Mathers, Bradley M MBChB*; Degenhardt, Louisa PhD*; Adam, Philippe PhD‡§; Toskin, Igor MD, PhD†; Nashkhoev, Magomed MD, PhD†; Lyerla, Rob PhD†; Rugg, Deborah PhD†
By and large, outside of sub-Saharan Africa, HIV epidemics are concentrated among certain subsets of the general population that are especially vulnerable to HIV infection. In particular, injecting drug use accounts for an increasing number of new HIV infections in several parts of the world.1 A recent systematic review by the Reference Group to the United Nations (UN) on HIV and Injecting Drug Use found that injecting drug use has been reported to occur in 148 countries around the world. HIV infection was reported to occur among injecting drug user (IDU) populations in 119 of 127 countries where this had been examined. Prevalence of HIV ranged from zero to as high as 90% in some urban samples of IDUs.2
Of the 147 Member States categorized as low-income and middle-income countries (LMICs) in 2007 by the World Bank,3 injecting drug use was reported to occur in 99.2 It was estimated that in 2007, there were a total of 11.8 million IDUs (range from 8 to 16 million) in LMICs.2
The size of this at-risk population is therefore substantial. It is crucial that the national HIV response in countries where injecting drug use occurs includes strategies that address HIV among injectors.
People who inject drugs are at risk of HIV infection both through unsafe drug injecting and unprotected sex with an HIV-infected person. Provision of sterile injecting equipment and condoms to people who inject are recognized as important components of a comprehensive strategy to prevent the transmission of HIV among IDUs and their sexual partners.4
Despite this, HIV prevention, in general and especially among most-at-risk groups, has been “marginalized” in the response to HIV,5 and behavioral approaches have not been pursued as diligently as antiretroviral therapy programs.6,7
In the Declaration of Commitment made at the 2001 United Nations General Assembly Special Session on HIV/AIDS4 (UNGASS), people who inject drugs were recognized to be one of the priority most-at-risk populations for HIV prevention interventions. In that Declaration, Member States made commitments to ensure expanded access to HIV prevention and care services for IDUs and to regularly report on progress.
Monitoring and evaluation of HIV prevention is a challenging but vital element of HIV programming.7,8 Most-at-risk populations, such as people who inject drugs, tend to be hidden or hard to reach. This makes data collection difficult and requires the use of specific epidemiological methods. In the 2006 UNGASS reporting round, the number of LMICs reporting on indicators related to IDU was low (n = 31, of 99 LMIC where injecting is reported).
The UNGASS indicators are intended to function as a set of key indices by which countries' progress can be assessed and compared, and with which a global assessment of progress can be made. The indicators were developed by the Joint United Nations Programme on HIV/AIDS (UNAIDS) Monitoring and Evaluation Reference Group and countries biennially report against them.
These indicators have evolved during the course of the previous 3 UNGASS reporting rounds and efforts have been made to improve their relevance and the quality of reporting.9 In addition, the number of countries providing data has increased in later reporting rounds.
After revision in 2007 of the UNGASS indicators related to most-at-risk populations, there is the potential that the data from the 2008 UNGASS reporting round could be used as a baseline against which data from future rounds can be compared. This would allow for improved assessment of progress in efforts to prevent the transmission of HIV among injectors. This article examines the data reported by LMIC in the 2008 UNGASS reporting round and estimates coverage of key HIV services across countries and estimated IDU populations.
We reviewed data submitted by LMICs to UNAIDS in January 2008 as part of the process established to monitor progress toward the targets outlined in the Declaration of Commitment made at the 2001 UNGASS4. Countries were classified as LMIC using the World Bank classification for 20073; a list of the categorized countries is presented in Appendix 1.
In the 2008 UNGASS reporting round, countries were requested to report on 5 core indicators measuring the provision and impact of HIV services for people who inject drugs; these indicators are detailed in Table 1.
Each indicator reported by a country was assessed to determine whether or not it was consistent with the indicator as defined by the UNAIDS guidelines.9 Indicators were judged to be inconsistent if the data provided were gathered from program monitoring rather than the required survey methodology as outlined in the UNAIDS guidelines9 (behavioral surveillance survey or an equivalent method). Data were also excluded if questions posed to survey participants differed substantially from those in the indicator guidelines.9 For example, in the case of the indicator assessing HIV testing, a question asking “have you ever been tested for HIV?” rather than the required question “have you been tested for HIV in the last 12 months?” was considered as being inconsistent-and those data were excluded. Similarly, the indicators on prevention programs and HIV prevention knowledge comprised a number of subcomponent questions (Table 1). If any subcomponent questions were inconsistent then the overall indicator was determined to be inconsistent. Inconsistent data were excluded from our analysis. Data were also excluded if samples contained fewer than 100 participants.
SPSS version 15.010 was used to produce descriptive statistics on these data. To provide some measure of the uncertainty around each indicator value, we calculated 95% confidence intervals based on the standard error of the percentage. For each indicator, the mean of the harmonized indicator data was determined and the mean weighted by the estimated size of the IDU population in each country was also calculated.
Estimates of the size of IDU populations in each country were drawn from the Reference Group to the UN on HIV and Injecting Drug Use.2 We used the country-level 2007 mid-point estimates of the numbers of people, aged between 15 and 64 years, who were estimated to have injected in the last 12 months. Where no direct estimate was available for a particular country, the population-weighted regional prevalence reported by the Reference Group was applied, using UN Population Division estimate of the country's population aged 15-64 years,11 to derive an estimate of the number of people who inject drugs in that country. It should be noted that such derived estimates of IDU population size have been used here for the purpose of “weighting” the estimated levels of various indicator data only. They should be regarded as approximate estimates of IDU population size and should not be used in preference to national level estimates derived from country-level studies.
Using the statistical software program Stata version 9.2,12 random effects meta-analyses were performed to determine 95% confidence intervals around the unweighted mean for each indicator. This allowed us to generate uncertainty bounds around the unweighted mean, accounting for the size of the different sample populations.
In addition to global weighted estimates, regional estimates for each indicator were weighted by estimated IDU population size in cases where reports from 2 or more countries were available for a region.
Forty LMIC reported on at least 1 of the 5 indicators examined here. There are an additional 59 LMIC that did not report on any of the IDU-related indicators in this reporting round but where injecting drug use has been reported to occur.2
Only 32 countries reported data on any of the 5 indicators that met the inclusion criteria for this current analysis-that is to say, data reported for indicators that are consistent with the UNAIDS indicator guidelines and that are drawn from samples of more than 100 participants. Those 32 countries account for 68% (8.07 million) of the estimated 11.86 million injectors in LMICs.2
Tables 2-6 present the data that were reported by countries and summary data for each indicator.
Eleven countries (Armenia, Bangladesh, Belarus, Bulgaria, China, Georgia, Latvia, Republic of Moldova, Nepal, Russian Federation and the Ukraine) reported data that met the inclusion criteria for all 5 indicators. Countries were more likely to report on some indicators than others: 38 reported on HIV testing (27 met the inclusion criteria); 26 reported on HIV prevention programs (15 met inclusion criteria); 32 reported on HIV prevention knowledge (19 met inclusion criteria); 34 reported on condom use (29 met inclusion criteria); and 34 reported on injecting practices (29 met inclusion criteria).
The size of samples from which countries drew data varied considerably across different indicators-from 159 to 21,490.
Twenty-seven countries (accounting for 59% of the estimated number of IDUs in LMICs) reported on HIV testing among injectors. In 6 of those 27 countries, more than 45% of the sampled injectors reported having undergone HIV testing in the previous year. The IDU population-weighted mean rate of testing among injectors in these 27 reporting countries was 36%. Regional IDU population estimates were calculated for the 6 regions, where data were available from 2 or more countries (Table 2). HIV testing among people who inject drugs was found to be lowest in the Middle East and North Africa region (2 countries, weighted mean of 9%) and highest in Latin America (2 countries, mean = 78%). Eastern Europe contained the largest number of reporting countries (10 countries), with a weighted regional mean of 50%.
HIV Prevention Programs
Reported levels of access to HIV prevention programs as measured by the UNGASS indicator varied considerably (5%-89%) across the 15 countries (representing 48% of all IDUs in LMICs) that reported data, which met the inclusion criteria of this review. Weighted by estimated IDU population size, the mean proportion in these 15 reporting countries of sampled injectors who knew where they could receive an HIV test and who had also received both condoms and sterile injecting equipment in the last 12 months was 26%. Only in 3 regions did 2 or more countries report data that met the inclusion criteria for this indicator (Table 3): Eastern Europe (8 countries) and East and South East Asia (3 countries) had similar weighted means (28% and 26%, respectively), whereas South Asia (2 countries) had a much higher mean of 80% for this indicator.
HIV Prevention Knowledge
Nineteen countries (accounting for 53% of IDUs from LMICs) reported data meeting the inclusion criteria for the indicator on HIV prevention knowledge among IDUs (Table 4). In 6 of these 19 countries, more than 50% of sampled injectors were able to correctly answer questions related to HIV prevention. South Asia (3 countries) had the lowest calculated regional IDU population-weighted mean (28%) and Central Asia (2 countries) the highest (63%).
Twenty-nine countries (containing 65% of the estimated number of IDUs in LMICs) reported on condom use among IDUs at the time of last sexual intercourse, with an overall IDU population-weighted mean of 37% (Table 5). Regional weighted estimates ranged from 11% for the Middle East and North Africa (2 countries) to 93% for Latin America (2 countries).
Safe Injecting Practices
Twenty-nine countries reported on the use of sterile injecting equipment at last injecting, with an IDU population-weighted mean of 63% (Table 6). The highest regional IDU population weighted mean was observed for Eastern Europe (11 countries; mean of 77%) and the lowest in the Middle East and North Africa (2 countries; mean 10%).
HIV prevention interventions must target and successfully reach most-at-risk populations such as people who inject drugs, especially in countries where epidemics are being fuelled by the sharing of contaminated injecting equipment.
The data provided in the 2008 UNGASS reporting round suggest that for each of the 5 IDU-related indicators, the LMICs reporting data represent approximately 50% or more of the total population of injectors estimated to live in all LMICs. For the 2 indicators assessing HIV prevention behaviors, this proportion approaches two-thirds (65% for the condom use indicator and 63% for the safe injecting practices indicator).
The overall level of coverage reported for the indicator assessing the reach of HIV prevention programs for IDU was especially low, with approximately 25% (IDU population-weighted mean) of IDUs reporting that they knew where to get tested for HIV and had received condoms and syringes in the last 12 months. This is far below what is considered universal access to HIV prevention services. However, it is important to note that harm reduction programs for injectors have been initiated relatively recently in many of these LMICs.
Levels of HIV testing in the last 12 months were higher than levels of access to HIV prevention services. HIV testing can serve as a gateway for access to antiretroviral and other treatment. It is also an HIV prevention education opportunity and an access point for other HIV prevention services. The data collected, however, do not indicate whether or not testing was voluntary. It is possible that the higher levels of testing reported in this round reflect, to some extent, mandatory testing of subpopulations of IDUs (such as those who are incarcerated or who attend health care facilities). Encouragingly, reports from the LMICs that did provide data indicate that a majority of IDUs report using a sterile syringe at the last injection. However, the behaviors of IDUs in countries that did not report data (home to an estimated 37% of people who inject drugs in LMICs) are not known.
There are limitations to the indicator data reported by countries and to the analysis of these data described here.
Because so few countries reported consistently across all the indicators we examined here, it was not possible to deduce regional trends.
Very little detailed information is available on the methods used to collect the data that countries submit. In particular, in the absence of information regarding how and where samples were recruited, it is difficult to interpret many of these indicators. The samples investigated may not be representative of the total population of injectors in a country. Sampling of sentinel populations that are commonly accessed via service sites introduces a likely bias when measuring service coverage, that is, availability of HIV prevention services tends to be higher than in the rest of the country. Thus, there is potential risk of overestimating levels of service coverage. In addition, it is possible that many countries may report data from samples gathered in a single or limited number of geographic locations. Few countries provided information on where samples were drawn from. But, among those that did provide such information, sampling invariably was reported to have occurred in large cities only. In the Russian Federation, for example, samples were gathered from 3 major cities only, whereas it is known that injecting occurs in other locations and the levels of services provided is inconsistent across these different settings. It is therefore possible that the heterogeneity of IDUs, of their behavior and of their access to services within a country may not be reflected in these data.
The reporting process itself is also subject to inaccuracy if countries make errors or omissions in the reports they submit. All data submitted are reviewed and checked for irregularities, but this does not exclude the possibility of incorrect data going undetected as the reviewers do not have access to original source material or information.
Furthermore, a lack of comparable data from other sources prevents verification of the data on these indicators reported by countries. An extensive search of the peer-reviewed and grey literature yielded relatively few data that were similar enough to allow comparison with the UNGASS reported data. There seemed to be a substantial amount of other data being collected on the reach of HIV prevention programs and HIV prevention behaviors of people who injected drugs. For the most part, however, these data were either from the very same source as submitted to UNAIDS (and thus identical) or, conversely, were too dissimilar to compare directly with the data from the UNGASS process.
In addition, the indicators in the 2008 UNGASS reporting round have some inherent limitations that are important to recognize. For example, the indicator assessing the reach of HIV prevention programs does not measure the quality of services provided. This indicator only measures whether or not an injector has accessed a service at least once in the last 12 months (but with no qualification of the nature of that contact).
The indicator definitions are necessarily strict to allow for comparison of data from different countries. However, this also leads to data from many countries being excluded from the analysis. This may reflect difficulties countries face when reporting against these indicators, or it may be that the UNGASS indicators are not regarded as appropriate measures in particular country contexts.
The situation in a large number of countries is not reported, however, it is important that efforts are made to gain understanding of the status of HIV prevention in those countries and to identify factors that might impede data collection or reporting.
Although the data reported for many LMIC countries suggest that many IDUs do have an understanding of how to prevent HIV transmission and that some are using condoms and, to some extent, inject with clean injecting equipment, it is clear that many remain at risk of sexual and injecting-related transmission of HIV. In addition, the goal of universal access to HIV prevention programs for IDUs is far from being achieved in LMICs. Given the role of IDUs in contributing to and in some regions, driving the HIV epidemics, failure to adequately address HIV transmission among IDUs is a serious barrier to achieve the targets of the Millennium Development Goal (MDG 6), which calls halting and reversing the spread of HIV/AIDS by 2015.
The UNGASS indicators are not intended to be a comprehensive set of measures for monitoring all aspects of countries' HIV prevention programs. Instead, they are proposed to provide an index with which countries' responses can be compared and to assist in monitoring progress of the overall global response. Despite these limitations and the difficulties inherent in monitoring and evaluating the current state of HIV prevention responses for people who inject drugs, the 2008 UNGASS reporting round does provide a useful baseline against which future progress might be measured. If in future rounds a greater number of countries report on these indicators and if there continues to be an improvement in the quality of the data submitted, this dataset will be of even greater value in understanding progress toward universal access and achievement of the Millennium Development Goals. This is critical given the continued absence of any comparable global assessment mechanism.
Tim Slade and Deborah Randall, (National Drug and Alcohol Research Centre, University of New South Wales, Sydney, Australia) provided advice on statistical analysis and Tim Slade performed the meta-analyses and meta-regressions.
1. UNAIDS. 2008 Report on the Global AIDS Epidemic
. Geneva, Switzerland: Joint United Nations Programme on HIV/AIDS; July 2008.
2. Mathers BM, Degenhardt L, Phillips B, et al. Global epidemiology of injecting drug use and HIV among people who inject drugs: a systematic review. The Lancet
3. World Bank. Historical Classifications
: The World Bank Group; July 25, 2008.
4. United Nations. Declaration of Commitment on HIV/AIDS-United Nations special session on HIV/AIDS
. Adopted by the UN General Assembly Twenty-sixth special session: June 25-27, 2001. New York, NY: United Nations; 2001.
5. Horton R, Das P. Putting prevention at the forefront of HIV/AIDS. Lancet
6. Coates T, Richter L, Caceres C. Behavioural strategies to reduce HIV transmission: how to make them work better. The Lancet
7. Wilson D, Halperin DT. “Know your epidemic, know your response”: a useful approach, if we get it right. Lancet
. August 2008 (HIV prevention series: special issue):4.
8. Piot P, Bartos M, Larson H, et al. Coming to terms with complexity: a call to action for HIV prevention. Lancet
. August 2008 (HIV Prevention series: special issue):15.
9. UNAIDS. Monitoring the Declaration of Commitment on HIV/AIDS: Guidelines for Construction of Core Indicators, 2008 Reporting
. United Nations General Assembly Special Session on HIV/AIDS
. Geneva, Switzerland: Joint United Nations Program on HIV/AIDS (UNAIDS); 2007.
10. SPSS for Windows
[computer program]. Version 15.0. Chicago, IL: SPSS Inc; 2007.
11. United Nations Population Division. World Population Prospects: The 2006 Revision
. New York, NY: United Nations; 2007.
12. Stata v9.2
[computer program]. College Station, TX; 2007.
APPENDIX 1. Low and Middle Income Countries (LMIC) in 2007 as classified by the World Bank3 (147 countries) and, of these, where injecting drug use has been reported marked with * (99 countries)
Central African Republic
Congo, Democratic Republic of The
Iran, Islamic Republic of*
Korea, Democratic People's Republic of
Lao People's Democratic Republic*
Libyan Arab Jamahiriya*
Macedonia, The FYRO*
Micronesia, Fed. States of*
Moldova, Republic of*
Papua New Guinea*
Saint Kitts and Nevis
Saint Vincent and the Grenadines
Sao Tome & Principe
Syrian Arab Republic*
Tanzania, United Republic of*
Trinidad and Tobago
© 2009 Lippincott Williams & Wilkins, Inc.