settings often find that there is a reluctance
among drug using respondents to recognize them-
All respondents were asked whether they had ever- selves that they have a need for treatment, even
wanted help for a drug problem but that they had where considerable evidence exists to suggest that
been unable to receive it. Overall, 66% of those inter- this is the case. Considerable efforts have been
viewed reported that this had been the case. Perhaps invested in many countries in attracting drug abusers
not surprisingly, those currently in treatment were and particularly heroin abusers into services.
significantly less likely to report a previous failure to It is therefore encouraging that in this study most
access help than those in the street or prison subjects recognized that they had a need for help.

27
D R U G A B U S E I N PA K I S TA N

This suggests that whilst considerable difficulties settings. Again heroin was most commonly used,
may exist in respect to the provision of care for this with virtually all respondents reporting using the
group, should appropriate services be available drug in the last 30 days (96%). This was followed by
many would seek to access them. hashish/charas and then alcohol. Other significant
current drug use included opium and tranquillizers,
Drug abuse history which were currently being used by a quarter of all
respondents.
All respondents were asked about their history of
using different types of illicit drugs and alcohol. The data on use of drugs reveal a picture, which is
These data can be found in table 16. Again the dif- fairly commonly found internationally, of multiple
ferences observed between the three samples were drug consumption amongst hard-core drug abusers.
not great and drug consumption repertoires Whilst the main focus of this study was on the use
appeared remarkably constant across the three of heroin, and this was the principal sample inclu-
groups. sion criterion, it can be observed that many heroin
abusers are also regularly consuming a range of other
In terms of lifetime prevalence, heroin was the drug psychoactive substances. This fact can complicate
most commonly used by respondents, followed by the development of effective interventions from both
hashish/charas, alcohol, opium, tranquillizers and the supply and demand reduction perspective.
synthetic opiates. The use of cough syrups (for the
purposes of intoxication) inhalants, and morphine, Although many respondents were consuming multi-
was relatively low (9%, 5% and 3% respectively) and ple drugs, the patterns of frequency of consumption
no significant use of barbiturate or amphetamine varied for specific substances. A simple measure of
use was detected. specific drug consumption among individuals who
had used the drug in the last month is the mean
Current drug use number of days on which they were consuming in
the 30 days prior to interview (note: the terms “last
When recent drug use was considered the con- month/30 days” refer to the month before entering
sumption patterns closely reflected patterns of life- prison or treatment for the sample groups contact-
time use. The reader should note that this informa- ed in those settings). Dependent drug use is typi-
tion applies to the month before entering prison or cally characterized by daily or near daily consump-
treatment for the samples contacted in those tion patterns.

Table 16. History of drug use (all respondents)

Used in past
Ever used 12 months Mean number of days
Drug (percentage) Age first used (percentage) used in last 30 days

Hashish/charas 86 18 77 19
Heroin 97 22 96 27
Opium 38 23 25 18
Synthetic opiates 14 27 11 20
Morphine 3 26 2 23
Cough syrups 9 22 3 14
Barbiturates 1 26 1 11
Tranquillizers 30 24 25 15
Amphetamines 1 24 >1 17
Alcohol 52 19 32 8.5
Inhalants 5 23 3 20

28
 INTERVIEWS WITH DRUG ABUSERS

Figure V. Usual route of heroin drug on a daily basis. It appears that overall the
administration users of these drugs are comprised of both a group
of intensive users and a significant, but slightly
Inject smaller proportion, of occasional users.

Sniff Daily use of hashish was also common with 41% of

those using this drug reporting daily use in the
month prior to interview, a further 34% were using
more occasionally and reported 10 days of use or
Smoke
less in this period. Of those consuming opium most
(32%) were using the drug regularly on five days a
week or more. The consumption of alcohol was more
The mean number of days on which heroin was con- varied but nearly all respondents (76%) who drank
sumed (27) indicates that heroin use was being more were drinking on two or three days a week or less
intensively used by respondents than the other psy- with only 10% five or more days a week.
choactive substances listed. Most (77%) of those
In addition to intensity of use, the route of admin-
who were using heroin reported using the drug on
istration (the way the drug is taken by the consumer)
a daily basis in the month prior to interview. This
can be an important factor in influencing the amount
figure was highest for the street users where 86%
of morbidity and mortality that results from a given
reported daily use as compared to 75% of treatment
level of prevalence. Data on the usual route of
attendees and 69% of the prison sample. Just over
administration for drugs data is presented in
6% of subjects where using heroin on a once per
table 17. It should be noted that these data reflect
week basis or less. Overall these data suggest that
the most common rather than an exclusive practice.
a highly dependent, chronic heroin abusing popula-
For many drug types, including heroin, drug injec-
tion has been accessed by this study.
tion is the route of administration typically thought
Far fewer individuals were using synthetic opiates of as resulting in the greatest levels of morbidity and
but of these just over half (57%) were doing so on mortality. In particular, this method of use is asso-
an intensive basis (four times a week or more). ciated with the transition of HIV and other blood
Intensive users of synthetic opiates were more com- borne diseases and has also been associated with
mon amongst the sample of street addicts. Although the elevated risk of opioid overdose. The injection
morphine was only consumed by a very small num- of drugs is therefore considered again in detail later
ber of subjects, two-thirds of these were using the in this section.

Table 17. Route of administration (selected drugs only—percentage of whole sample)

Drug Swallow Snort/sniff Smoke Inject

Hashish/charas 4 96
Heroin <1 11 73 15
Opium 98 <1 1
Synthetic opiates 20 80
Morphine 12 88
Cough syrups 100
Barbiturates 100
Tranquillizers 92 8
Amphetamines 92 8
Alcohol 100
Inhalants 100

29
D R U G A B U S E I N PA K I S TA N

Respondents most commonly smoked heroin, with adequate psychometric properties. The scale con-
73% reporting that this was their usual mode of sists of 5 questions each scored on a continuum of
administration. Eleven per cent usually sniffed the 0-3 with a maximum score of 15. Higher, positive
drug and 15% injected it. Patterns of heroin admini- scores are associated with higher levels of depend-
stration can change dramatically over relatively short ence. Other research studies suggest that scores at
periods of time. One of a number of important around the 5 or 6 mark would be comparable with
factors here is the nature of the available illicit a population requiring drug treatment. The mean
heroin, both in terms of composition and purity. It score on this measure in this present study is 8.4
would therefore appear important that any ongoing across all subjects. This figure varies by sample
information system include monitoring of the usual group, with scores of 8.1 for treatment attendees, 7.9
route of administration of heroin abusers and heroin for the prison group and a significantly higher figure
preparations available in the illicit marketplace. of 9.4 for the street sample. Full data for each of the
Some drugs are typical used in one particular fash- sample groups can be found in figure A.II.I in annex
ion, so it not surprising that hashish and charas are II to this report and is presented for all subjects in
largely smoked, inhalants sniffed and that cough figure VI. If a cut off point of 5.0 is taken as a simple
syrup and alcohol are exclusively drunk by respon- indication of dependence, this measure would sug-
dents. More worrying is the high level of injection gest that 87% of all respondents had levels of drug
for those using synthetic opiate and morphine. dependence that made them suitable for inclusion
Whilst the use of synthetic opiate remains low at in a drug treatment programme. If a cut off point of
present, a concern exists that any increased use of 6.0 is taken then this figure falls to 78%. In either
this drug group is likely to be accompanied by an case this can be regarded as a high percentage
increase in injecting prevalence rates. requiring treatment. Whist this can be expected for
the treatment group, who almost by definition are
likely to be dependent users, the street and prison
recruited samples of drug abusers exhibited a
Dependence and problems broader continuum of dependence scores. In this
study street users scored statistically significantly
Respondents were asked to identify the drug that higher on this measure and the prison sample
had caused them the most problems in the year was statistically equivalent (F = 31.27 (df2,1032),
prior to interview (or, when relevant, in the 12 p<0.001).
months before entering prison or treatment). No dif-
ferences were observable between the groups on this These data support the results presented earlier that
question. Overwhelmingly (94%), respondents suggest that drug taking repertoires of the heroin
reported that heroin was the drug that had caused
them the most problems. Other drugs, such as;
hashish, opium, morphine, cough syrup and tran-
Figure VI. Severity of dependence
quillizers were mentioned by 1% or less, of respon-
(all respondents—per cent and
dents as their major problem drug, and slightly more
SDS score)
(3%) cited synthetic opiates. Again, whilst the num-
20
bers of individuals in this study using this type of
18
drug was small, the potential for synthetic opiates 16
to cause problems of a comparable nature to heroin, 14
among those who abuse them, is worthy of note. 12
Per cent

10
8
A standardized measure of self-reported drug 6
dependence, the Severity of Dependence Scale 4
(SDS), was administered to all respondents. This 2
0
measure has been widely used with heroin and other 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
drug abusers elsewhere, and is considered to have SDS

30
 INTERVIEWS WITH DRUG ABUSERS

abusing population are actually rather narrow. The was designed to access hard-core and regular heroin
group is characterized by individuals in their late 20’s abusers, the homogeneity of the sample is still
and early 30’s with chronic and dependent drug surprising. The street recruited sample, where
habits, and of whom nearly all are appropriate for arguably a wider range of behaviours could have
admission to formal drug treatment programmes. been expected, is marginally more chronic and
Whilst the sampling strategy employed for this study needy even than those currently in treatment.

31
5

Drug injecting

Drug injection among the sample

Traditionally drug-injection has not been a common route of administration for Pakistani
heroin abusers. In the 1993 exercise, 92.5% of heroin users were smoking the drug, usually by
“chasing the dragon”—where heroin is heated on aluminum foil and the fumes inhaled. It was
noted with concern at this time that a small number (1.8%) of heroin users reported inject-
ing, as were a small number of the users of synthetic opiates.

By comparison, in the present 2000 study, 15% of heroin users reported that injecting was their
usual mode of heroin administration. Moreover, when asked if they had ever injected, a sur-
prisingly high 31% of the total sample reported they had. Of those subjects who had ever
injected a drug, 88% (27% of the entire sample) had done so in the year prior to interview (or,
where relevant, in the year prior to entering treatment or prison). This mode of administra-
tion is associated with particularly deleterious consequences for the individual. It is therefore
one of the most worrying finding of this study that injecting drug use is now a common prac-
tice among Pakistani heroin abusers. Data on injecting rates are summarized in table 18.

Table 18. Summary of injecting practices

Percentage

Heroin users who usually injected as main route of administration 15

Whole sample who had ever injected 31
Treatment sample who had ever injected 30
Street sample who had ever injected 30
Prison sample who had ever injected 33
Karachi respondents who have ever injected 55
Lahore respondents who have ever injected 30
Quetta respondents who have ever injected 14
Peshawar respondents who have ever injected 12
Whole sample who had injected in the last year 27

Injecting use varied between the cities studied. The majority (65%) of those who had ever
injected a drug were in Karachi, where they comprised more than half (55%) of all drug addicts
interviewed in the city. In Lahore, 30% of all drug users interviewed had injected (24% of all
injectors). This figure fell to 14% in Quetta (12% of the total injectors), and 12% in Peshawar
(9% of all injectors). Previous recent research reports had identified drug injection in some
areas of Pakistan, notably Karachi, but data from this study suggest that this behaviour is now
widespread, be it at varying rates of prevalence. Injecting drug use is a behaviour that is known
to diffuse quickly under certain conditions.

33
D R U G A B U S E I N PA K I S TA N

It is therefore desirable both to explore those fac- Age, education and employment
tors that are associated with initiation into injecting
drug use in Pakistan and to monitor the level of The mean age of injectors in the sample was
injecting among the heroin using population as a 32 years, with a minimum of 14 and maximum age
whole. of 74 years. There were no significant differences
between the ages of those respondents who had
Figure VII. Percentage of sample who injected a drug and non-injectors, nor between the
inject (lifetime, last 12 months, average ages of the injectors in the four cities.
and usual route of Injectors had initiated heroin use at the same age
administration) as non-injectors.
35
In terms of education, the injectors reported more
30
years of education than non-injectors. The mean
25 years of education for injectors within the sample
20
were 6.5, as compared to 5.6 years for non-injectors.
Within the cities, the injectors in Lahore (7.42) and
15
Quetta (7.39) had higher mean years of education
10 than those in Peshawar (6.6) and Karachi (5.9).
5
With regard to usual type of employment, one
0 observes major differences among employment cat-
months
Ever
Usually

Last 12

egories of injectors within the cities. The percentage

of injectors within each employment category how-
ever differs slightly from categories of non-injectors
in all the cities. In Karachi the large majority of injec-
tors consisted of skilled (36%) and unskilled (29%)
Drug injecting amongst women workers. The other employment categories reported
were business (10%), professionals and students (4%
As noted earlier a very small number of women were each). In Lahore the major categories for usual type
interviewed for this study. For clarity of interpreta- of employment among injectors were skilled work-
tion these interviews were excluded from the analy- ers (24%) and those affiliated with business (22%).
sis presented earlier in this section. Of the total of The other important employment categories in
316 injectors interviewed two were female—these Lahore were unskilled workers (15%), and profes-
individuals were interviewed in Lahore. For the rea- sionals (9%). The major employment categories for
sons discussed earlier this data tells us little about injectors in Quetta were professionals and business
injecting practice among women in Pakistan. A pre- (15% each), skilled (12%) and unskilled workers (9%).
liminary conclusion may be that drug injection The injectors in Peshawar matched more closely the
among women may not be common but does exist. injectors in Lahore with regard to their employment
However, comments made here on this topic are categories, but differed significantly from the cate-
largely speculative as little hard data exist to inform gories of non-injectors in the city. The employment
the debate. The best conclusion that can be drawn categories of injectors in Peshawar were skilled
from this study is that this remains an important workers (24%), business (21%), professionals (10%)
topic for further investigation. However, women’s and unskilled workers (7%).
drug use is likely to be a particularly hidden behav-
iour in Pakistan, and any future studies will have to With regard to employment status in the preceding
be sensitive to the cultural factors that make dis- 12 months there was a similar pattern of employ-
cussing this topic with female respondents particu- ment observed among injecting drug users in
larly challenging. Nonetheless this remains an Karachi and Peshawar and among injectors in
important subject for future investigation. Lahore and Quetta. More than half of drug injectors

34
 DRUG INJECTING

in Karachi had been unemployed and another quar- or before entering treatment or prison. In Karachi
ter doing part time work, while in Peshawar the heroin was the reported to be most commonly
majority (60%) of injectors were unemployed, 27% injected (87%), followed by synthetic opiates (23%),
doing part time and 12% full time work. In Lahore other drugs (14%) and tranquilizers (11%). Only one
the picture was slightly different, around 39% of respondent in Karachi had injected barbiturates.
injecting drug users were unemployed, whereas 32%
were employed full time and 29% doing part time In contrast to other cities synthetic opiates were the
work. In Quetta around 36% of injectors were unem- most commonly reported injected drug (65%) in
ployed, while similar proportions of injectors were Lahore, followed by heroin (28%), morphine (22%),
engaged in part time and 28% were working full time. tranquilizers (17%), and other drugs (9%). The inject-
ing drug use pattern in Quetta seems similar to
Karachi, as heroin was the most commonly cited
drug (89%) followed by synthetic opiates (14%) and
Patterns of use tranquilizers being injected by just one respondent.
In Peshawar heroin and synthetic opiates were
Overall 88% of injectors had injected a drug in the reported equally as the most commonly injected
12 months prior to interview. This figure varied drugs (40% each), followed by tranquilizers (27%),
across the four cities. In Quetta all those who had and other drugs (20%).
ever injected had also injected in the last 12 months,
this applied to 90% of injectors in Karachi, 87% in
Lahore and over just over half (52%) of the injectors
in Peshawar. Injection risk behaviour
Of those who had injected in the last 12 months One of the major concerns about drug injecting is
prior to interview, many had done so on multiple the risk of the transmission of HIV and other viral
occasions. A four point scale was used on this meas- infections through the sharing of contaminated
ure which categorized injecting once or twice, inject- equipment. In many parts of the world injecting drug
ing 10 times or fewer, injecting 11 to 24 times, and use is a major cause of new HIV infections. Among
injecting 25 times or more. This variable was intend- populations of injecting drug users where the shar-
ed to distinguish experimental injectors from those ing of equipment is commonplace, HIV infection
for whom it was a regular occurrence. Over 85% of rates can rise to epidemic levels in relatively short
injectors reported injecting 25 times or more in the time periods. Understanding the risk behaviour
month prior to interview. However, on a city level associated with drug injecting, in combination with
major variations in this pattern were observable. An an assessment of the prevalence of this behaviour,
overwhelming majority (93%) of the injectors in is therefore a critical aspect of assessing the poten-
Karachi had injected at least 25 times in this per tial for increased HIV infections within a country. In
iod, compared to 87% in Lahore, 60% in Quetta and this study the results are worrying, since consider-
just over half (54%) of injectors in Peshawar. A pat- able high risk behaviour was detected and the risk
tern is therefore discernable in which injecting is of widespread infection among drug injectors must
less common in some cities, both in terms of the now be considered a major public health issue for
number of heroin users who have engaged in the Pakistan.
behaviour and also in terms of the intensity in which
they engage in the behaviour. On a typical injecting day an individual was inject-
ing four times. A wide variation in the number of
With regard to the kind of drugs injected in the past times injecting equipment was used before being
twelve months, again different patterns of use are replacing was reported. Only 31% of respondents
observable between the cities. Respondents were reported that they used a clean needle/syringe on
asked on a multi-response question to list all drugs each occasion; 28% of injectors reported never or
they had injected in the 12 months before interview rarely cleaning a syringe before use; a further 22%

35
D R U G A B U S E I N PA K I S TA N

reported occasional cleaning and nearly half regu- never passing on a needle or syringe (28%), and
larly cleaning their injecting equipment before use. slightly under half (42%) reporting doing so 20 times
Methods used for cleaning included the use of or more (the highest category on the scale used).
spirit (25%), bleach (3%), boiling water (22%), and These data can be found in table 19.
cold water (49%). The actual efficacy of the proce-
dures used is not known and is very probably
questionable. The use of cold water in particular is Karachi profile
not likely to be effective.
Patterns of injection risk behaviour differed between
Most injectors (68%) reported that they usually used the cities studied. In Karachi almost two thirds of
drugs with other injectors. In the 12 months prior to injectors injected drugs in the company of others,
interview (or, where appropriate, entering prison or and 80% did not use new needles every time they
treatment) over half (53%) of injectors reported shar- injected. Although 60% said they “very often” cleaned
ing a needle/syringe after someone else had used it. their needles before use, a significant minority of
This figure rose to 69% for those in the street sam- injectors (22%) had never cleaned equipment and
ple. For many, using others’ equipment was not an 17% had only occasionally cleaned their needles
isolated incident. Nearly a quarter reported (23%) before injecting. More than half (58%) of the injec-
that they had done so 20 times or more, and this tors in Karachi cleaned their needles with cold or
figure was again higher (40%) for injectors in the warm water whereas 26% cleaned them by boiling.
street sample. Admitting sharing equipment after With regard to needle and syringe sharing it is
somebody else has used it is often difficult for drug interesting to note that 45% said they never used
injectors. A subsequent question therefore asked others’ needles while almost 31% said they had
about the risk behaviour of other injectors by ask- used needles used by others for more than 20 times
ing how often they had passed on a needle or in the past 12 months. On the other hand, more than
syringe they had used to another injector. half (57%) said that others had used their needles
Worryingly, on this question sharing rates were or syringes more than 20 times and only 22% said
higher yet. Under a third of the sample reported that other injectors had never used their needles.

Table 19. Sharing of needles and syringes in the last year (percentage)

Using equipment after Treatment Street Prison

another injector sample sample sample All injectors

Never 54 31 53 47
Once or twice 20 9 4 11
3-5 times 3 7 10 7
6-10 times 2 7 15 8
11-20 times 3 6 4 4
20 times + 18 40 14 23
Passing on equipment
to another injector
Never 41 27 17 28
Once or twice 22 7 4 11
3-5 times 2 6 12 7
6-10 times 7 2 15 8
11-20 times 2 5 5 4
20 times + 26 53 47 42

Note: Percentages based on those reporting “ever injecting”—lifetime prevalence measure.

36
 DRUG INJECTING

Lahore profile Correspondingly, almost half of the injectors said

they never cleaned their needles, while the other half
In Lahore a little more than half (52%) of the injec- either cleaned them very often or occasionally. More
tors said they injected in groups, the remaining 48% than half boiled their syringes to clean them, where-
said they injected alone. Sixty-three per cent did not as one third cleaned them with cold or warm water.
use a new needle or syringe every time they inject- Some 17% cleaned their needles and syringes with
ed. With regard to cleaning their syringes 35% said spirit. Apparently the majority (64%) had never used
they had never cleaned them, 60% reportedly others’ needles and syringes, while 21% had used
cleaned them very often and/or occasionally. In con- others’ needles just once or twice. Similarly, 43% of
trast with the other cities, the injectors in Lahore the injectors said others had never used their nee-
seem to use a wide range of methods for cleaning dles and syringes while some 29% said this might
their needles and syringes from cold and warm water have happened three to five times in the past 12
(32%), to boiling (23%), bleach (29%) and spirit months.
(18%). It is worth noting that the injecting drug users
in Lahore are the only group who reportedly were Drug treatment history
using bleach to clean their equipment. In respect to
the sharing of needles and syringes more than half The majority of injectors (63%) had undergone treat-
of the injectors said they had never used syringes ment for drug problems. The mean number of
used by others. Thirteen per cent reported using episodes of treatment for drug problems, excluding
three to five times after others, and 15% had used cannabis and alcohol, was 5.6 for the injectors,
more than 20 times after others. Similarly, 45% said which is significantly higher than the mean times of
others had never used their syringes, while 22% said 3.6 for non injectors. Within cities, the injecting drug
others had used their syringes more than 20 times. users in Peshawar and Lahore had been treated
more times (mean 7.89 and 7.71 respectively) than
Quetta profile in Karachi (mean 4.96) and Quetta (mean 1.9).

In Quetta the highest percentage (84%) of all injec- The mean age for first treatment for drugs other than
tors used in groups of other users. However, 47% alcohol and cannabis for injecting drug users is
had never used others’ syringes, while 25% said they 25 years, whereas the minimum age was 11 and
had used once or twice and 14% had used more than maximum 48 years. The injectors’ mean age for first
20 times syringes used by others. On the other hand treatment is lower than that of non-injectors, for
38% of injectors said others had used their syringes whom the reported mean age is 26 years. Within
once or twice, 27% said more than 20 times and cities the mean age for first treatment of any drug
about 24% said other injectors had never used their for injectors in Karachi and Lahore was reported
syringes. Thirty-eight per cent of injectors in Quetta around 25 years, whereas it was around 24 for Quetta
cleaned their syringes very often, while 34% of injec- and 23 years for injectors in Peshawar. The mean
tors cleaned then occasionally. About 29% of the duration time of their last treatment was reported
injectors had either never or almost never cleaned as 30 days ago for injectors—around 28 days in
their syringes. With regard to cleaning, 96% cleaned Karachi, 34 and 33 days in Lahore and Quetta
their syringes with cold or warm water and a mere respectively and around 20 days in Peshawar.
4% cleaned them by boiling.
Arrest and imprisonment history
Peshawar profile
More than half of the injecting drug users inter-
The injectors in Peshawar seem to prefer injecting viewed had been imprisoned. The mean age for first
alone as indicated by 67% of the respondents. imprisonment was 27 years whereas the mean time
Similar percentage of injectors indicated that they period between starting injecting drug use and first
used a new needle every time they injected. imprisonment was over 3 years, with over 2 times as

37
D R U G A B U S E I N PA K I S TA N

the mean number of times of imprisonment—these injectors in Peshawar had been arrested in the pre-
do not differ significantly from other drug users nor ceding 12 months.
across cities, except for Peshawar where the mean
number of times of imprisonment is around 5 for Whereas the mean number of times, injecting and
injectors. Similarly, the mean overall time served in other drug users had been arrested for drug related
jail does not differ significantly between injectors offences in the preceding 12 months does not differ
and other drug users, which is around 2 years. in other cities, the injectors in Peshawar have been
arrested significantly more times in the past twelve
The drug related offences for which injecting drug months than other drug users in the city (over four
users were imprisoned vary by offences among the times versus around two times).
four cities. Except for Karachi, where only 22% of the
injectors had been imprisoned for possession, the Living and support status
majority of injectors in the other three cities (Lahore
85%, Quetta 78%, and Peshawar 74%) had been In Karachi, around one third of injecting drug users
imprisoned for possession of drugs along with other lived mostly on the street, whereas for Lahore and
offences There are no significant differences in pro- Peshawar this proportion was around a quarter of
portion of injectors and other drug users imprisoned injectors. This proportion does not differ signifi-
for possession of drugs in all the cities. cantly among other drug users in these cities.
However, in Quetta whereas the percentage of inject-
In Karachi, Lahore and Peshawar around 60% or ing drug users who lived on the street is lower than
more of the injectors had been imprisoned for sell- in the other cities, it was significantly higher than
ing drugs. The percentage of injecting drug users for non-injectors in the city—16% for injectors ver-
imprisoned for selling drugs in Quetta is 21% only. sus 5%.
The proportion of injectors (61%) imprisoned for
selling drugs in Peshawar is significantly higher from Around 40% of injectors in Lahore lived mostly with
other drug users (24%) imprisoned for the same their parents, 26% with friends, 13% with children
offence in the city. and spouse and 10% alone. In Quetta, more than
half (57%) of injectors lived mostly with their par-
A significantly higher proportion of injectors in ents, 17% with children and spouse, 14% with other
Lahore (15% injectors vs. 5% other drug user) and family members and around 5% lived alone and 4%
in Peshawar (30% injectors vs. 6% other drug users) with friends. A similar living pattern is observed
had been imprisoned for trafficking of drugs. In among non-injectors in these cities.
Karachi and Quetta this proportion was negligible.
Except for Karachi where only a quarter of injectors In Karachi and Peshawar, there are significant dif-
had ever been imprisoned for intoxification, in the ferences in respect to the living status of drug injec-
other cities nearly half of the injectors had been tors compared to non-injectors. While 42% of the
imprisoned for these offences as well. Over half of injectors in Karachi lived with their parents, a lower
the injectors in Karachi had been imprisoned for percentage (9%) than non-injectors (22%) lived with
other drug related offences whereas this proportion their children and spouse. Similarly, while a quarter
was 12% in Lahore, and 9% in Peshawar. of injectors lived mostly with their friends, only 14%
of non-injectors reported living with their friends in
In the preceding 12 months, one third of the injec- Karachi. Less than half (48%) of injectors in
tors in Lahore had been arrested for drug related Peshawar lived with their parents compared to 60%
offences, while in Karachi a significantly higher pro- of other drug users. Similar proportion of injectors
portion (45%) of injectors as opposed to other drug (21%) and other drug users (16%) lived with their
users (26%) had been arrested for drug related children and spouse in Peshawar.
offences. The same is true for Quetta where more
than 60% of injectors as opposed to 30% of other In the preceding three months, more than 80%
drug users had been arrested. More than half of of injectors in Karachi and Peshawar had been

38
 DRUG INJECTING

unemployed, compared to 62% in Lahore and 58% (22%), by begging (18%), wages and salary (13%),
in Quetta. Around 11 weeks was the mean period by selling drugs (12%), and by theft and pick
that both injecting and non-injecting drug users pocketing (8%).
remained unemployed.
The main source of support for the majority (40%)
During this period around one third of injectors in of unemployed injectors in Quetta, as in other cities
Karachi were supported by family and/or a partner, was either family or partner. The other means by
24% supported themselves by begging, 18% by casu- which the injecting drug users supported themselves
al work, 10% by wages or salary, 6% by self employ- were casual work (24%), wages and salary (21%), by
ment, and 14% by other means. These percentages friends (18%), self-employment (12%), and selling
do not differ significantly among injectors and other drugs (9%). A similar proportion of non-injectors in
drug users in Karachi. However, significantly higher Quetta supported themselves by these means.
proportions of injectors in Karachi supported them- However significantly higher proportion (15%) of
selves by selling drugs (27% of injectors vs. 9% of injectors supported themselves by theft and pick
other drug users), were supported by their friends pocketing than other non-injectors (6%) in Quetta.
(24% of injectors vs. 10% of other drug users) and/or
supported themselves by theft or pick pocketing Family or partner supported almost half of the
(23% of injectors vs. 7% of other drug users). injecting drug users in Peshawar. The other means
of support were by theft and pick pocketing (21%),
In Lahore, more than a third of injectors were friends (18%), begging (14%), casual work (14%), and
supported either by their families or partners. The self-employment (7%). Significantly higher propor-
other means by which the injectors supported them- tion of injectors (21% injectors vs. 8% of other drug
selves during this period of unemployment were by users) in Peshawar supported themselves by selling
casual work (27%), friends (25%), self-employment drugs.

39
6

Data on prisons activity in the four cities

Prison data collection

An audit of local prison and treatment activity data was conducted in the four selected cities.
This information was needed to inform the prevalence estimation exercise and was also use-
ful for exploring the impact that heroin abuse was having on the local prison services. Prison
activity data were collected for the main prison dealing with drug offences for each of the four
cities. This would audit a high proportion of total cases in each area and give some informa-
tion on the kinds of offences that drug abusers were incarcerated for. The audit focused on
collecting detailed information on prison occupancy by drug addicts and those involved in
lower levels of selling drugs, specifically, the number of people sent to prison in the last
year for drug use and drug-related offences for each city. This data is summarized in table 20
and 21.

Table 20. Data of four prisons (all inmates)

Drug related charges

Total prison (Both under trial
population Under trial Convicted and convicted)

Lahore 2 834
Peshawar 2 782 2 286 496 1 222
Karachi 3 500 Not available Not available 545
Quetta 840 Not available Not available 168
Total 4 769

Table 21. Prison data record (inmates who have drug related charges)

Multiple
3/4 P.O 6/9 CNSA charges 11 P.O 8 P.O 9c 9b 9/15 14/15 Total

Lahore 1 090 589 1 063 32 8 1 51 2 834

Peshawar 912 150 83 68 1 1 222
Karachi 307 238 545
Quetta 22 146 168
Total 2 331 1 123 1 146 100 8 1 51 1 4 769

Notes: P.O = Haddod Ordinance (Islamic Injunctions), CNSA = Control of Narcotics Substance Ordinance.
Multiple charges = drug charges + other charges, 9c, 9c,9/15, 14/15 are all Control of Narcotics Substance
Ordinance.

41
D R U G A B U S E I N PA K I S TA N

In Pakistani prisons large numbers of drug addicts purchase, distribute, delver on any terms whatsoever,
can be found and at any given time, the prison pop- transport, dispatch, any narcotic drug, psychotropic
ulation of drug abusers represents the largest num- substance or controlled substance, expect for medi-
ber of individuals to be found in any institution in cal, scientific or industrial purposes in the manner
the country. Many addicts spend prolonged periods and subject to such conditions as may be secified
in custody and drug rehabilitation and treatment by or under this ordinance or any other law for time
facilities are extremely limited within the prison sec- being in force.”
tor. Most of the drug addicts are incarcerated for
charges of drug use, and/or the possession of small The punishment for contravention of section 6 is:
quantities of drugs. Interestingly, it appeared that “Whoever contravenes the provision of section 6
many addicts were in prison through the involve- shall be punishable with:
ment of their families, who having been frustrated “(a) imprisonment which may extend to two years,
by their behaviour and their addiction, had arranged or fine, or with both, if the quantity of the narcotic
for them to be incarcerated. Drug addicts are usu- drug, psychotropic substance or controlled sub-
ally placed in a separate barracks and where care is stance is ten grams or less;
available, usually this consists only of short-term
“(b) imprisonment which may extend to seven years
symptomatic treatment for the relief of withdrawal
but shall be not less than three and shall be liable
distress. When medication is provided it is on a
to fine, if the quantitiy of the narcotic drug, psy-
reduced dosage basis lasting for approximately three
chotropic substance or controlled substance
to four weeks. Although it was not the subject of this
exceeds ten grams but does not exceed:
inquiry, it was suggested that, like many other coun-
tries, some limited abuse of drugs also occurred “i) one hundred grams, in the case of
within the jail system. heroin, cocaine, opium derivative or coca
derivative, with whatever name or com-
position;

Overview: drug charges in “ii) two hundred grams, in case of opium,

coca leaf or psychotropic substance, with
Pakistan whatever name or composition; and

Drug offenders are charged under two main laws in “iii) one kilogram or equivalent, in case of any
Pakistan. These are: (a) Islamic Law (Hadood Ordi- other narcotic drug or controlled sub-
nance) and (b) Control of Narcotics Substance Abuse stance not specified in sub-clauses (i)
(CNSA) Act of 1997. The most relevant sections of and (ii); and
the 1997 legislation is Section 6: Prohibition of “(c) death or imprisonment for life, and shall also
possession of narcotics drugs etc: (According to be liable to fine which shall not be less than one
chapter II of the Control of Narcotics Substance million rupees, if the quantity; of narcotic drug, psy-
Abuse (CNSA) Act of 1997). “No one shall produce, chotropic subsatnce or controlled substance exceeds
manufacture, extract, possess, offer for sale, sell, the limit specified in clause b.”

42
7
Drug treatment in Pakistan:
the national treatment register update

Updating the register of drug treatment in Pakistan

The provision of therapeutic interventions for individuals with drug problems is a core com-
ponent of a comprehensive demand reduction approach and the provision of drug treatment
is one of the central issues addressed in the Declaration on the Guiding Principles of Drug
Demand Reduction. A number of studies have shown that drug treatment, when delivered
appropriately, is a cost-effective component of a drug control strategy. One aspect of the 2000
assessment study was therefore to update an earlier register of agencies offering drug treat-
ment in Pakistan. Such a register is a valuable resource for conducting future training, net-
working and development activities to improve the quality and delivery of services for those
with drug problems in Pakistan. Furthermore, this information is important for considering the
coverage and availability of services and also for beginning a debate on their most effective
configuration. A wide range of activities and therapeutic regimes can fall under this general
heading. The need exists in most countries to ensure that a range of appropriate interventions
is available, that they reflect client needs, and that they are evaluated to demonstrate effec-
tiveness. UNDCP is currently working with the Government of Pakistan to help ensure that
services meet these objectives.

Data on treatment attendance, usually collected in an anonymous register, is also the central
component in many ongoing drug abuse surveillance systems. This source of information can
provide a valuable trend indicator. Treatment registers are also low cost, can provide useful
information to those delivering services and can also inform government planning.

A special questionnaire was developed for the purpose of updating the earlier treatment agency
list. As well as auditing what provisions existed, this form also collected information on cur-
rent activities, patient mix, average occupancy rates and other services provided. This infor-
mation is discussed below.

Overview and history of drug treatment in Pakistan

Drug treatment programmes in Pakistan formally began with the proclamation of Hadd (Religious
Injunctions) in 1979, which saw scores of opium addicts who could not get their daily dosage
of opium from the traditional opium vend system, coming to government hospitals for “help”.
Most of the hospitals and the staff available at that time were ill equipped and did not have
the necessary training to deal with such cases. Thus the onus of treating those with drug prob-
lems fell on the departments of Psychiatry within the Government hospitals. With the emer-
gence of heroin epidemic in the early 1980s, again the demand for treatment services grew. At
this time there was an increase in provision, with Government agencies, hospitals, NGOs and
private facilities all expanding in size. The focus of much of this activity was detoxification

43
D R U G A B U S E I N PA K I S TA N

procedures for drug withdrawal which was viewed as treatment at the supported centres. Every month the
a key first response to the increasing number of drug treatment centres being financed in the proj-
heroin addicts seeking help. ect sent in forms filled in providing information on
the social and demographic profile and drug use his-
With regard to institutional and capacity building for tory of every patient that came for treatment. These
delivery of drug treatment and rehabilitation ser- were then complied, analysed at intervals and final-
vices, one can see two distinct phases in Pakistan. ly printed as a PNCB publication with the same title.
These were, At the end of the project, the responsibility of run-
❏ Setting up of drug treatment facilities at ning the treatment facilities was handed over to the
selected locations with the support of Provincial Health Departments under whose admin-
UNFDAC (1982-1988) istrative control fell the government run facilities.
❏ UNDCP funded Integrated Drug Demand Over the years due to financial constraints and other
Reduction Project (IDDRP) efforts at im- factors, most of the treatment facilities had become
proving the service delivery of treatment dysfunctional and capacity appeared to decline, as
programmes (1991-1996) did the availability of specialized and trained staff.

UNFDAC funded treatment Integrated drug demand

facilities reduction project
The erstwhile UNFDAC provided the first financial The second distinct phase in improving the service
and technical support in the 1980s (1982-1988) in delivery of treatment programmes was of the Inte-
setting up 32 model drug treatment centres in the grated Drug Demand Reduction Project (IDDRP)
country, most of them situated in government teach- supported by UNDCP. In its over five years of imple-
ing hospitals (departments of psychiatry) and some mentation the project developed materials and pro-
within NGO run facilities. The financial support pro- vided training to service providers with the objective
vided to these facilities by UNFDAC included provi- of improving service delivery of treatment pro-
sion for part salaries of the staff involved in the grammes beyond the short-term medical interven-
treatment, cost of medicines, and other operational tions. The concepts introduced focused on social
costs related to running the treatment centres. The rehabilitation and integration of drug dependent
material support included provision of vehicles for persons and included assessment of drug related
bringing in clients for treatment or for their follow problems, counselling, relapse prevention, and after-
up, TLC (Thin Layer Chromatography) equipment for care. In all, through 23 workshops, over 600 health
drug testing, beds, etc. Technical support was pro- workers, psychologists, social workers, paramedics,
vided in training medical doctors not only from the and NGO workers were trained in the application of
supported treatment facilities, but also from other these approaches. The impact of this activity was
institutions as well as in treatment procedures, prima- most apparent in the non-governmental sector where
rily symptomatic treatment of withdrawal symptoms. some agencies were motivated and able to develop
modern and comprehensive treatment regimes.
The Pakistan Narcotics Control Board (PNCB) also However, such agencies are by no means common
published a booklet as a result of technical support in Pakistan and rely on extremely limited resources.
in the project titled “Guidelines for Heroin Impact on the Government run treatment sector has
Detoxification in Pakistan”. As the name suggests, largely not been apparent and this sector appears to
this publication provided broad guidelines for pri- have declined in importance as a primary provider of
mary health care physicians for detoxification of care for those with drug problems. Where services
people with heroin dependence. The project also exist they are usually limited in approach and it is
introduced a “National Case Monitoring System” unclear how successful they are in attracting and
which looked at the profile of patients coming in for retaining patients.

44
 D R U G T R E AT M E N T I N PA K I S TA N

Similarly, in the absence of inpatient rehabilitation, regimes used at the Rawalpindi General Hospital,
the IDDRP introduced the community based treat- Rawalpindi, Lady Reading Hospital, Peshawer, and
ment approach through supporting community Mayo Hospital, Lahore for their relative effective-
intervention teams (CIT). Each team consisting of ness in relieving withdrawal symptoms. All of three
two members and based within an NGO, were estab- regimes proved more or less equally effective in
lished, funded and monitored by the project. The helping a patient through heroin withdrawal.
teams were trained in social reintegration, rehabili-
tation and community development concepts,
including concepts of community involvement and
linkage of community resources for social reintegra-
National survey of drug
tion of their clients. Each team worked in a defined treatment and rehabilitation
geographical area making referrals for treatment, services in Pakistan
and applying the concepts they were trained in.
These teams over the years proved a good model of This second study commissioned by IDDRP in 1994
providing effective community based rehabilitation looked at the extent and type of services available
and social reintegration services to the drug depen- for drug treatment and rehabilitation all over the
dent persons. Again, with the termination of the country. According to this survey, there were 203
project, most of the NGOs could not sustain the organizations operating with varying capacities pro-
level of activities, the services provided, or the two viding drug treatment services in some 89 cities and
member team and therefore in most of the places towns across the country. Of these, 47% were private
these became dysfunctional. clinics, 42% government hospitals and 11% were
facilities run by NGOs. Eighty per cent of the organ-
Around 1995, using the concept of Community izations were providing inpatient care while 68% pro-
Intervention Teams, two pilot projects were also vided outpatient care only or in addition to the inpa-
launched to provide drug treatment and rehabilita- tient care. Concerning the duration of detoxification
tion services to prison inmates with drug problems and total treatment stay more than 80% of those
in Lahore and Rawalpindi jails. Again, these projects reporting provided one to two weeks for detoxifica-
could not be sustained after the end of support from tion, whereas around 60% of facilities reported one
the project, mainly due to lack of interest and own- to two weeks for total treatment stay. Only fewer
ership by the concerned authorities. Only one pro- than 2% reported a total stay of three to four weeks
gramme that was started around 1998 with support for their clients. More than 90% used symptomatic
from other donors continues to operate in the treatment of withdrawal symptoms as their preferred
Peshawar Jail. method of detoxification of clients. In 61% of the
clinics, the head of the facility had no formal train-
As an outcome, these efforts at improving the ser- ing in drug treatment and rehabilitation. The remain-
vice delivery of drug treatment programmes in ing 39% had some formal training. The survey teams
Pakistan produced not so significant results or had felt that detoxification was the process the
changes in delivery of services in Pakistan. respondents were most familiar with. Although the
organizations were providing other services as well,
some respondents appeared to lack a clear under-
Past assessments of drug standing of some of the services asked about in the
treatment programmes survey.

In the past two decades, prior to the current assess- This survey was a first ever effort in this area and as
ment, two main efforts were made to access the a result it was possible to develop a directory of
quality and type of treatment services offered in drug treatment services available in Pakistan that
Pakistan. The first one titled “Heroin Detoxification could be used by those interested in knowing the
in Pakistan” was undertaken by PNCB with location and types of services available in a parti-
UNFDAC’s support and compared three treatment cular area.

45
D R U G A B U S E I N PA K I S TA N

National treatment registry view, was 264. This ranged from between 7 and
3,000 clients being admitted during the relevant
update 2000 period and indicates that some services had con-
siderably higher client contacts than others. The
As part of the current research exercise, the National total number of admissions in the 73 centres was
Treatment Registry was updated. During this exer- calculated as 17,425 in the 12 months prior to inter-
cise the field workers visited 18 selected major urban view. This can be regarded as a proxy yearly total for
centres in the country and identified 73 centres that client admissions. It should be noted that this
were providing drug treatment and rehabilitation figure excludes alcohol users, and includes repeat
services on a regular basis. Out of the 73 centres, attendees. It should also be noted that this data
38 facilities had been enlisted in the 1994/95 survey does not allow one to comment on the level of
of drug treatment services, whereas 35 centres and double-counting between treatment facilities. If re-
facilities enlisted in the current exercise are either peat attendees (those with more than one episode
newly established facilities or that had not been of treatment at the agency in the last 12 months)
enlisted in the previous exercise. Twenty-eight of are excluded, the total estimate falls to 11,454 with
these 73 facilities (38%) are government funded and the mean of 176 different patients per agency admit-
administered hospitals, 25 (34%) are NGO run facili- ted in the previous 12 months.
ties and 20 (27%) are private (for profit) run drug
treatment centres.
Young people and women were less commonly seen
Forty-three (59%) of the 70 three facilities provide in treatment. Only 30% of clients less were reported
both in and outpatient services, whereas 27 (37%) as being under 25 years of age and only 3% of
provide only inpatient services and three (4%) pro- clients were women. How far this reflects a dis-
vide only outpatient services. The mean capacity, i.e., proportionate higher proportion of male drug use
number of clients treated at any given time, for in- and how far it reflects reluctance or difficulties
patient clients in the centres is around 37 whereas women have in accessing services is unclear. How-
for outpatients it is 59. The range for clients treated ever, the social stigma on drug use among women in
as inpatient was 1-1,000 (one facility only) and for Pakistan is considerable. It would therefore appear
outpatient 1-450 (one facility only) clients. highly likely that women with drug problems would
be reluctant to enter mixed-sex services. Assessing
All services reported providing detoxification servi- the prevalence and nature of drug problems among
ces. It should be noted that even within this narrow women in Pakistan, and developing culturally appro-
category of response considerable differences may priate treatment responses, remains an important
exist in terms of therapeutic practices. Ninety per challenge.
cent (66) of agencies reported providing counselling,
and over half of all agencies (50/68%) also offered As treatment services have ongoing contact with
religious counselling. Forty-one (56%) reported that drug abusers, they can be considered a useful source
they had an outreach service. This kind of provision of information on trends over time. Around 40% of
is regarded as an important component of an over- treatment centre staff interviewed thought that the
all treatment approach as it can provide a conduit proportion of well-off clients coming for treatment
to more formal treatment or deliver services to drug has lowered over the last five or six years although
abuses who are unable or unwilling to access other 20% considered that this proportion had increased.
treatment options. Just over half (59 of the 73-80%) Some staff (26%) were also of the opinion that “sniff-
included a relapse prevention element in their after ing drugs” (the use of inhalants) had increased espe-
care facilities. A range of other treatment options cially among adolescents, that the use of other opi-
were mentioned by 26 (35%) of facilities. ates and cough syrups had increased due to a
decline in quality of illicit heroin (25% of the respon-
The mean number of admissions (at each centre) dents) and that injecting drug use was increasing
for drug treatment, in the 12 months prior to inter- among treatment attendees.

46
 D R U G T R E AT M E N T I N PA K I S TA N

No major changes are evident from earlier studies practice standards, and the identification of training
on the pattern of service provision in respect to and other developmental needs.
Government and non-government facilities. NGO’s
appeared to be proving the broadest range of care All but one of the treatment facilities indicated that
and where Government treatment centres were pro- they kept records of age, gender, education, occu-
viding services this tended to be only on a detoxifi- pation, primary drug, injection status, marital status
cation basis. and treatment history of their clients for monitoring
purposes. However, most of these records are not
Some diversification and improvement is suggested summarized in a form that would allow easy analy-
in respect to the range of services offered by some sis and storage facilities often mean that data
private and NGO managed facilities. They have retrieval would be problematic. Currently some peri-
developed programmes that provide a range of treat- odic reviews do take place within agencies but peri-
ment and rehabilitation services, and have made odically, no actual analysis of trends over time is
innovations in their interventions to meet the chang- attempted. Nonetheless, the fact that agencies col-
ing requirements of their clients. A few organizations lect this information means that developing a treat-
have experimented with concepts of therapeutic ment reporting system is a realistic aspiration. As
communities as well as of community based reha- the information currently collected is poorly utilized
bilitation of clients. This area is not addressed in at present, such a system could supply the treat-
detail by this current study but the results do sug- ment services themselves with a useful review of
gest the need for a more in-depth assessment of the their own clients. This data could also provide a
kinds of treatment provision available for those with valuable national indicator of patterns and trends in
drug problems in Pakistan, the development of good illicit drug abuse in Pakistan.

Table 22. Comparison of drug treatment services available in 1994/1995 and 2000

1994/1995 2000

Total provision identified Total provision identified

Number of treatment centres identified was 203 in Number of treatment centres identified was 73 from
89 cities and towns 18 major urban centres of the country. Thirty-eight
of these facilities were also listed in the 1994/1995
survey whereas 35 facilities were newly established
or not identified in the previous exercise.

Centres providing: Centres providing:

Outpatient services only = 35 (17%) Outpatient services only = 3 (4%)
Inpatient services only = 61 (30%) Inpatient services only = 43 (59%)
Inpatient and outpatient services = 97 (48%) Inpatient and outpatient services = 27 (37%)

Organization: Organization:
Government hospitals = 67 (33%)* Government hospitals = 28 (38%),
Private (for profit), = 93 (46%) Private (for profit), = 20 (27%)
NGO = 43 (21%) NGO = 25 (34%)

Almost all provide detoxification services, nearly all All provide detoxification and over 90% counselling
claimed to be providing counselling services

Over 66% reported to be providing relapse preven- Relapse prevention and aftercare services provided
tion therapy services by over 80% of services

*Government hospitals including teaching hospital, district headquarter hospitals or Tehsil headquarter hospitals.

47
8
National drug contour mapping: estimating the
prevalence of hard-core heroin use in Pakistan

Data structures of the contour mapping exercise

The Pakistan National Survey consists of three component data collection studies, each of
which is designed to stand alone as a separate research study. These component studies have
been designed in a consistent way and carefully structured to allow a mapping of drug use
patterns across the whole country, extrapolating from the actual geographical locales that have
been studied in detail. This fourth exercise—the National Drug Contour Mapping—uses
key data from the other three components to produce national prevalence estimates of drug
abuse.

The component studies have been constructed to collect information in hierarchically increas-
ing detail, but with increasingly restricted geographical coverage. The studies are:
Key informant study component;
Treatment register update component;
Four cities study component.

The structure of the overall National Survey is shown below in table 23 and table 24. Table 23
gives the geographical locales selected, indicating the principal urban locale and the rural
locale selected to pair it. This structure is intended to give adequate national coverage, bal-
ancing rural and urban populations.

Method of estimation: treatment multiplier method

The construction of overall prevalence rates for the country as a whole of hard drug addicts
(heroin users and injectors) has never been an easy matter, as previous surveys attest. In this,
Pakistan reflects the difficulties found in other countries, where identifying the extent of an
illegal behaviour that is generally conducted out of contact with official information sources
is recognized as requiring special techniques of estimation. These specialist techniques use
“indirect estimation methods”, not attempting to count directly the number of addicts, but
instead putting together different sources of information that allow the prevalence of drug
abuse to be estimated indirectly.

In this exercise, the procedure selected is one of the principal indirect techniques used inter-
nationally, in the United States and in Europe, the Treatment Multiplier Method. It is a two-
step procedure that requires in essence two separate figures to be produced:
(a) information on the number of addicts that receive treatment during the year, and
(b) an estimate of the proportion of the entire addict population that this represents.

49
D R U G A B U S E I N PA K I S TA N

Table 23. Locales used in the key informant study component

Key informant study (urban) Key informant study (rural)

Study Urban locale No. of key Rural pairing No. of key

Province ID No. (districts) informant interviews locale informant interviews

Punjab 7 Lahore 22 Kanna 3

6 Multan 10 Mattial 5
10 Rawalpindi 12 Dahmial 3
8 Sialkot 10 Rangpura 6
9 Faislabad 11 Jaranwalla 4
Sindh 3 Karachi 20 Malir 5
4 Hyderabad 14 Tando Allah Yar 5
5 Sukkar 15 Abad Jageer 5
Balochistan 2 Quetta 13 Kachlaq 5
1 Turbat 6 Godaan 6
NWFP 12 Peshawar 10 Bad Bair Village 5
14 Bannu 5 Mandan 5
13 Maradan 5 Tahkat Bai 2
17 Chitral 5 Garrama Chasma 5
15 Dir 7 Rural Timergrah 5
16 Gilgit 3 Ghaok Gaza 5
18 Haripur 7 Rehana 6
(FATA)* 20 Mohmand Agency 5
21 Khyber Agency Ladi Kotal 5

Totals 175 90

*For prevalence estimation purposes FATA is incorporated with NWFP.

In any given survey, deriving the requisite informa- and this decision was made for two reasons. Firstly,
tion requires methods suited to and adapted to the it provided a clearer focus for determining the nature
social and geographical structures of the country at of the drug addiction for which the person was being
that time. Precise operational definitions implemen- treated, and so allowed a more accurate definition
ted in the survey need to be tailored to the avail- of the target group of addicts to be implemented,
able information and data collection possibilities. than would be the case had broader general health
The immediate steps in providing a prevalence esti- settings for treatment been included. Secondly, the
mate in the Pakistan National Survey are therefore enumeration of all addicts treated at specialist units
as follows. over the 12 months is more accurately and more
easily carried out than could be the case in more
broadly defined and more widespread health treat-
Estimating the number of addicts ment settings.
in treatment
Although therefore some addict treatment episodes
Data from the Treatment Register Update study were are not included in the figure, this shortfall can be
used to determine the number of addicts receiving corrected at the second stage of the calculation. In
treatment over the 12 months prior to the survey. determining the number of addicts so treated, the
This figure was restricted to those addicts undergo- distinction has been made between the number of
ing inpatient treatment in specialist drug clinics, treatment episodes and the underlying number of

50
 N AT I O N A L D R U G C O N T O U R M A P P I N G

Table 24. Locales in the treatment register study and in the four cities study

Locales in Treatment register

overall study update study Four cities

Study Urban locale Specialist treatment No. of treatment Addict interviews

Province ID No. (districts) centres centres conducted

Punjab 7 Lahore Lahore 14** 256*

6 Multan Multan 5
10 Rawalpindi Rawalpindi 6
8 Sialkot Sialkot 2
9 Faislabad Faislabad 1
Sindh 3 Karachi Karachi 17** 264*
4 Hyderabad Hyderabad 6
5 Sukkar Sukkar 2
Balochistan 2 Quetta Quetta 3 255*
1 Turbat
NWFP 12 Peshawar Peshawar 7 262*
14 Bannu Bannu 1
13 Maradan Maradan 1
17 Chitral Chitral 2**
15 Dir Dir 1
16 Gilgit
18 Haripur
(FATA)*** 20 Mohmand Agency
21 Khyber Agency
Totals 68 1 037

*A total of six additional interviews with female addicts were conducted.

**Three locales reported a treatment centre with no inpatient facilities.
***For prevalence estimation purposes FATA is incorporated with NWFP.

people who generate the episode total through in specialist drug clinics during the year. This infor-
repeated treatment episodes during the same year. mation was collected from a combination of data
sources, namely from the four cities study and from
The information from the Treatment Register Update the key informants study. The addicts interviewed in
study generated a profile of specialist drug treatment the four cities study—those who were not currently
clinics across Pakistan. The information on numbers in treatment—were asked whether they had been in
of inpatient beds and numbers of male hard drug treatment in the 12 months prior to their interview.
users treated during the year are given in summary The proportion of addicts who had been treated
form in table 27 below. could therefore be estimated. This was done in all
four cities where addict interviews were conducted.
Estimating the multiplier to represent
addicts not receiving treatment The key informant study provided more widely
spread information—by direct questioning of the
The appropriate multiplier for the calculation, as a respondent—that gave the proportion of addicts
consequence of the above decisions on inclusion, treated that year in the various locales used in the
represents all addicts who did not receive treatment study. Informants who were not directly involved

51
D R U G A B U S E I N PA K I S TA N

with treatment of addicts were asked how many extrapolation are required to produce an estimate.
addicts they had contact with in the past 12 months, The first level requires extrapolation of the preva-
and how many of these had received treatment in the lence estimates from the locales with identified spe-
past 12 months. In using this information to produce cialist treatment clinics to other locales in the
an estimate of the proportion of addicts treated in province that have none. The second level requires
the year, account was taken of the informants’ expe- an extrapolation from the locales selected for the
rience and suitability in providing a valid estimate. study to the province as a whole. The estimates for
Pakistan as a whole then are calculated as a combi-
These studies produce a range of likely estimates of nation of the four provincial figures.
the proportion of addicts treated and the associat-
ed appropriate multiplier factors that are used in the Extrapolation at each stage described above can be
calculation of the total number of addicts. The rele- made by a variety of methods, and requires simple
vant information is presented in table 28 below. This assumptions to be made. In this report the reason-
is a difficult estimation exercise and considerable able assumption is made that prevalence rates that
levels of variability are likely to be introduced into are estimated for the population in the clinic catch-
the estimation, although in methodological terms it ment areas within a province—that is, in those loca-
is probably smaller than likely errors produced by les with identified specialist drug treatment clinics—
other non-multiplier methods. Reliance on a single can be applied to the remainder of the population
multiplier value is likely to be hazardous in terms of in the province.
accuracy of the resulting estimates and figures are
presented in table 28 for a likely minimum and maxi-
mum value of the appropriate multiplier.
Estimation of prevalence
To focus the estimation and reduce the potential for
Method of estimation: sampling variation and error, the estimation exercise
geographical coverage has been carried out on a target population of males
aged 15-45 years. For many reasons, discussed in the
Estimation of the prevalence of drug addiction report, estimating drug addiction prevalence amongst
through a treatment multiplier benefits from taking women has not proven possible. Restricting the
account of geographically local variation both in prevalence to the age band of 15 years to 45 years
treatment facilities and in the proportion of addicts of age covers the vast number of addicts in the
treated. Data have been collected in four cities and population; selecting this age band where addicts
in 36 locales spread across the four provinces of are most densely found allows a more accurate esti-
Sindh, Punjab, NWFP and Balochistan. The FATA mate to be made than would be possible if the less
were included with NWFP for purposes of the preva- densely found, very young and very old were in-
lence estimation exercise. With the exception of cluded in the prevalence denominator. Choosing this
Balochistan, it has been possible to estimate drug age band is justified both from previous Pakistan
addiction prevalence on a province-by-province studies and from the present four cities study of
basis. In Balochistan, information was too sparse to interviewed addicts.
allow satisfactory estimation of the prevalence level.
Figures are presented in tables 25 to 29 for:
Within the provinces, all locales provided informa-
❏ The locales in which the four cities study
tion on the proportion of addicts treated over the pre-
was carried out, where the most detailed
ceding year (for step (b) of the estimation proce-
information is available;
dure). Enumeration of the number of addicts in
treatment over the preceding year (step (a) of the ❏ The remaining locales, incorporating an
estimation procedure) can be carried out in the sub- extrapolation from clinics with treatment
set of locales where specialist treatment clinics have clinics to those without clinics;
been identified. Within each province two levels of ❏ All locales studied in the survey.

52
 N AT I O N A L D R U G C O N T O U R M A P P I N G

Figures for each province separately are presented repeatedly refine the estimates of prevalence made
in the annex tables A.III.1 to A.III.5 in an identical by extrapolation. These data are:
fashion. Relationship of drug use to more detailed
breakdown of age distribution;
Table 25 constructs a demographic profile of the four
Relationship of drug use to more detailed
cities locales and the remaining locales, using data
description of the rural/urban distinction;
from the National Census.
Relationship of drug use to more detailed break-
Table 26 constructs the relevant target population of down of the catchment areas of the clinics;
males between 15 and 45 years of age, using data Relationship of drug use to numbers of house-
on age and sex from the National Census. The hold or implied household size;
national age ratios are presumed to apply across all
Relationship of drug use to other measures of
locales.
social structure such as poverty, literacy.
Table 27 presents a profile of specialist clinic access, The initial estimates provided here use:
in terms of inpatient beds provided, the number of
(a) the broad age-groups of 15 to 45 years of
admission episodes in the year, the number of
age, where most drug addiction occurs;
patients admitted in the year, and the number of
these that are male heroin addicts or injectors in the (b) the distinction between the four major
target age range. cities selected from each of the four pro-
vinces and the rest of the locale (district)
Table 28 presents the estimates from the key inform- in which they are located;
ant study of the likely proportion of addicts who have (c) the assumption that the catchment area for
been treated in the identified specialist clinic, and the clinics is essentially the locale (district)
the associated multipliers. The number of addicts in in which it is located—key informant infor-
each locale is estimated, using a correction factor of mation suggests that about 10% of addicts
90% to allow for addicts who are treated but come are treated from outside the locale;
from outside the locale. In addition to the median
(d) the characterization of each local as either
estimate, a likely maximum value and a likely mini-
“rural” or “urban”, as described in the study
mum value are shown, derived from the extent of vari-
design section of the report.
ability in key informant reporting in each locale.
A subsequent wave of analysis refining these initial
Table 29 shows the estimated prevalence rate (esti-
estimates can take recognition of
mated from those locales with specialist clinics, and
assumed to apply also to locales without identified (a) a greater treatment attendance by those
clinics), the number of addicts in the locales with addicts over 25 years of age;
clinics, and the estimated number of addicts in the (b) the urban and rural male population ratios
province, assuming the same prevalence rate applies within each city’s district;
throughout the province. (c) the differential treatment attendance from
the city rather than the district where the
clinic is located;
Improving the initial estimates (d) the extent of the rural versus urban popu-
lations difference within in each of the rural
The present report uses the simplest and most direct or urban locales.
of possible assumptions. Subsequent refining of the
estimates can be carried out in phases that intro- The procedures used in this report give rise to a range
duce more sophisticated use of available data at of likely estimates of overall prevalence and overall
each of stages reported above. The relationship—if numbers of heroin addict or injectors. In an analytical
any—of drug use to these features is exploited to exercise of this sort, where the procedures are not, nor

53
D R U G A B U S E I N PA K I S TA N

could be, simply to count the numbers of addicts, between the data elements and the relationships
assumptions must be made about the viability of any between the component surveys. These initial esti-
calculation with regard to both validity and reliability. mates reported here give a range that should be robust
The resulting range of different estimates must neces- first approximation to the overall prevalence of hard
sarily depend upon the details of the different assump- drug addiction. Further analysis refining and possibly
tions that can be made concerning the relationships improving these analyses can be carried out.

Table 25. Basic demographic profiles provided by the census data, using figures for
each of the locales in the study

Total of the Karachi Lahore Peshawar Quetta Total of

Total of four city division district district district other
all locales locales (Sindh) (Punjab) (NWFP) (Balochistan) locales

(i) Locale demographic profiles

Households 6 119 483 2 813 698 1 531 234 901 558 281 456 99 450 3 305 785
Male 22 121 618 10 017 487 5 261 712 3 262 904 1 067 397 425 474 12 104 131
Female 20 023 358 8 795 236 4 540 422 2 949 811 971 232 333 771 11 228 122
Total 42 144 976 18 812 723 9 802 134 6 212 715 2 038 629 759 245 23 332 253
Urban males 13 036 019 8 513 290 4 978 253 2 695 022 522 940 317 075 4 522 729
1981 population
Total 21 867 553 10 477 795 5 437 984 3 544 942 1 113 303 381 566 11 389 758

Table 26. Reduced overall target population of males in the age band of 15 to 45 years

Total of the Karachi Lahore Peshawar Quetta Total of

Total of four city division district district district other
all locales locales (Sindh) (Punjab) (NWFP) (Balochistan) locales

(ii) Locale target age-band: males aged 15-45

Assumed % of
males in target 42.20% 42.20% 42.20% 42.20% 42.20% 42.20%
Number of males in
target population
(millions) 9.335 4.227 2.220 1.377 0.450 0.180 5.108

54
 N AT I O N A L D R U G C O N T O U R M A P P I N G

Table 27. Clinic profiles derived from the treatment register update study

Total of the Karachi Lahore Peshawar Quetta Total of

Total of four city division district district district other
all locales locales (Sindh) (Punjab) (NWFP) (Balochistan) locales

(iii) Local clinic profiles

Total number of
beds provided in
specialist clinics 2 564 2 042 1 466 353 149 74 522
Total number of
treatment admissions 17 053 13 464 7 887 2 531 2 076 970 3 589
Total number of
inpatients admitted
in year 11 166 9 015 5 012 2 356 1 377 270 2 151
Proportion of
admissions that are
male inpatient
heroin cases 0.9 0.71 0.81 0.85
Number of admissions
of male inpatient
heroin cases 9 311 7 528 4 511 1 673 1 115 230 1 783

Refinements to initial multipliers informants and their suitability in this

regard.
for estimating the number
In further more detailed waves of analysis these esti-
of addicts mates could be refined by:

The initial estimation of the number of addicts (a) comparing the number of addicts in treat-
requires a count of the addicts in treatment over the ment with the number of (possibly repeat)
past year and a multiplier reflecting, for each one treatment episodes generated by the
addict in treatment the number of addicts not in treat- addicts over a year, and the total number
ment. Within the data collected, variations in the of days spent in treatment;
detail of the definitions and calculations of these two (b) triangulating the number of addicts against
quantities can be made. The current initial estimates the numbers in treatment and the numbers
reported here are based on: in prison over the previous year;
(a) the estimated number of male inpatients (c) taking different ranges of feasibility for the
in the specialist clinics who are being multipliers, depending upon the number
treated for hard drug use—heroin use or of addicts with whom the key informants
drug injecting; have had contact in the last year, or
(b) the multiplier estimated from the propor- depending upon the reported treatment
tion of interviewed addicts treated for hard ratios amongst the acquaintances of the
drug addiction and the key informants' interviewed addicts;
direct experience of this proportion—this (d) incorporating key informant assessments of
generates a range of possible multipliers the difficulty of obtaining treatment in the
depending on the experience of the key locales.

55
D R U G A B U S E I N PA K I S TA N

Table 28. Treatment multipliers derived from key informant and addict interview data

Total of the Karachi Lahore Peshawar Quetta Total of

Total of four city division district district district other
all locales locales (Sindh) (Punjab) (NWFP) (Balochistan) locales

(iv) Key informant and interview treatment multipliers

Range of estimated proportion of addicts who received treatment in year (see text)
(a) Maximum likely
proportion who
received treatment 0.2633 0.3000 0.1000 0.4722 0.2567 0.0224 0.2567
(b) Median likely
proportion who
received treatment 0.1225 0.1000 0.0800 0.2056 0.1693 0.0217 0.1464
(c) Minimum likely
proportion who
received treatment 0.0500 0.0400 0.0400 0.0500 0.0954 0.0212 0.0600

Range of estimated derived multiplier, representing addicts not receiving treatment in year
(a) Minimum multiplier
(applies to clinics’
catchment area) 3.80 3.33 10.00 2.12 3.90 44.64 3.90
(b) Median multiplier
(applies to clinics’
catchment area) 8.16 10.00 12.50 4.86 5.91 46.08 6.83
(c) Maximum multiplier
(applies to clinics’
catchment area) 20.00 25.00 25.00 20.00 10.48 47.17 16.67

Range of estimated number of addicts in population (see text)

Correction factor for addicts
out-of-catchment (see text) 0.90 0.90 0.90 0.90
(a) Likely estimated
minimum number of
addicts in catchment 62 995 56 917 40 597 3 188 3 911 9 221 6 078
(b) Likely estimated
median number of
addicts in catchment 89 157 73 517 50 747 7 322 5 929 9 518 15 640
(c) Likely estimated
maximum number of
addicts in catchment 189 319 151 868 101 493 30 110 10 522 9 743 37 451

Interpretation of the Supplementary calculation sheets show information

for each locale that is used to derive the summary
initial estimates and their calculation sheets.
extrapolations
Various technical details are explicated in a later
Calculation sheets are included that show the esti- document, but these calculations provide a first pro-
mation procedure step-by-step as described above. visional estimate of prevalence of heroin addiction,
They produce a range of possible estimates of preva- by making a variety of assumptions. The main fea-
lence for the country as a whole, broken down by tures of the results are summarised below, with some
the cities in the four cities study and by province. cautions and guidelines on their interpretation.

56
 N AT I O N A L D R U G C O N T O U R M A P P I N G

Table 29. Extrapolated estimates of the numbers of addicts in the country as a whole

Total of the Karachi Lahore Peshawar Quetta Total of

Total of four city division district district district other
all locales locales (Sindh) (Punjab) (NWFP) (Balochistan) locales

(v) Prevalence estimates and numbers of addicts (extrapolated)

(a) Minimium likely

Minimum likely target
population prevalence
rate (in clinics’
catchment area) 0.67% 1.35% 1.83% 0.23% 0.87% 5.14% 0.12%
Estimated minimum
number of addicts
(in locales with
identified clinics) 62 995 56 917 40 597 3 188 3 911 9 221 6 078
Estimated minimum
number of heroin
addicts in total
(see text) 97 637 56 917 40 597 3 188 3 911 9 221 40 721

(b) Maximium likely

Maximum likely target
population prevalence
rate (in clinics’
catchment area) 2.03% 3.59% 4.57% 2.19% 2.34% 5.43% 0.73%
Estimated minimum
number of addicts
(in locales with
identified clinics) 189 319 151 868 101 493 30 110 10 522 9 743 37 451
Estimated maximum
number of heroin
addicts in total
(see text) 398 302 151 868 101 493 30 110 10 522 9 743 246 434

In this regard the following points are to be noted. The range of estimates from this study suggests as
a possible upper limit up to 500,000 heroin users
There are some less populous geographic areas that and other drug injectors, in round figures, amongst
have been excluded from the prevalence estimates males in the 15 to 45 year old age band. Whilst
for a number of technical reasons in these provi- this figure is considerably lower than previous
sional figures. For example Balochistan province national estimates, there are good reasons for the
outside the Quetta district proved too difficult to difference.
estimate by these preliminary methods.
The overall prevalence in this study is for hard-core
Prevalence rates quoted in the provisional figures heroin users and drug injectors only, with no inclu-
are for males aged in the target age-band of 15 to sion of alcohol or charas abusers. The reasons for
45 years of age, which according to this research and advisability of excluding these other abusers from
to previous government figures constitutes the vast this study are detailed in the preceding sections. In
majority of the heroin-users in the population. previous figures reported for Pakistan the definitions

57
D R U G A B U S E I N PA K I S TA N

of addiction have not been so precisely expressed would be considered unrealistically high, but it
nor were they built into the calculations from the should be borne in mind that these are likely maxi-
outset, as this study has done. mum figures that are being quoted for the worst
affected Pakistan cities. Even so, previous reports of
There is large geographic variation in the density of heroin addiction in Pakistan were considerably
hard-core drug use across Pakistan, in particular the higher again, and as a result fell a long way short
difference between urban and rural rates in some of achieving international credibility.
provinces. To allow for these variations in prevalence
rates is difficult, and failure to do so adequately The overall prevalence expressed in terms of the whole
could produce misleading results. The results in this population of Pakistan is around one third of one per
report have been adjusted as closely as possible cent and is not out of line with other countries
using available data to capture this variation. expressing prevalence rates of one quarter or one half
of one per cent. There are various special considera-
Within some cities the prevalence rate is estimated tions relating to Pakistan in term of the country’s age
to be very high, possibly as great as 4%, compared structure that make such international comparisons
with about 0.5% in some rural areas, for males in difficult to interpret, but the provisional figures sug-
the target age-band. These figures—1 male in every gest that, in the international setting, Pakistan has
25 in the age-band—by most countries’ standards one of the highest rates of drug addiction.

58
9
Drug abuse in Pakistan: the implications of
the Pakistan national assessment study

Comparisons with the earlier national assessment studies

Comparisons of the data presented here can be made with the findings of the earlier assess-
ment studies but caution is required as the earlier work tended not always to differentiate
between drug types (including alcohol) and the sampling strategy varies between these stud-
ies. Nonetheless, sufficient comparable information is available to draw some conclusions on
changes in patterns and trends in drug abuse in Pakistan.

With respect to the overall pattern of drug abuse in the country, in 1982 hashish and charas,
which we will refer to here for convenience as cannabis, was estimated to be the most com-
monly consumed illicit substance. The 1982 study estimated that 3.4% of the adult male pop-
ulation was using this drug. The later surveys suggest that heroin use overtook cannabis use
around 1986. Cannabis abuse was reported to have declined between 1982 and 1986 but sub-
sequently to have slowly risen after this date, although remaining at a lower consumption level
than for heroin. The current study does not stand in conflict with this analysis, neither does
it necessarily support it: with respect to the direction of trends it may well be that cannabis
use declined slightly during the mid 1980s and then subsequently increased.

Equally difficult to resolve are questions concerning the relative consumption levels of cannabis
and heroin in Pakistan. Did cannabis abuse actually fall lower, as reported, than heroin abuse
or was this finding due to estimation errors? The present data do not allow us to determine
the truth of this matter, as either scenario would be consistent with the current findings.
Relevant, though, is that although the results presented here give no definite answer, they do
raise the question as to what extent earlier estimates of heroin abuse might have been inflat-
ed. Addressing these issues cogently is greatly complicated by the time intervals between the
assessments exercises, a fact which serves to emphasise that the accurate identification of
consumption trends over time relies on the development of a continuous surveillance mech-
anism. This issue is discussed in detail below.

The key informant data in the 2000 exercise strongly suggest that cannabis in one form or
another (marijuana, charas, etc) is the most commonly used drug in the country in terms of
lifetime use and prevalence over the last year. Whilst this study was not designed to give an
estimate of the number of cannabis consumers in Pakistan, it can be posited that this figure
is likely to be considerable. Any estimate will depend specifically on the period prevalence
measure used (lifetime, last year, last month etc). However, if period prevalence for either life-
time or last year is considered, the total numbers of individuals is likely to be higher than a
million. For a point of reference, if the 1982 prevalence estimate is taken in conjunction with
today's figures it would suggest there would be around 1.3 million current users. The general
opinion would be that consumption today is probably considerably higher than in 1982, so

59
D R U G A B U S E I N PA K I S TA N

this figure is itself likely to be an underestimate. which can provide intensive and long-term support
Both the scale of cannabis use and its relationship to those that require it.
to health and other problems merit further research
attention in Pakistan.
Table 30. Comparisons of heroin abusers
Of particular concern is the use of illicit drugs by in the 1993 and 2000 studies
young people. Sixty-five per cent of key informants 1993 2000
identified hashish and charas as drugs used by
young people. This finding is consistent with the Mean age at interview 32 33
1993 study and moreover is concordant with inter- Age first use heroin 26 22
national patterns of cannabis use, where prevalence Percentage of sample
levels are usually found to be higher among the age 15-20 years of age 24 5
young. Future studies may productively explore atti- Percentage female 4 1
tudes to, and patterns of use of, cannabis and other Sample size 506 1 049
substances among young people. School surveys
have extensively been used for this purpose else-
where. Whether a school survey of drug use in The findings of the 2000 study also broadly sup-
Pakistan is feasible and what resources would be ported the findings of the 1993 survey in respect to
required are appropriate questions to address in the provincial breakdown, with North West Frontier
future discussions about developing a permanent Province being identified as the province with the
drug information system for Pakistan. lowest prevalence of heroin use. Opium abuse was
reported to be at low levels and in decline in both
The early studies suggested that heroin use was studies and poly-drug abuse was common among
“common in the younger population 16-30 years” drug abusers in both research exercises.
(1993 study). In the 2000 study, fewer young respon-
dents were interviewed, even though the mean age
of respondents in both samples were very similar Drug use by women
(see table 30). In the 1993 exercise, 24% of heroin
users interviewed were between 15 and 20 years old. None of the assessment exercises conducted in
In the 2000 study this figure fell to 5%. This suggests Pakistan has been able to explore drug use among
that the age distribution of the earlier sample was women in any detail. The 2000 assessment exercise
more evenly spread but, given that the sampling does indicate that whilst drug abuse by women is
strategy used in the studies is not identical, caution not, for most drug types, likely to be found at the
should be used in drawing conclusions from this same levels as among the male population, it clear-
with respect to the overall makeup of the heroin ly exists. In the key informant data, a small but sig-
abusing population in Pakistan. More informatively, nificant number of respondents reported heroin
the mean age of first heroin use has fallen from 26 abuse among women in their locales. The estima-
to 22 years; and as the mean age of subjects inter- tion of psychotropic use was even higher. If women
viewed has remained constant, the 2000 sample are abusing drugs in Pakistan, they are not access-
therefore represents more long-term abusers of ing treatment in any great numbers. The treatment
heroin. audit suggested that only 3% of current patients
were women.
In any respect the data suggest that street, treat-
ment and prison populations of heroin abusers are Two things are then clear from the current study.
characterized by a substantial proportion of long- First, more information on drug use by women in
term users whose needs are therefore likely to be Pakistan is required to allow a better understanding
both considerable and varied. Successfully interven- of the extent of problems among women and to
ing with such a group is likely to require services gauge their related needs. Such studies will be
that can provide a broad range of interventions and methodologically challenging and require sensitivity

60
 I M P L I C AT I O N S O F T H E N AT I O N A L A S S E S S M E N T S T U D Y

to the difficulties of accessing and interviewing provision are reliant on the delivery of a high qual-
women on the topic of their drug consumption. It ity service. Staff training and the development and
may well be that qualitative inquiries using female implementation of appropriate therapeutic proce-
interviewers will be the way forward here. dures are all important here. The term “treatment”
is used to describe a wide range of activities that
Second, attention should be given to developing address a diverse set of needs and patient charac-
treatment facilities that would prove more attractive teristics. Some positive evidence of the development
to women with drug problems. The Guiding of comprehensive, high quality treatment services is
Principles on Drug Demand Reduction identify the identified in this study. However, other evidence also
need for gender sensitive treatment services. This suggests that in some areas treatment services are
issue is an international one, which poses a chal- less well regarded. For example, overwhelmingly key
lenge for both developed and developing countries. informants agreed—or strongly agreed—that treat-
In Pakistan, cultural factors may make achieving this ment services in their area did not offer a good serv-
goal particularly demanding. Nonetheless this ice to those with drug-problems.
remains an important issue for the future develop-
ment of drug treatment services. The UNDCP is currently working with the Government
of Pakistan on measures to improve the quality of
services in the country. A more detailed investigation
of current practices and procedures would facilitate
Access to and delivery of this work. The long-term goal must be to broaden the
drug treatment range of treatment options and to ensure that all
services provide the high quality care that is currently
A finding of both the key informant exercise and the found in some of the best facilities in the country.
addict interviews is that increased access to drug
treatment is urgently required in Pakistan. Nearly all Many of those on drug charges interviewed in prison
key informants agreed that many drug users in their were severely dependent and would benefit from the
area were too poor to get treatment. All the addicts provision of therapeutic services that address long-
in the interview samples were suitable for treatment standing drug problems. This need is emphasized by
and furthermore three-quarters expressed a self reflection on the fact that some of those incarcerat-
identified need for help, yet many respondents (66%) ed have arrived in prison through the intervention
reported previous failed attempts to access treat- of their families, who are no longer able to cope with
ment, overwhelmingly because of financial reasons. the behaviour of their family member. Drug abusers
Street recruited addicts appeared least able to are housed in separate barracks in Pakistani prisons,
access help; in general those addicts interviewed in which could facilitate the development of services
treatment settings were better educated and more within the prison setting.
likely to be in employment, supporting the sugges-
tion that income levels influence treatment access. Many of those interviewed in the street or treatment
Some evidence was found that current treatment setting had previously passed through the prison
capacity was not being fully utilized despite the system. The importance of developing post-release
identified need of many drug abusers for therapeu- rehabilitation services for drug abusers has been noted
tic interventions. A successful expansion of treat- elsewhere. In both instances—for drug abusers in
ment provision is therefore likely to be dependent prison and for those leaving prison with a history of
at least in part, on delivery mechanisms that allow drug abuse—the potential exists to further develop
uptake of services by a wider section of the addict demand reduction programmes. Delivering effective
population. drug treatment and prevention programmes within the
prison setting is not an easy undertaking. Working in
Drug treatment has been repeatedly shown to be a this setting presents a number of challenges but this
cost-effective response at a national level to drug remains an area in which considerable potential exists
problems. However, clearly the benefits of treatment for the development of drug services.

61
D R U G A B U S E I N PA K I S TA N

Drug injection Developing a permanent drug

information system to inform
In the 1993 exercise, concern about drug injection policy and action
and its link to HIV infection were expressed, but
injecting appeared to be very rare in Pakistan. The year 2000 national drug assessment study has
Perhaps the most worrying conclusions of the 2000 provided an overview of drug abuse problems for
assessment exercise is that drug injection has now Pakistan. The study's aims were ambitious and the
become widely established, that it is continuing to fact that they have been largely achieved owes much
grow, and that many drug injectors are engaging in to the hard work of the research team in Pakistan, to
high risk practices that make them vulnerable to the implementation organizers and to the good will
infection by HIV and other blood borne diseases. of all those individuals in the country whose support
was necessary for collecting the required information.
To date, what evidence that exists on HIV infection That so many people concerned with drug problems
suggests that it has not become widespread in Pakistan were prepared to give up their time to
amongst Pakistani drug injectors. Against this opti- help with the assessment exercise demonstrates that
mistic note, though, two notes of caution must be the importance of having sound information on the
sounded. Firstly, infection rates have been shown in drug abuse phenomenon is widely accepted. That this
other countries to rise dramatically over short peri- project was implemented jointly with the Anti-
ods of time, when the behaviour of drug injectors is Narcotics Force, and that it would not have been pos-
favourable to its spread. Secondly, HIV surveillance sible without their work and support, reflects the
studies have not been conducted with any intensity importance the Government of Pakistan has invested
among drug injectors in Pakistan and therefore the in basing its policy and actions on sound evidence.
possibility must be considerable that unrecorded In Pakistan, national expenditure on drug problems
infections exist. has to be considered in relation to other pressing
calls on the public purse, and it is therefore particu-
It is therefore important that any future drug larly important that investments be guided by an
research and surveillance activities include HIV sero- accurate assessment of the situation.
prevalence monitoring wherever possible. Some
work is being currently developed in this area, sup- However, whilst the mass of data collected by this
ported by UNAIDS and UNDCP. More intensive activ- study is impressive, it represents a only a snap-shot
ities are urgently required. The 2000 study has of the drug abuse situation. Studies conducted many
demonstrated that the research capacity exists in years apart cannot provide the on-going relevant
Pakistan to efficiently access drug injectors in street information required for policy formation. The assess-
and institutional settings at relatively low costs. ment study should be viewed as an important first
Further research is required with drug injectors not step in a permanent ongoing drug information
only to monitor infection rates but also to under- system for Pakistan. The data presented here can
stand the dynamics of drug injection practices in provide a valuable baseline for future continuous
order to inform prevention strategies and other monitoring exercises. UNDCP through its Global
interventions. There is also a priority need to devel- Assessment Programme has already begun discus-
op interventions that provide support to drug injec- sions with the Government of Pakistan on how such
tors and that address the HIV risk behaviours of this provision can be established. Drug information sys-
group. UNAIDS, UNDCP and the government of tems typically rely on the collection of a range of rou-
Pakistan have been developing a strategy in this tine indicators and other data and then provide a
respect. Given the relative size of the drug abusing forum for experts to discuss regularly the implications
population in the country, the high levels of risk of the information.
behaviour detected, and the corresponding potential
for a future epidemic, urgent and effective interven- The four city sites included in this study would make
tions are required in Pakistan. a good starting point for further work: city-based

62
 I M P L I C AT I O N S O F T H E N AT I O N A L A S S E S S M E N T S T U D Y

networks can provide locally relevant information, Other research needs

are sensitive to emerging trends, and can contribute
information to develop a national picture of drug Understanding the composition and purity of drugs
abuse trends. Successful networks based on this available on the illicit market and monitoring changes
city-network design are found elsewhere including in this area have direct relevance to understanding
the United States and Europe. patterns of abuse and changes in consumption. It was
suggested during the pilot work for this project that
One information source with clear potential for declining heroin quality was impacting on the con-
development in Pakistan is the reporting of the sumption patterns of the abusing population, either
records of treatment agencies. Many countries in the by changing dosing habits, affecting the choice of
world have developed treatment attendance regis- route of administration or encouraging the substitu-
ters that provide basic details in a summarized form tion of street heroin for other pharmaceutical drugs.
of those accessing treatment facilities for drug prob- In the study, over half of respondents agreed with the
lems. Such information is cost effective and easy to statement that “heroin users were changing to other
collect in a standardized form, provided the require- pharmaceutical drugs in their locale”. This question
ments are modest and that the system works on the was not addressed in further detail by the current
basis of anonymous records. As almost all of study and it remains worthy of further investigation.
Pakistan's treatment facilities collect relevant infor-
mation on client profiles, considerable potential It would be helpful to know more about the com-
exists for the development of monitoring in this position and nature of the drugs consumed by those
area. with drug problems in Pakistan. UNDCP has been
working with the government of Pakistan to improve
In developing a continuous drug information sys- the facilities available for the analysis of drug sam-
tem, it is important to remember that, as the key ples and a number of laboratory services have been
informant exercise reveals here, a considerable developed. Any future drug information system that
amount of information is available from the tech- might be set up would benefit from including an
nical experts working with drug abusers and that analysis of the drugs available on the illicit market.
Pakistan is fortunate to have many well-qualified A number of models exist for this type of activity,
professionals working in the drug abuse field. One including the analysis of seizures taken from differ-
important role of an information system is to pro- ent levels of the supply chain.
vide a forum for professionals in which to discuss
and analyse the current situation. It is also impor- Another question that emerged during the course of
tant to remember that drug use remains a hidden the study was the suggestion that there was growing
behaviour and that assessments of drug abusers in solvent abuse among young people and street chil-
street and other natural settings can often provide dren in particular. Street children are recognized as
a unique insight into the extent of currently exist- group who are particularly vulnerable to drug prob-
ing and newly developing problems. Whilst there is lems. Data from this study do not allow comment on
clearly much to do, Pakistan is fortunate in that suf- the issue in any detail. Solvent abuse was most com-
ficient capacity does exist within the country to monly reported from urban areas and from the Punjab
develop the capacity to monitor and to better province in particular. The extent to which solvents are
understand drug problems. Such understanding can being abused by this marginalized and vulnerable
be translated into well targeted, timely and effec- group of street children deserves further attention, as
tive responses. The Pakistan 2000 national assess- does the wider question of what can be done to alle-
ment study has made a valuable contribution to this viate the drug and other social problems amongst this
endeavour. particularly needy group of young people.

63
 Annex I.
Methodological discussion

The method used for the national assessment exercise is to carry out a set of surveys on particular aspects
of the drug problem that each in its own right provides vital information on drug use in Pakistan; and
which when taken together also can provide a prevalence estimate of drug use. Below we discuss some
of the methodological issues this approach raises in more detail.

Drugs of abuse considered by this study

Drug abuse is a complex and multi-faceted phenomenon. It is therefore of critical importance to be clear
and precise about the particular aspect of the problem that is being investigated by any study. This issue
is particularly important for the question of prevalence estimation. For example, depending on which of
the commonly used prevalence measures is selected—lifetime prevalence (ever use) or current use (usu-
ally last month)—the estimates produced will vary considerably. Similarly, different drugs often have very
different use profiles. One problem apparent in aspects of some of the previous National Assessments
was the need to clearly delineate between different drugs of abuse.

One of the first tasks of the research team in the present study was to produce a summary list of drugs
of abuse that was appropriate for inclusion in the research questionnaires but also covered the main
issues of concern about drug abuse in Pakistan. As the questionnaires would require the same list of
drugs to be used in a number of repeat format questions, from a design point of view it was desirable
that the list be as short and clear as possible. Questionnaires with overly long and detailed drug lists
tend to perform poorly in fieldwork. However, at the same time it was important to address the main
areas of drug abuse. The final drug list selected for used in this study after discussions with local experts
contains the following items. Some drug types have been grouped and alcohol, which is an illicit sub-
stance in Pakistan, is also included.

Hashish and charas (cannabis type)

Heroin
Opium
Other opioids (e.g. morphine, temgesic, sosegon)
Cough Syrups (when used for intoxication)
Psychotropics (e.g. tranquillizers, such as ativan, valium, lexotonil and amphetamines)
Alcohol use
Drug use by injection (any drug)
Inhalants (glue, solvents, etc.)

Component study 1. National contour mapping exercise

The initial aim of the study was to provide an overview of the contours of the national drug abuse situ-
ation in Pakistan. Whilst this exercise could not explore every aspect of drug abuse in detail, it could be
expected to provide an overview of the situation. In particular, patterns and trends since the last natio-
nal exercise, could be explored and expert assessments on the scale of relative problems associated with

65
D R U G A B U S E I N PA K I S TA N

different aspects of the drug abuse situation produced. As non-government organizations (NGOs) and those involved
a national mapping exercise, the study would also help to with providing treatment to drug abusers.
assemble a contemporary picture of regional differences
in the nature and scale of the drug abuse problem. Critical A questionnaire was produced especially for the purposes
questions for further research activity could also be iden- of this study. The questionnaire was prepared in both
tified although not necessarily explored in detail. English and Urdu. To ensure that the translation and the
Nonetheless, producing an overview of the current situa- English versions of the questionnaire were identical in
tion is critically important for planning and targeting meaning each question was reviewed in detail in both lan-
demand reduction programmes. guages at a special training workshop held before inter-
viewing commenced.
In detail, the mapping exercise would map and explore
regional difference in respect of the following issues: A team of specially trained interviewers administered the
❏ The relative popularity of different drug types in questionnaire. Interviewers were selected from across
Pakistan; Pakistan and were for the most part associated with either
treatment and rehabilitation services or academic insti-
❏ Patterns of use for different drug types among
tutes. Provincial coordinators and an official of the
rural and urban populations;
Anti Narcotics Force supervised the interview teams.
❏ Sex and age differences in drug consumption pat- Completed interviews were checked for accuracy and ver-
terns; ification measures were also used to ensure interviewer
❏ Trends in the use of different drug types; compliance with the study protocol.
❏ The existence of drug injecting;
Data entry was conducted by a Pakistan-based informa-
❏ Routes of administration for opioid users;
tion technology company. Additional data cleaning and
❏ Differences in the extent of local problems relat- the preparation of tabulations were conducted under the
ing to drug abuse; supervision of the statistical advisor to the project at the
❏ Relative extent to which different drug types were National Addiction Centre (UK).
causing problems;
❏ The need for educational programmes;
❏ Opinions on the availability of treatment; Sampling strategy for key informant
❏ Changes in treatment uptake over time. interviews
Method: Two research strategies were employed for this The rationale for this part of the study was to produce a
part of the study: national picture of drug use in Pakistan. Therefore a sam-
❏ A desk review was undertaken of all the available pling strategy was devised that would provide national
research material on drug abuse in Pakistan. coverage. Study locations (hereafter referred to as locales)
❏ A key informant study was conducted to provide across Pakistan were selected for inclusion in the study.
an overview of the contours of the drug abuse This list of locales has been drawn up by relevant experts
problem in Pakistan. for the authorities in Islamabad (ANF). To ensure that the
study reflected difference between rural and metropolitan
Materials for the desk review were obtained from the files areas, locales were stratified in this dimension. In total,
of the Anti-Narcotic Force, UNDCP regional office and from the Anti-Narcotics Force (ANF) in discussions with the
other governmental and non-governmental sources. In study coordinator selected 18 urban and 18 rural matched
particular, the previous national assessments (from 1982) locales across Pakistan to be used in this study. The
were reviewed to place the current study in context. Data locales were selected as being broadly representative of
from the 1998 National Population and Housing Census the country as a whole and reflected the sampling stra-
of Pakistan were also audited along with a number of spe- tegy used in previous national assessment exercises.
cialist ad hoc studies and reports. Practical as well as statistical issues were important in
selecting locales. For the purposes of analysis it was there-
Key informants were selected who could be reasonably fore accepted that some adjustments might have to be
expected to have an informed view of drug abuse in made, based on the 1998 census data, to ensure the popu-
their locale. They included community leaders, teachers, lation parameters of study sites reflected the national
business professionals, police officials, and members of population characteristics.

66
 ANNEX I. METHODOLOGICAL DISCUSSION

In each locale a minimum of five key informants were skills and backgrounds to access street-addicts was rec-
selected. Key informants were selected on the basis of ognized and it was possible to identify suitable inter-
obtaining differing perspectives across occupation groups viewers and to include them in the fieldwork team.
and on the basis of them having some knowledge about
drug abuse in their communities. It was important for the The objective of this module of the study was to produce
accuracy of the study that those interviewed had life expe- detailed information from each province on one selected
riences that allow them to comment on drug abuse pat- locale from the original 18 urban locales. The aim was to
terns. Training and instructions were given to interviewers characterize the nature of hard-core drug abusers in the
on the appropriate selection of these respondents. locale within three population groups of heroin users and
injectors, namely those in prison, those in treatment, and
The key informant data would be used to produce a street addicts.
national map of drug use patterns for Pakistan as a whole.
This involved the projection of information contours from There were three reasons for selecting these population
the 36 locales where interviews took place onto a nation- groups. First, drug abusers not in institutional settings are
wide map, giving due regard to geographical and demo- often different from those who are, and therefore it is
graphic details of the locales and the nation as a whole. important to include street recruited samples in any
assessment exercise. Second, drug treatment and criminal
justice represent two of the major responses to hard-core
drug abuse and therefore understanding more about the
Component study 2. Provincial city characteristics of the drug using populations in these
studies: an exploration of the heroin areas is important. Third, an analysis of the extent of over-
abusing and drug injecting population lap between these three sections can be identified and,
given that suitably strong information results from this,
Heroin use and the use of any drug by injection are widely local prevalence estimates can then be produced. This
recognized as among the most damaging patterns of drug issue is discussed in more detail below.
abuse both for society as a whole and for the individual.
Earlier in this report we have discussed the emergence of Four major provincial cities were chosen as the sampling
widespread heroin use in Pakistan from the early 1980s. sites for this part of the study. Key informant interviews
We also note concern in recent reports of the beginnings were also conducted in these locals. The cities selected
of an injecting subculture in Pakistan and therefore the were:
increased potential for HIV transmission amount drug Karachi,
abusers. This group of drug abusers also impact consid- Lahore,
erably on the health and criminal justice infrastructure. Quetta, and
For all these reasons, understanding more about heroin Peshawar.
use and drug injection was clearly a policy relevant prior-
ity. As such this topic was selected as for a more detailed The design of the provincial four city study focused study
investigation and as the focus for the subsequent preva- comprised therefore the following activities in each city:
lence estimation exercise. It should be noted that this (a) Structured face-to-face interviews with:
does not mean that other aspects of the drug abuse sit-
(i) Heroin users or injectors who were interviewed
uation were considered unproblematic or not meriting
on the streets,
detailed investigation. Rather, the resources available for
(ii) Heroin users or injectors who were currently in
this study only permitted a focused exercise with one
drug treatment,
group of abusers. In the conclusion to this report we high-
light other important research areas that data from the (iii) Heroin users or injectors who were currently
general assessment exercise would indicate require future within the prison population
attention. (b) An audit of local prison and treatment activity data
was required in the four selected cities, as this infor-
Successfully interviewing drug abusers is never an easy mation was needed for the prevalence estimation
task. This study's success in this area is largely due to the exercise and is also useful for exploring the impact
dedication and skills of the fieldwork team. The work also that heroin abusers were having on local services. The
benefited from recent experiences in Pakistan of conduct- treatment data were collected as part of the larger
ing small-scale studies with drug abusers. For example, National Audit (see below). Prison activity data were
the importance of using interviewers with the necessary only collected for the main prison dealing with drug

67
D R U G A B U S E I N PA K I S TA N

offences for each of the four cities. In particular it was The questionnaires for the three samples covered the
necessary to collect detailed information on prison same topics and only varied for those issues specifically
occupancy by drug addicts and drug peddlers. Speci- relevant to the individual sample group. Topics covered in
fically, the number of people sent to prison in the last the questionnaires included the following:
year for drug use and drug-related offences was Demographic information (sex, age, employment, and
required for each of the four cities. An audit form was living situation)
prepared for this purpose and used by specially History of attendance at drug treatment facilities
trained members of the fieldwork team who visited Attitudes to drug treatment
the prison and transferred information onto the form History of prison attendance
from the prison records. This process was facilitated Patterns of current drug use
by ANF officials contacting prison governors in ad- History of drug taking
vance and explaining the purpose of the study. The Injecting and HIV risk behaviour
consent of prison governors was required both for this Current problems
audit exercise and for the face-to-face interviews with Financial situation
the prison addicts (described below). The fact that Knowledge of other drug users
information for the audit was required only on those Severity of drug dependence
prisoners serving sentences for drug-offences and that
drug abusers are housed in separate barracks within
the prison, made achieving the objectives of the study
less demanding than would otherwise have been the Component study 3. A national
case. treatment register update

The third data collection module of the study was to update

Face-to-face interviews with the 1994 national register of drug treatment facilities in
heroin abusers Pakistan. Again the rationale for this exercise was that this
information was important in its own right and also that the
The target for this exercise was to conduct 1,000 interviews data produced could assist with an analysis of the scale of
in total with drug abusers. This would comprise 400 sam- hard-core heroin use in Pakistan. The 1994, register was con-
pled from patients currently in treatment (100 in each sidered in urgent need of an update as it was strongly sus-
city), 300 sampled from prisoners currently in jail (75 in pected that considerable inaccuracies now existed in the
each city), and 300 sampled from the street addict popu- register. The provision of drug treatment is one of the key
lation (75 from each city). This number of interviews was components of a comprehensive demand reduction strategy.
considered to be sufficient for the objectives of the study The debate on how best to develop treatment opportunities
and achievable within the available resources. requires as a first step an understanding of exiting provi-
sion. This exercise was intended to provide this under-
As with the key informant study, all interviews were con- standing by collecting data on current treatment capacity.
ducted using a set of structured questionnaires especially
prepared for the study. Questionnaires were in English and The 1994 report was used as a starting point for this exer-
Urdu and, as with the key informant interviews, con- cise. Interviewers then contacted the treatment centres
siderable effort was made to ensure accuracy in transla- and completed a brief, specially designed, audit form. This
tion. Methods for interviewer training and supervision process was facilitated by the fact that most treatment
were also the same as those described above for the key agencies keep relatively good records, although central
informant work. Data entry was conducted in Pakistan with collation of this information is poor.
additional data cleaning in the United Kingdom under the
supervision of the statistical advisor to the project. To help identify new treatment centres that had been
established post 1994, and those that had ceased operat-
Each questionnaire contained an introductory paragraph ing in this period, staff at those agencies that were suc-
explaining that the purpose of the study to the respon- cessfully contacted were asked if in their area:
dent, that all information collected was confidential and
anonymous and that participation in the study was ❏ they knew of any new centres stating up since
entirely voluntary. The importance of observing appropri- 1994, or
ate social research ethical standards was also discussed ❏ they knew of centres operating in 1994 that had
at length during the training of the field work team. subsequently closed down?

68
 ANNEX I. METHODOLOGICAL DISCUSSION

In addition, appropriate key informants were also asked contact, where the drug problem impinges directly on social
these questions during the national contour mapping part resources, are monitored as systematically as possible.
of this study (component study 1). In this manner the Ideally this should be done on a continuous basis, to pro-
study team were confident that the national coverage of vide a future steady flow of information on trends in the
this exercise was acceptably good and was estimated by drug abuse situation. This is a topic we return to in this
a local treatment expert at 90% or better. report's concluding section where some of the options for
the ongoing monitoring of the drug abuse situation are
considered.

Making use of the component studies: It was considered important in this study, however, to pro-
estimating the scale of heroin abuse duce a prevalence estimate for heroin abuse and drug
in Pakistan injecting. One reason for this was that current estimates
varied greatly and therefore there was considerable bene-
fit in generating a more reliable baseline figure. The meth-
In making a practical and policy-relevant assessment of
ods used in this study to generate a prevalence estimate
the country's drug-abuse situation the first concern must
for hard-core heroin use in Pakistan is described in detail
be the impact of illicit drug use on society with respect
in the section of this report that contains the estimates
to the problems caused and the social resources neces-
themselves. Here we provide a more general overview of
sary to remedy them. Some of these issues are addressed
the principles behind the exercise.
in the component studies of the national assessment exer-
cise as discussed above. In many respects the critical pol-
icy-relevant questions are “how does the drug abuse prob- In simple terms, two approaches are available to those who
lem make it self felt?” (i.e. problems) and “do current wish to generate drug prevalence estimates for the general
responses address these problems?” Lying behind these population. The first is to conduct some form of conven-
different aspects is the question of the current overall tional household survey using established survey methods
prevalence of drug use, which is a useful summary guide and probability sampling. This approach has a number of
to and non-specific indicator of the overall problem. problems. Most importantly, household surveys are costly,
and as drug use is a stigmatized and low frequency behav-
This importance of the value of prevalence estimates in iour, surveys do not always produce convincing estimates
themselves should not be overstated. The development of even when considerable resources are invested. This is
sound policy need not be reliant on an exact measure of especially a problem for drugs like heroin, whose use is
the current size of the abusing population. Nonetheless, often particularly stigmatised. A pilot household survey
having a reasonable estimate of the size of different approach was tested in the 1986, national assessment study
aspects of the drug abusing population is useful. It was in Pakistan. The results were extremely disappointing at it
decided that this analysis should be included in addition was concluded that this method was not appropriate to
to the other aspects of the study, rather than constitute recording drug abuse in Pakistan. Among the reasons for
the only objective of the assessment exercise. this were the difficulties in generating an appropriate sam-
pling frame and the simple fact that is was not considered
In Pakistan, in common with most countries in the world, likely that respondents would admit to drug use in the
the prevalence figure for drug abuse is difficult to estimate household setting. These problems are not specific to
with any degree of accuracy. This difficulty underlines the Pakistan and, arguably, household survey work never per-
need to focus on obtaining good data directly from the pri- forms well for the measurement of hard-core drug abuse.
mary problem areas, such as those individuals needing These problems lead the research team in discussions with
treatment. For example, indicators such as the level of first- Pakistan Government to rule out this approach as a sensi-
time demands for drug treatment are often in practice bet- ble option for this study.
ter than overall prevalence estimates as a tool for moni-
toring drug abuse trends over time. This is in part because The second approach to prevalence estimation rests on
they can be measured more accurately and more cost effec- more recent work to develop indirect statistical techniques
tively than actual prevalence can be estimated, in part for this purpose. These “benchmark/multiplier” methods
because they often serve as a better indicator of incidence remain relatively new but are now seen as having merit for
(new cases), and in part because they are a measure of estimating the prevalence of the use of drugs like heroin,
direct impingement of the drug problem on national even in countries with a long tradition of conducting house-
resources. It is important therefore that these points of hold surveys.

69
D R U G A B U S E I N PA K I S TA N

This approach has distinct advantages for the current realistic option in this study to attempt to generate robust
study as it is based on extrapolations from the compo- prevalence estimates for all drug types. Rather, the use of
nent studies described above, each of which also provides regular use heroin and drugs by injecting was selected as
valuable information in its own right. This makes the the core topic of the estimation work. There is consider-
method cost effective and means that the estimates pro- able overlap between these groups and for summary pur-
duced are only one of several useful results from the poses we will refer to them as hard-core addicts. The
assessment exercise. Given the difficulties in producing benchmark-multiplier methods used here are particularly
credible drug prevalence estimates, regardless of the appropriate for use with this group of drug abusers.
methods used, this approach was adopted as the strategy
most likely to guarantee the maximum amount of useful The prevalence estimation aspect of the study relies on the
information generated by the study. joint analysis of the component surveys. Each survey is
analysed separately to produce the relevant information and
A third, but less conventional approach to prevalence esti- assessments. These figures are then utilized in a joint analy-
mation, was the method employed by the previous natio- sis that makes use of the benchmark/multiplier indirect tech-
nal estimation exercises in Pakistan. This involved map- niques. The joint analysis takes the information on treatment
ping descriptive data produced by snowball samples on figures in the past year and information on the ratio of
to national estimates generated by key informants. This untreated to treated addict numbers to generate an estimate
method and the results from the previous surveys are dis- of the total number—treated and untreated—of heroin
cussed in detail elsewhere in this report. Whilst these addicts. A similar joint analysis of the component surveys
studies were competently conducted and did provide use- using imprisonment figures and imprisonment ratios is also
ful information on patterns of use, the prevalence figures carried out.
produced were open to question. Subsequently, the
method used to generate those estimates had been widely The methods used produce estimates that need to be har-
criticized and it was therefore decided that this was not a monised not only across the different geographic regions
fruitful approach for the current study to repeat. This issue but also across the different methods themselves. They are
is explored in more detail in the prevalence estimation supplemented by further indirect estimation techniques
section of this report. where the information collected is appropriate. The face-to-
face interviews with drug abusers provide information on
Some of the problems noted with the earlier estimation drug use patterns, frequency of treatment and frequency of
work in Pakistan stem from a failure to provide sufficient imprisonment. In particular these interviews establish the
clarity in respect of issues relating to operational defini- ratio of treated to untreated addicts over the last year and
tions. As we noted earlier, prevalence estimates for drug the ratio of those imprisoned to those not imprisoned in
abuse are virtually un-interpretable unless clear opera- the previous year. These interview data can be used in har-
tional definitions are used. This is especially the case with mony with the key informant data to give greater strength
respect to defining the drug type covered. It was not a in establishing the ratios.

70
 Annex II.
Supplementary figure and tables
Figure A.II.I. Severity of dependence scores—by sample group

(a) Treatment sample

20
18
16
14
Per cent

12
10
8
6
4
2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
SDS

(b) Street sample

20
18
16
14
12
Per cent

10
8
6
4
2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
SDS

(c) Prison sample

20
18
16
14
12
Per cent

10
8
6
4
2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
SDS

71
D R U G A B U S E I N PA K I S TA N

Table A.II.1. Respondents perceptions to the scale of drug use in their locale:
rural/urban comparisons

Urban areas Rural areas

Commonly Some Rarely Don't Commonly Some Rarely Don't
Drug type used use used know used use used know

Hashish and charas 127 (71%) 43 (24%) 7 (4%) 2 (1%) 67 (69%) 24 (25%) 6 (6%) —
Heroin 96 (54%) 69 (39%) 10 (6%) 4 (2%) 31 (32%) 42 (44%) 11 (11%) 12 (12%)
Opium 18 (10%) 58 (33%) 83 (47%) 18 (10%) 10 (11%) 33 (35%) 38 (40%) 13 (14%)
Other opiates 21 (9%) 28 (16%) 52 (30%) 74 (42%) 3 (3%) 6 (7%) 23 (26%) 56 (64%)
Cough syrups 19 (1%) 22 (12%) 56 (32%) 80 (45%) 12 (14%) 2 (2%) 17 (20%) 56 (64%)
Psychotropics 44 (25%) 62 (35%) 47 (26%) 25 (14%) 10 (11%) 26 (28%) 33 (36%) 23 (25%)
Drug injection 23 (13%) 56 (32%) 42 (24%) 52 (30%) 7 (8%) 17 (19%) 21 (24%) 43 (49%)
Solvents (glue, etc) 6 (4%) 30 (18%) 44 (26%) 90 (53%) 4 (5%) 6 (7%) 17 (20%) 57 (68%)
Alcohol 84 (47%) 66 (37%) 26 (15%) 1 (>1%) 37 (40%) 26 (28%) 28 (30%) 2 (2%)

Note: Sub-sample N's = Rural areas (84–97), Urban areas (170–179).

Table A.II.2. Respondents perceptions to the scale of drug use in their locale:
provincial comparisons

Punjab Sindh
Commonly Some Rarely Don't Commonly Some Rarely Don't
Drug type used use used know used use used know

Hashish and charas 62 (67%) 26 (28%) 5 (5%) — 36 (56%) 25 (39%) 2 (3%) 1 (2%)
Heroin 53 (58%) 31 (34%) 4 (4%) 4 (4%) 36 (56%) 24 (38%) 3 (5%) 1 (2%)
Opium 11 (12%) 33 (37%) 35 (39%) 10 (11%) 5 (8%) 15 (24%) 35 (56%) 8 (13%)
Other opiates 19 (22%) 19 (22%) 22 (26%) 26 (30%) 3 (5%) 8 (13%) 22 (34%) 31 (48%)
Cough syrups 24 (28%) 15 (17%) 27 (31%) 21 (24%) 1 (2%) 3 (5%) 19 (30%) 41 (64%)
Psychotropics 25 (28%) 38 (43%) 18 (20%) 8 (9%) 14 (22%) 21 (33%) 22 (34%) 7 (11%)
Drug injection 19 (21%) 41 (46%) 21 (23%) 9 (10%) 8 (13%) 15 (24%) 21 (33%) 19 (30%)
Solvents (glue, etc) 9 (11%) 25 (29%) 26 (31%) 25 (29%) 1 (2%) 7 (11%) 19 (31%) 35 (56%)
Alcohol 62 (71%) 22 (25%) 3 (3%) — 26 (41%) 26 (41%) 12 (19%) —

Balochistan NWFP
Commonly Some Rarely Don't Commonly Some Rarely Don't
Drug type used use used know used use used know

Hashish and charas 27 (90%) 2 (7%) — 1 (3%) 56 (75%) 13 (17%) 6 (8%) —

Heroin 23 (77%) 7 (23%) — — 9 (12%) 43 (57%) 12 (16%) 11 (15%)
Opium 7 (23%) 13 (43%) 8 (27%) 2 (7%) 4 (5%) 27 (36%) 37 (49%) 7 (9%)
Other opiates — 2 (7%) 7 (25%) 19 (68%) 2 (3%) 4 (6%) 18 (25%) 48 (67%)
Cough syrups — 2 (7%) 8 (30%) 17 (63%) 5 (7%) 4 (5%) 18 (25%) 46 (63%)
Psychotropics 1 (4%) 7 (25%) 9 (32%) 11 (39%) 11 (15%) 21 (28%) 23 (31%) 20 (27%)
Drug injection 3 (10%) 13 (45%) 6 (21%) 7 (24%) — 3 (4%) 12 (17%) 54 (78%)
Solvents (glue, etc) — 3 (11%) 15 (54%) 10 (36%) — 1 (1%) 1 (1%) 68 (97%)
Alcohol 18 (60%) 4 (13%) 6 (20%) 2 (7%) 14 (19%) 33 (44%) 27 (36%) 1 (1%)

Note: Sub-sample N's = Punjab (85–93), Sindh (62–64), Balochistan (27–30), NWFP (69–75).

72
 Table A.II.3. Key informant perceptions of long-term trends in drug use in their locale (last 5 years):
rural/urban comparisons

Urban areas Rural areas

Decreased Decreased No Increased Increased Don’t Decreased Decreased No Increased Increased Don’t
Drug type a lot a little change a little a lot know a lot a little change a little a lot know

Hashish and charas 4 (2%) 21 (12%) 24 (13%) 64 (36%) 61 (34%) 4 (2%) 3 (3%) 11 (11%) 9 (9%) 33 (34%) 41 (42%) —
Heroin 13 (7%) 42 (24%) 17 (10%) 50 (28%) 52 (29%) 4 (2%) 6 (6%) 15 (16%) 11 (12%) 26 (28%) 26 (28%) 9 (10%)
Opium 12 (7%) 21 (12%) 86 (50%) 23 (13%) 9 (5%) 20 (12%) 3 (3%) 16 (17%) 46 (49%) 13 (14%) 10 (11%) 5 (5%)
Other opiates 3 (2%) 7 (4%) 33 (20%) 32 (20%) 16 (10%) 72 (44%) 4 (5%) 1 (1%) 10 (12%) 9 (11%) 3 (4%) 55 (67%)
Cough syrups 6 (4%) 13 (8%) 19 (12%) 27 (16%) 18 (11%) 81 (49%) 1 (1%) 3 (4%) 8 (10%) 7 (9%) 8 (10%) 52 (66%)
Psychotropics 4 (2%) 6 (3%) 23 (13%) 70 (40%) 47 (27%) 23 (13%) — 6 (7%) 8 (9%) 30 (35%) 21 (24%) 21 (24%)
Drug injection — 9 (5%) 23 (13%) 52 (30%) 30 (18%) 57 (33%) 3 (4%) 2 (2%) 8 (10%) 23 (28%) 7 (9%) 38 (47%)
Solvents 2 (1%) 4 (2%) 22 (14%) 42 (26%) 10 (6%) 83 (51%) — 1 (1%) 8 (11%) 17 (23%) 1 (1%) 46 (63%)
Alcohol 2 (1%) 15 (9%) 54 (31%) 47 (27%) 52 (30%) 6 (3%) 5 (5%) 11 (12%) 17 (18%) 37 (39%) 21 (22%) 3 (3%)
A N N E X I I . S U P L E M E N TA RY F I G U R E A N D TA B L E S

73
74
Table A.II.4. Key informant perceptions of long-term trends in drug use in their locale (last 5 years):
provincial comparisons

Punjab Sindh
Decreased Decreased No Increased Increased Don’t Decreased Decreased No Increased Increased Don’t
Drug type a lot a little change a little a lot know a lot a little change a little a lot know

Hashish and charas 2 (2%) 7 (8%) 11 (13%) 46 (55%) 15 (18%) 2 (2%) 2 (3%) 10 (16%) 10 (16%) 21 (33%) 19 (30%) 1 (2%)
D R U G A B U S E I N PA K I S TA N

Heroin 6 (7%) 27 (33%) 15 (18%) 29 (35%) 5 (6%) 1 (1%) 7 (11%) — 7 (11%) 19 (30%) 29 (46%) 1 (2%)
Opium 2 (3%) 7 (9%) 51 (66%) 9 (12%) 1 (1%) 7 (9%) 3 (5%) 9 (15%) 34 (57%) 7 (12%) 1 (2%) 6 (10%)
Other opiates 2 (3%) 2 (3%) 12 (16%) 20 (26%) 15 (19%) 26 (34%) 1 (2%) 4 (7%) 14 (25%) 7 (12%) 2 (4%) 29 (51%)
Cough syrups 1 (1%) 4 (5%) 7 (9%) 21 (28%) 19 (25%) 23 (31%) 5 (9%) 5 (9%) 4 (7%) 3 (5%) 3 (5%) 35 (64%)
Psychotropics 2 (3%) 1 (1%) 6 (8%) 44 (56%) 21 (27%) 4 (5%) 1 (2%) 4 (7%) 7 (12%) 20 (34%) 20 (34%) 7 (12%)
Drug injection — 6 (7%) 14 (17%) 37 (45%) 17 (20%) 9 (11%) 2 (3%) 1 (2%) 4 (7%) 16 (27%) 11 (19%) 25 (42%)
Solvents — 3 (4%) 13 (17%) 42 (55%) 6 (8%) 13 (17%) 1 (2%) 2 (4%) 8 (14%) 7 (13%) 4 (7%) 34 (61%)
Alcohol — 2 (3%) 30 (38%) 24 (31%) 20 (26%) 2 (3%) 2 (3%) 3 (5%) 13 (21%) 21 (33%) 23 (37%) 1 (2%)

Balochistan NWFP
Decreased Decreased No Increased Increased Don’t Decreased Decreased No Increased Increased Don’t
Drug type a lot a little change a little a lot know a lot a little change a little a lot know

Hashish and charas — — 3 (10%) 9 (30%) 17 (57%) 1 (3%) 2 (2%) 11 (13%) 7 (8%) 18 (21%) 47 (55%) —
Heroin — — — 5 (17%) 24 (80%) 1 (3%) 11 (13%) 20 (24%) 6 (7%) 21 (25%) 17 (20%) 10 12%)
Opium — 3 (10%) 8 (27%) 3 (10%) 12 (40%) 4 (13%) 10 (12%) 16 (19%) 31 (36%) 16 (19%) 5 (6%) 7 (8%)
Other opiates — — 8 (30%) 1 (4%) — 18 (67%) 4 (5%) 2 (3%) 9 (12%) 11 (14%) 1 (1%) 50 (65%)
Cough syrups — — 6 (21%) 6 (21%) — 17 (59%) 1 (1%) 6 (8%) 8 (10%) 3 (4%) 4 (5%) 55 (71%)
Psychotropics — 1 (3%) 5 (17%) 10 (34%) 3 (10%) 10 (34%) 1 (1%) 6 (7%) 12 (15%) 20 (24%) 22 (27%) 21 (26%)
Drug injection — — 3 (10%) 12 (41%) 8 (28%) 6 (21%) 1 (1%) 3 (4%) 9 (12%) 8 (11%) — 53 (73%)
Solvents — 1 (4%) 6 (22%) 10 (37%) — 10 (37%) — — 2 (3%) 1 (1%) — 67 (96%)
Alcohol — 2 (7%) 7 23%) 9 (30%) 8 (27%) 4 (13%) 3 (4%) 14 (16%) 19 (22%) 29 (34%) 18 (21%) 2 (2%)
 Table A.II.5. Key informant perceptions of short-term trends in drug use in their locale (last year): rural/urban comparisons

Urban areas Rural areas

Decreased Decreased No Increased Increased Don’t Decreased Decreased No Increased Increased Don’t
Drug type a lot a little change a little a lot know a lot a little change a little a lot know

Hashish and charas 6 (3%) 15 (8%) 39 (22%) 60 (34%) 55 31%) 4 (2%) 3 (3%) 9 (9%) 18 (19%) 30 (31%) 36 (38%) —
Heroin 14 (8%) 46 (26%) 33 (19%) 47 (26%) 35 (20%) 3 (2%) 8 (9%) 16 (17%) 17 (18%) 25 (27%) 18 (19%) 9 (10%)
Opium 10 (6%) 19 (11%) 96 (56%) 22 (13%) 4 (2%) 19 (11%) 4 (4%) 13 (14%) 47 (52%) 16 (18%) 4 (4%) 7 (8%)
Other opiates — 8 (5%) 36 (22%) 27 (17%) 17 (11%) 73 (45%) 3 (4%) 1 (1%) 15 (19%) 3 (4%) 3 (4%) 56 (69%)
Cough syrups 1 (1%) 10 (6%) 31 (19%) 24 (15%) 15 (9%) 79 (49%) 4 (5%) 1 (1%) 9 (12%) 7 (9%) 8 (10%) 49 (63%)
Psychotropics 1 (1%) 3 (2%) 27 (16%) 73 (43%) 39 (23%) 28 (16%) 1 (1%) 3 (3%) 15 (17%) 25 (29%) 22 (25%) 21 (24%)
Drug injection 1 (1%) 8 (5%) 24 (14%) 56 (33%) 24 (14%) 55 (33%) — 2 (3%) 14 (18%) 19 (24%) 3 (4%) 41 (52%)
Solvents 2 (1%) 3 (2%) 26 (16%) 35 (22%) 13 (8%) 81 (51%) — — 13 (16%) 13 (16%) 4 (5%) 51 (63%)
Alcohol 1 (1%) 11 (6%) 63 (36%) 51 (29%) 46 (26%) 4 (2%) 3 (3%) 12 (13%) 32 (34%) 26 (28%) 18 (19%) 3 (3%)
A N N E X I I . S U P L E M E N TA RY F I G U R E A N D TA B L E S

75
76
Table A.II.6. Key informant perceptions of short-term trends in drug use in their locale (last year): provincial comparisons

Punjab Sindh
Decreased Decreased No Increased Increased Don’t Decreased Decreased No Increased Increased Don’t
Drug type a lot a little change a little a lot know a lot a little change a little a lot know

Hashish and charas 4 (5%) 4 (5%) 16 (20%) 45 (55%) 10 (12%) 3 (4%) 1 (2%) 10 (16%) 14 (22%) 15 (23%) 23 (36%) 1 (2%)
D R U G A B U S E I N PA K I S TA N

Heroin 9 (11%) 23 (28%) 22 (27%) 22 (27%) 6 (7%) 1 (1%) 1 (2%) 14 (22%) 15 (23%) 15 (23%) 18 (28%) 1 (2%)
Opium — 7 (9%) 54 (70%) 9 (12%) — 7 (9%) 3 (5%) 5 (8%) 38 (63%) 5 (8%) 2 (3%) 7 (12%)
Other opiates — 4 (5%) 15 (19%) 18 (23%) 15 (19%) 25 (32%) 2 (3%) 1 (2%) 12 (21%) 6 (10%) 3 (5%) 34 (59%)
Cough syrups — 2 (3%) 13 (17%) 22 (29%) 18 (24%) 20 (27%) 3 (5%) 3 (5%) 8 (14%) 2 (4%) 1 (2%) 40 (70%)
Psychotropics — 1 (1%) 7 (9%) 41 (55%) 21 (28%) 5 (7%) — 1 (2%) 5 (8%) 29 (47%) 19 (31%) 8 (13%)
Drug injection 1 (1%) 3 (4%) 14 (17%) 39 (48%) 16 (20%) 8 (10%) — 1 (2%) 12 (19%) 12 (19%) 9 (15%) 28 (45%)
Solvents — 2 (3%) 12 (15%) 40 (51%) 12 (15%) 13 (16%) 1 (2%) 1 (2%) 12 (21%) 5 (9%) 4 (7%) 34 (60%)
Alcohol 1 (1%) 1 (1%) 29 (37%) 29 (37%) 17 (22%) 1 (1%) — 5 (8%) 14 (22%) 20 (31%) 24 (38%) 1 (2%)

Balochistan NWFP
Decreased Decreased No Increased Increased Don’t Decreased Decreased No Increased Increased Don’t
Drug type a lot a little change a little a lot know a lot a little change a little a lot know

Hashish and charas — — 13 (43%) 9 (30%) 8 (27%) — 3 (4%) 8 (9%) 10 (12%) 20 (24%) 44 (52%) —
Heroin — — 4 (14%) 12 (41%) 13 (45%) — 10 (12%) 22 (26%) 9 (11%) 20 (24%) 14 (16%) 10 (12%)
Opium — 3 (11%) 10 (36%) 9 (32%) 3 (11%) 3 (11%) 11 (13%) 15 (18%) 35 (42%) 12 (14%) 3 (4%) 8 (10%)
Other Opiates — — 9 (35%) — — 17 (65%) 1 (1%) 4 (5%) 14 (19%) 4 (5%) 2 (3%) 50 (67%)
Cough syrups — 1 (4%) 11 (39%) 1 (4%) — 15 (54%) 2 (3%) 4 (6%) 7 (10%) 4 (6%) 4 (6%) 51 (71%)
Psychotropics — 2 (7%) 10 (36%) 4 (14%) 2 (7%) 10 (36%) 1 (1%) 3 (4%) 17 (21%) 20 (24%) 17 (21%) 24 (29%)
Drug injection — — 5 (15%) 16 (62%) 1 (4%) 5 (19%) — 5 (7%) 8 (11%) 5 (7%) — 54 (75%)
Solvents 1 (4%) 13 (48%) 3 (11%) — 10 (37%) — — 1 (1%) 1 (1%) — 71 (97%)
Alcohol — 2 (7%) 16 (55%) 2 (7%) 6 (21%) 3 (10%) 2 (2%) 11 (13%) 32 (38%) 24 (28%) 14 (16%) 2 (2%)
 Table A.II.7. Key informant perceptions of problems caused by drug use: urban/rural comparisons

Urban areas Rural areas

Major Some Few No Don’t Major Some Few No Don’t
Drug type problems problems problems problems know problems problems problems problems know
Hashish and charas 31 (17%) 86 (48%) 47 (26%) 15 (8%) — 15 (15%) 43 (44%) 26 (27%) 13 (13%) —
Heroin 156 (87%) 16 (9%) 3 (2%) 3 (2%) 1 (1%) 66 (71%) 12 (13%) 3 (3%) 1 (1%) 11 (12%)
Opium 11 (6%) 41 (24%) 61 (35%) 45 (26%) 15 (9%) 11 (12%) 15 (16%) 27 (29%) 36 (39%) 3 (3%)
Other opiates 14 (8%) 22 (13%) 38 (23%) 23 (14%) 70 (42%) — 9 (11%) 7 (9%) 16 (20%) 50 (61%)
Cough syrups 16 (10%) 20 (12%) 21 (13%) 25 (15%) 82 (50%) 6 (7%) 4 (5%) 5 (6%) 14 (17%) 52 (64%)
Psychotropics 19 (11%) 29 (17%) 57 (33%) 41 (24%) 26 (15%) 1 (1%) 11 (13%) 28 (32%) 23 (26%) 24 (28%)
Drug injection 45 (26%) 40 (23%) 21 (12%) 16 (9%) 49 (29%) 13 (16%) 15 (18%) 7 (9%) 8 (10%) 39 (48%)
Solvents 17 (10%) 20 (12%) 34 (21%) 19 (12%) 72 (44%) 3 (4%) 10 (13%) 11 (14%) 7 (9%) 46 (60%)
Alcohol 52 (30%) 60 (34%) 51 (29%) 11 (6%) 2 (1%) 32 (34%) 30 (32%) 20 (22%) 7 (8%) 4 (4%)
A N N E X I I . S U P L E M E N TA RY F I G U R E A N D TA B L E S

77
78
Table A.II.8. Key informant perceptions of problems caused by drug use: provincial comparisons

Punjab Sindh
Major Some Few No Don’t Major Some Few No Don’t
Drug type problems problems problems problems know problems problems problems problems know
Hashish and charas 17 (20%) 45 (54%) 19 (23%) 2 (2%) — 8 (13%) 20 (31%) 23 (36%) 13 (20%) —
Heroin 73 (88%) 9 (11%) — 1 (1%) — 50 (78%) 10 (16%) 2 (3%) 2 (3%) —
D R U G A B U S E I N PA K I S TA N

Opium 1 (1%) 24 (31%) 31 (40%) 18 (23%) 3 (4%) 3 (5%) 4 (7%) 13 (21%) 35 (57%) 6 (10%)
Other opiates 10 (13%) 19 (24%) 20 (25%) 7 (9%) 23 (29%) 3 (5%) 3 (5%) 7 (12%) 17 (30%) 27 (47%)
Cough syrups 19 (25%) 17 (22%) 9 (12%) 9 (12%) 23 (30%) 1 (2%) 2 (4%) 2 (4%) 19 (33%) 33 (58%)
Psychotropics 11 (14%) 17 (22%) 26 (34%) 20 (26%) 3 (4%) 4 (7%) 7 (12%) 20 (33%) 20 (33%) 9 (15%)
Drug injection 28 (34%) 32 (39%) 10 (12%) 7 (8%) 6 (7%) 10 (16%) 10 (16%) 10 (16%) 10 (16%) 21 (34%)
Solvents 18 (24%) 23 (30%) 22 (29%) 5 (7%) 8 (11%) 1 (2%) 3 (5%) 11 (19%) 15 (26%) 28 (48%)
Alcohol 37 (48%) 33 (43%) 6 (8%) — 1 (1%) 19 (30%) 18 (28%) 20 (31%) 6 (9%) 1 (2%)

Balochistan NWFP
Major Some Few No Don’t Major Some Few No Don’t
Drug type problems problems problems problems know problems problems problems problems know

Hashish and charas 6 (20%) 13 (43%) 4 (13%) 7 (23%) — 12 (14%) 46 (54%) 22 (26%) 5 (6%) —
Heroin 30 (100%) — — — — 61 (72%) 8 (9%) 4 (5%) 1 (1%) 11 (13%)
Opium 10 (33%) 3 (10%) 13 (43%) 2 (7%) 2 (7%) 8 (9%) 22 (26%) 28 (33%) 21 (25%) 6 (7%)
Other Opiates — 1 (3%) 8 (27%) 1 (3%) 20 (67%) 1 (1%) 8 (10%) 7 (9%) 14 (18%) 47 (61%)
Cough syrups — — 5 (17%) 3 (10%) 22 (73%) 2 (3%) 5 (7%) 7 (9%) 7 (9%) 54 (72%)
Psychotropics — 3 (10%) 8 (27%) 4 (13%) 15 (50%) 5 (6%) 12 (15%) 26 (32%) 17 (21%) 21 (26%)
Drug injection 16 (53%) 5 (17%) 2 (7%) 1 (3%) 6 (20%) 1 (1%) 7 (10%) 6 (8%) 6 (8%) 53 (73%)
Solvents — 2 (7%) 10 (36%) 2 (7%) 14 (50%) — 2 (3%) 2 (3%) 4 (6%) 63 (89%)
Alcohol 1 (3%) 7 (23%) 14 (47%) 6 (20%) 2 (7%) 23 (27%) 26 (31%) 29 (34%) 5 (6%) 2 (2%)
 A N N E X I I . S U P L E M E N TA RY F I G U R E A N D TA B L E S

Table A.II.9. Key informants: attitudes to drug abuse

Strongly Strongly Overall

Attitude statements disagree Disagree Unsure Agree agree rating

Many drug users are too

poor to get treatment. 1 (>1%) 27 (10%) 1 (>1%) 106 (38%) 141 (51%) 1.30
Drug users are responsible
for a lot of crimes in this area. 3 (1%) 53 (19%) 3 (1%) 141 (51%) 75 (27%) 1.32
Many treatment services in
this area do not offer a
good service to drug users. 8 (3%) 17 (6%) 23 (8%) 103 (37%) 124 (45%) 0.80
Heroin use is becoming
less popular than it used
to be in this area. 70 (25%) 77 (28%) 22 (8%) 90 (33%) 18 (6%) 0.44
There is a need in this area for
more educational campaigns
about the dangers of drug abuse. 1 (>1%) 4 (1%) — 60 (22%) 211 (76%) 0.33
The police and prison system
should be more severe in
dealing with drug abusers. 51 (18%) 99 (36%) 2 (>1%) 59 (21%) 65 (24%) 1.01
Awareness campaigns against
drug abuse have been an
effective way of discouraging
drug abuse. 57 (21%) 70 (25%) 26 (9%) 80 (29%) 42 (15%) 0.09
There is not sufficient help
available for people with
drug problems in this area. 11 (4%) 10 (4%) 1 (>1%) 99 (36%) 155 (56%) 0.55
Heroin users are increasingly
changing to other pharmaceutical
drugs in this locale. 10 (4%) 31 (11%) 88 (32%) 88 (32%) 59 (21%) 1.26

Note: sub-sample N's (273–276). Note overall rating score positive values = attitude positive to statement/negative values
indicate negative attitude. Possible values run from –2 (all strongly disagree) +2 (all strongly agree).

79
Annex III.
Tables supplemental to prevalence calculation

1. Notes on calculation methods used in the

prevalence estimation tables

CALCULATION SHEET SUMMARY (1)

Overall calculation and presentation scheme

The locales for the four cities study component and the other study locales are estimated separately.

The overall estimation scheme has been to estimate the prevalence of addiction in each of the four
provinces of Sindh, Punjab, NWFP and Balochistan separately, as far as possible, in order to derive the
prevalence for Pakistan as a whole. The F.A.T.A. has not been included in the prevalence estimate as a
separate figure.

Within each of these provinces, prevalence has been estimated (i) for the principal locale that contributed
to the four cities study component of the National Assessment, namely Karachi, Lahore, Peshawar and
Quetta respectively; and (ii) within each province also for the remaining locales chosen for study.

The tables therefore present data for each of the four cities locales and for the totals of 'all other' locales
within each province—giving eight 'locale-groupings'—and for all study locales combined.

Data structures and estimate breakdowns used:

1. Locale demographic Profiles
Basic data provided by the Census is used, using figures for each of the locales in the study

Section 1 of the table establishes the key figure of the number of males in the locale-groupings, derived
from the figures provided in the 1998 National Census, Census Bulletin-1, along with other demographic
characteristics.

Data structures presented

Households
Male
Female
Total
Urban males
1981 Population Total

2. Target age-band (in millions of population)

This next step reduces the overall male population to target the most important age band of 15 to 45 years

Section 2 of the table presents the estimated percentage of the male population that lies in the target age-
band of 15 to 44 years of age. The initial figures were derived from the 1998 National Census Bulletin 6.

81
D R U G A B U S E I N PA K I S TA N

(Subsequently figures were derived separately for each in each of the locale-groupings. Key informants, excluding
province, from Census Bulletins 7-10.) This percentage is used police and treatment-related personnel, who had contact
in conjunction with the section 1 figure for the toal male popu- with at least ten addicts over the previous year reported
lation to provide the number of males in the target age-band, on how many of these had received inpatient treatment
for each locale-grouping. during the year.

Data structures presented

These reports from all included Key Informants in the
Percentage of males in target age-band locale-groupings gave a median figure for the percentage
Number of males in target population of addicts receiving inpatient treatment over the previous
year. Amongst these reports the 25%-tile (1st quartile) is
3. Locale clinic profiles used as a minimum likely estimate and the 75%-tile is
used as a maximum likely estimate to provide a likely range
Data from the treatment Register Update Study are used to estimate
for the estimate of the percentage of addicts who were
the number of male heroin addicts in treatment
treated.
Section 3 of the table presents information collected in the
Treatment Register Update component of the National These treatment percentages enable the calculation of a
Assessment. This information gives the key figures of (i) the range of 'treatment multipliers'—factors that give the
total numbers of inpatient cases in the past year who have required inflation factor of local inpatient treatment
attended the listed specialist (inpatient) clinics; and (ii) the figures to give the total number of local treated and un-
percentage of inpatient cases who were male heroin addicts, treated addicts. These inflation multipliers are applicable
along with other information on the listed clinic services. to the estimated number of local addicts receiving in-
patient treatment for each (of 7) locale-grouping.
Combining these figures allows the calculation of the num-
ber of male inpatient heroin addiction cases, although only
The Treatment Register Update interviews in the clinics
in 7 of the locale-groupings, since the locales in Balochistan
provided informally data on the percentage of treated
other than Quetta had no recorded inpatient services. So in
addicts at the listed specialist clinics who are from the
Balochistan only the Quetta city locale could provide a
local catchment areas, and this catchment 'correction
figure for the number of male inpatient heroin addicts
factor' has been taken at 90% in the initial estimation
treated in the past year.
procedures.
Data structures presented
Total number of beds in clinics Combining, for each of the locale groupings separately,
Total treatment admissions the figures for the multiplier and the catchment correction
Total number of inpatients in year factor with the numbers of inpatient treated addicts
Percentage male inpatient Heroin cases derived in section 3 of the Summary Sheet gives the
Number of male inpatient Heroin cases estimated total number of heroin addicts in the study
locales.

CALCULATION SHEET SUMMARY (2)

Data structures presented

The second Calculation Summary Sheet extends the esti- (a) Maximum likely % in treatment
mates made in Summary Sheet 1 of the numbers of (b) Median likely % in treatment
treated heroin addicts to give estimates of the likely range (c) Minimum likely % in treatment
of total numbers of heroin addicts for each locale- (a) Derived multiplier (for catchment)
grouping, and for Pakistan as a whole.
(b) Derived multiplier (for catchment)
(c) Derived multiplier (for catchment)
4. Key Informant and Interview Treatment Multipliers
The range of multipliers used to represent the number of addicts not
catchment correction
in treatment in the year is a composite of all study data
(a) Estimated number of addicts in catchment
The Key Informant component of the National Assessment (b) Estimated number of addicts in catchment
provided information on the percentage of addicts treated (c) Estimated number of addicts in catchment

82
 A N N E X I I I . TA B L E S S U P P L E M E N TA L T O P R E VA L E N C E C A L C U L AT I O N

5. Prevalence estimates and numbers of addicts Minimum likely and maximum likely estimates are given
(extrapolated) in each instance, to provide a plausible range of estimates.
Minimum and maximum likely numbers extrapolated to whole
Data structures presented
country
Ext. coverage (non-clinic locales)
The final section 5 extrapolates from the estimated Ext. coverage (province exc locales)
number of addicts estimated to be in each locale's clinic Minimium likely ...
catchment area to the overall figure for each of the provinces (a) Target prevalence (in catchment)
and for Pakistan in total. This extrapolation is in two stages: (a) Number of addicts (clinic locales)
(i) from the locales in the study that have clinics listed for
(a) Likely minimum number of heroin addicts
male heroin inpatient treatment to all locales used in
Maximium likely ...
the study; (ii) from the locales used in the study to the
whole province, keeping the distinction between the locale (c) Target prevalence (in catchment)
for the principal city and the remaining locales in each (c) Number of addicts (clinic locales)
province. (c) Likely maximum number of heroin addicts

2. Tables showing prevalence estimation data by province

Tables correspond to tables 25 to 29 in chapter 8, giving data outside the principal cities of the four cities study for
each province.

Table A.III.1. Basic data provided by the census, using figures for each of the locales
in the study

Locale data (exc four cities) summed for each province

Total of
other locales Sindh Punjab NWFP FATA Balochistan

(i) Locale Demographic Profiles

Households 3 305 785 625 540 2 112 993 470 016 97 236 0
Male 12 104 131 1 947 639 7 484 506 2 062 258 454 492 155 236
Female 11 228 122 1 770 872 6 979 950 1 924 769 411 487 141 044
Total 23 332 253 3 718 511 14 464 456 3 987 027 865 979 296 280
Urban males 4 522 729 997 809 3 191 978 268 185 28 587 36 170
1998 population total 11 389 758 2 612 874 7 334 489 994 206 448 189 0

Table A.III.2. Reduced overall target population to males in the age band of 15 to 45 years

Locale data (exc four cities) summed for each province

Total of
other locales Sindh Punjab NWFP FATA Balochistan

(ii) Locale Target Age-Band:

males aged 15-45 (in millions
of population)
Assumed % of males in target 42.20% 42.20% 42.20% 42.20% 42.20% 42.20%
Number of males in target population 5.108 0.822 3.158 0.870 0.192 0.066

83
D R U G A B U S E I N PA K I S TA N

Table A.III.3. Data from the treatment register update study

Locale data (exc four cities) summed for each province

Total of
other locales Sindh Punjab NWFP FATA Balochistan

(iii) Locale Clinic Profiles

Total number of beds provided
in specialist clinics 522 243 269 10 0 0
Total number of treatment admissions 3 589 1 316 2 118 155 0 0
Total number of inpatients
admitted in year 2 151 603 1 393 155 0 0
Proportion of admissions that are
male inpatient heroin cases 0.89 0.83 0.58
Number of admissions of
male inpatient heroin cases 1 783 537 1156 90

Table A.III.4. Treatment multipliers derived from key informant and addict interview data

Locale data (exc four cities) summed for each province

Total of
other locales Sindh Punjab NWFP FATA Balochistan

(iv) Key Informant and Interview

Treatment Multipliers
Range of estimated proportion of
addicts who received treatment
in year (see text)
(a) Maximum likely proportion
who received treatment 0.2567 0.2000 0.3000 0.4167 —
(b) Median likely proportion
who received treatment 0.1464 0.1000 0.1000 0.2000 —
(c) Minimum likely proportion
who received treatment 0.0600 0.0300 0.0500 0.1500 —
0.2567 0.2000 0.3000 0.4167 —

Range of estimated derived multiplier,

representing addicts not receiving
treatment in year
(a) Minimum multiplier
(applies to clinics' catchment area) 3.90 5.00 3.33 2.40 —
(b) Median multiplier
(applies to clinics' catchment area) 6.83 10.00 10.00 5.00 —
(c) Maximum multiplier
(applies to clinics' catchment area) 16.67 33.33 20.00 6.67 —

Range of estimated number of addicts

in population (see text)
Correction factor for addicts
out-of-catchment (see text) 0.90 0.90 0.90
(a) Likely estimated minimum
number of addicts in catchment 6 078 2 415 3 469 194
(b) Likely estimated median
number of addicts in catchment 15 640 4 830 10 406 405
(c) Likely estimated maximum
number of addicts in catchment 37 451 16 100 20 811 539

84
 A N N E X I I I . TA B L E S S U P P L E M E N TA L T O P R E VA L E N C E C A L C U L AT I O N

Table A.III.5. Extrapolated estimates of the numbers of addicts in the country as a whole

Locale data (exc four cities) summed for each province

Total of
other locales Sindh Punjab NWFP FATA Balochistan

(v) Prevalence estimates and numbers

of addicts (extrapolated)
Minum and maximum likely numbers
extrapolated to whole country
ext.coverage (non-clinic locales) 1.00 1.59 2.76 — —
ext.coverage (province exc locales) 5.42 4.63 3.83 — —
(a) Minimium likely ...
(a) target prevalence (in catchment) 0.12% 0.29% 0.17% 0.06% —
(a) Number of addicts (clinic locales) 6 078 2 415 3 469 194 —
(a) Likely minimum number
of heroin addicts 40 721 13 096 25 570 2 055
(b) Maximium likely …
(c) target prevalence (in catchment) 0.73% 1.96% 1.05% 0.17% —
(c) Number of addicts (clinic locales) 37 451 16 100 20 811 539 —
(c) Likely maximum number
of heroin addicts 246 434 87 305 153 421 5 707

85
Printed in Austria
V.02-55763–October 2002–1,100