Background
Illicit drug use is a significant global health problem. Annually, a quarter of a billion people use at least one illicit drug, with cannabis, amphetamines, cocaine, and heroin being the most common [
1]. Evidence from a recent systematic review suggested that illicit drug dependence directly accounted for 20 million disability-adjusted life years (DALYs) in 2010, representing 0.8% of global all-cause DALYs [
2]. This burden of disease is exacerbated by the intersection between the use of illicit drugs, poor mental health, and suicides [
2,
3].
In Cambodia, people who use drugs (PWUD), particularly people who inject drugs (PWID), are considered an important group for HIV prevention and harm reduction programs. The current Cambodia 3.0 framework aims to eliminate new HIV infections in the country by 2020, by accelerating prevention among all key populations including PWUD and PWID, female entertainment workers, men who have sex with men, and transgender women [
4]. To achieve this goal, a key emphasis in the national response is an aspiration to strengthen the availability of data and strategic information related to the HIV epidemic among the key populations [
5].
In relation to illicit drug use, however, greater emphasis in Cambodia has been placed on PWID with limited attention to people who use drugs via non-injecting modes [
5,
6]. While injection of illicit drugs clearly increases the risk of HIV and viral hepatitis infections [
7,
8], ignoring non-injecting consumption of illicit drugs will prevent the achievement of the Cambodian national goal of eliminating HIV infection. We argue that focusing on wider substance abuse is important for two reasons. First, the use of illicit drugs, particularly stimulants, has been shown to increase risk of HIV, regardless of injection [
7]. Secondly, the use of illicit drugs such as cannabis, amphetamines, cocaine, or heroin frequently occurs alongside the consumption of other substances such as alcohol and tobacco [
9,
10], which have their own impacts on health and health behaviors. For instance, the concomitant use of alcohol alongside the illicit drugs has been shown to further compromise HIV prevention by reducing condom use [
11].
In the Cambodian context, PWUD are specifically defined as people who have used any types of illicit drugs, as specified in the Cambodian Law on Control of Drugs, via any route of administration other than injection in the past 12 months [
12]. In 2012, it was estimated that there were 13,000 PWUD and 1300 PWID nationally [
4]. A more recent size estimation exercise conducted alongside this national survey concluded that there were 22,374 PWUD and 4136 PWID in the country in 2017 [
13]. In both of these estimates, PWID were not subsumed under PWUD. Despite the large numbers of PWUD, studies focusing on this population are limited. Rare studies have shown that PWUD in Cambodia have an increased risk of mental health problems [
14]. However, further data related to this population are needed to inform HIV prevention and other public health interventions. Accordingly, characterization of the extent and risk factors of non-injecting drug use was purposely emphasized in this Integrated Biological and Behavioral Survey (IBBS 2017). In this paper, we reported the prevalence of HIV and factors associated with HIV infection among PWUD in Cambodia. We also discussed the potential implications of these findings on HIV prevention policies and programs.
Methods
Study design, sites, and participants
This cross-sectional IBBS 2017 among PWUD and PWID in Cambodia was conducted from June to December 2017 in the capital city of Phnom Penh and 11 other major provinces purposively selected by the research team and approved by the National Center for HIV/AIDS, Dermatology and STD (NCHADS) and National Authority for Combating Drugs (NACD). These study sites were selected based on results from a feasibility assessment conducted as part of the study designing and protocol development. About 70% of the Cambodian population resides in the 12 provinces out of 25 provinces in Cambodia. The 12 provinces included: Banteay Meanchey, Battambang, Kampong Cham, Kampong Chhnang, Kampong Speu, Kandal, Phnom Penh, Preah Sihanouk, Prey Veng, Siem Reap, Svay Rieng, and Tbong Khmum. According to the national epidemiological data from NCHADS and size estimation studies, the selected sites consisted of 21 operational districts with a high burden of HIV and drug use, i.e. more than 80% of estimated drug use in Cambodia. To be included in the study, an individual must: (1) be at least 18 years old; (2) have a valid study coupon; (3) have not participated in this survey before; (4) meet the criteria defining PWUD; and (5) be able and willing to provide a written informed consent.
Sample size and sampling procedures
The estimated PWUD population size of 13,000 [
15] and the assumption that the prevalence of HIV would be lower compared to the prevalence in the national survey in 2012, justified on the basis of the scaled up provision of community-based harm reduction services in recent years [
4,
6], informed the sample size calculation for this survey. Using an estimated HIV prevalence of 3.5%, a margin error of 1.5%, a confidence interval of 95%, a response rate of 90%, and a design effect of 1.5, we derived a minimum sample size of 1390 for this study. The calculated sample size was further stratified by study sites, and roughly 15% of the estimated PWUD in each site were recruited. Data collection was conducted in 21 locations: six locations in Phnom Penh and 15 locations in the remaining 11 provinces. The precise number of the selected locations was determined based on the required sample size in each study site. The Respondent Driven Sampling (RDS) method was used to recruit study participants.
The RDS was implemented in five stages. First, eligible four seeds who were well connected to other PWUD in each location were invited to participate in the survey with support from a non-governmental organization (NGO) working in the area. A written informed consent was obtained from them. Second, each seed was given a Personal Identification Number (PIN) and enrolled as a participant. Third, each seed received three coupons and were asked to refer three additional PWUD. They would receive US$2 for a successful referral. Each seed was expected to extend up to three to six “recruitment waves” at each site. If the initial seeds did not recruit participants or if the enrollment has been halted because all recruitment chains have “dried up” (i.e. stopped recruiting), additional seeds would be selected based on the inclusion criteria. Finally, recruited individuals were provided the same opportunity as seeds to recruit other PWUD. In total, 84 seeds were selected for the entire survey.
Data collection training
All data collection team members participated in a three-day training workshop, which covered a review of the study protocol, data collection methods, tool pretesting, and reflection. The training also included interview techniques, confidentiality, and privacy and provided the study teams with an opportunity to practice questionnaire administration. The training facilitators explained the study protocol and tools in details during the sessions to ensure that all team members were thoroughly familiar with it. A regular daily review session with data collection teams were conducted during the survey period to review progress and communicate any issues encountered during the data collection.
Data collection procedures
HIV and syphilis test
HIV and syphilis screening were performed by a laboratory technician using SD Bioline HIV/Syphilis Duo test. A non-HIV reactive result established that an individual was not HIV-infected. In the case of an HIV reactive result from the SD Bioline HIV/Syphilis Duo test, a confirmatory test was conducted on site using HIV 1/2 STAT-PAK® Assay. The participants were given an option to receive their screening results verbally in a post-test counseling session after the questionnaire interview. A follow-up on newly detected HIV-infected individuals was pursued closely by a local NGO working in the area. Participants received a token gift, which costed approximately US$5, to compensate them for their time and transport expenses.
Questionnaire development and measures
A structured questionnaire was developed using standardized tools adapted from previous studies among HIV key populations in Cambodia and the most recent Cambodia Demographic and Health Survey [
14‐
18]. The questionnaire was initially designed in English and then translated into Khmer, the national language of Cambodia. Another translator then back-translated it into English to ensure that the “content and spirit” of every original item were maintained. A consultative meeting was held with representatives of PWUD and key stakeholders working on HIV and harm reduction in Cambodia. A pilot study was conducted with 20 PWUD in Phnom Penh, who were later excluded from the main study.
Socio-demographic characteristics included geographical location of the study site (urban or rural), age (continuous), gender, level of education attained (continuous), average monthly income in the past 6 months (continuous), living situations, and employment status. Regarding drug use, we collected information on types of illicit drugs most commonly used and frequency of use in the past 3 months. We also collected information regarding exposure to community-based HIV, harm reduction, and other related services in the past 6 months. To measure HIV risks, participants were asked about their sexual behaviors in the past 3 months including number of partners and condom use with commercial (defined as a partner with whom the participant had sex in exchange for money or goods) and non-commercial partners in the past 3 months. Other collected information included HIV testing history, STI symptoms, and care seeking behaviors as well as the use of other substances (tobacco and alcohol) in the past 3 months.
Data analyses
The prevalence of HIV was calculated by dividing the total number of participants tested positive for HIV with the total number of participants tested. In bivariate analyses, we compared characteristics and behavioral variables of participants who were HIV positive to those of participants who were HIV negative. Chi-square test (or Fisher’s exact test for an expected cell value of ≤5) was used for categorical variables, and Student’s t-test or Mann-Whitney U test for continuous variables. To facilitate the analyses, age, level of education, and income were transformed into categorical variables. A multivariable logistic regression model was constructed to identify independent factors associated with HIV infection. Variables with a significance level of p < 0.05 in the bivariate analyses were simultaneously included in the model. Age, gender, level of education, and income were included in the model regardless of the significance level in bivariate analyses because they have been found associated with HIV infection in previous studies.
Ethical considerations
This study was approved by the National Ethics Committee for Health Research (NECHR) of the Ministry of Health, Cambodia (No. 193 NECHR). Privacy and confidentiality of the participants were protected by conducting the data collection in a private room and removing all personal identifiers from the study documents. Participants were fully informed about the voluntary nature of the study as well as risks and benefits they may expect from their participation in the study. A written informed consent was obtained from each participant.
Discussion
This paper presents crucial data regarding the prevalence of HIV and its risk factors among Cambodian PWUD sampled in the most recent national IBBS conducted in 2017. The findings present useful information related to the profile of substances used by the participants. Methamphetamines were particularly prominent in the profile of substances recently consumed (95.1%), while ecstasy, cannabis, and other inhalants feature less, having been used by 3.0 to 5.4% of the participants. These findings concur with previous studies that have reported the ubiquity of methamphetamines in Cambodia [
19‐
22]. As expected, most PWUD used drugs by way of smoking, while fewer others snorted or swallowed them. A small proportion (0.2%) had a history of drug injection in their lifetime, even though this was more than 12 months prior to the survey.
Overall, we found that the prevalence of HIV was 5.7%, about 10 times higher than the prevalence in the general adult population [
6]. This finding confirms that non-injecting drug users are highly vulnerable to HIV infection. Independent risk factors of HIV infection identified in this study included older age, lower level of education, homelessness, perception of high level of HIV risk, previous drug injection, and STI symptoms. These findings warrant further discussion.
To start with, these findings suggest that older age may be related to the risk of HIV infection. This is in contrast to other studies in China and South Africa, which did not find a relationship between HIV infection and age among methamphetamine users [
23,
24]. However, a previous study among transgender in Cambodia reported that older age increased the risk of HIV infection [
25]. Bivariate findings in our study suggested that duration of use is predictive of HIV infection, and the duration was also identified as predictive of acquisition of HIV and hepatitis C virus infections in China [
23]. As PWUD continue to consume drugs, the cumulative chances of HIV acquisition would be expected to increase over time, especially if these drugs are consumed in environments that promote risky sexual practices. Our study suggests that this could be the case. For instance, 31.5% of the participants reported alcohol drinking at least four times per week, and 33.0% reported binge drinking at least on 4 days per week in the past 3 months. Furthermore, the prevalence of transactional sex and STIs in our sample was considerably high, and genital ulcers were independently predictive of HIV infection. Stimulants such as methamphetamine and ecstasy are particularly associated with risky sexual practices such as unprotected or transactional sex with multiple partners [
7,
26,
27] through enhanced sexual desire [
28].
Not surprisingly, history of having used injecting drugs was associated with higher likelihood of being diagnosed with HIV. This is consistent with assertions that injecting drug use is a potential risk factor for HIV due to the risk of sharing contaminated needles [
7]. Together with the above assertions, this finding suggests that users of stimulants predominantly expose themselves to HIV sexually, while drug injection can also determine HIV acquisition among those who transition to injecting, even temporarily. The finding that a history of drug injection tripled the odds of HIV infection provides a compelling reason to reach out to people who are currently using drugs by routes other than injection, as some of them may have injected in the past, often over a year ago as was the case in our study. Taking the history of lifetime drug injection provides a programmatically useful way of stratifying risk among current PWUD. Stratifying profiles of HIV risk could be an important and cost-effective strategy for ensuring that every HIV infection is identified within the current 90–90-90 global goals and the Cambodia’s 3.0 strategy.
A quarter of the respondents in this study perceived that they were at higher risk of HIV infection compared to the general population. HIV was more likely to occur among these PWUD who perceived that they were at high HIV risk, perhaps based on their sexual behaviors, which is not necessarily surprising. Perception of risk has been shown to influence the uptake of HIV prevention interventions [
29], and to act in association with contextual factors to influence the risk of HIV infection [
30]. As noted in Table
5, the majority of the respondents in this survey had been reached with a range of community-based HIV programs, including risk reduction education interventions and education related to their risks. It is therefore reasonable to expect that these populations would be aware of risks of HIV infection. The importance of our findings, however, is that participants who clearly thought they were at high risk might have not been taken measures to reduce those risks, leading to their acquisition of HIV infection. This discordancy might be explained by a lack of self-efficacy to reduce their exposure to HIV, or a range of external determinants that could reduce their ability to manage their exposure to HIV despite being aware of risks.
Determinants of health behaviors such as low self-efficacy, low motivation, and external barriers have been shown to prevent PWID in India from accessing health services [
31]. Data showing the complex interplay between recognition of vulnerability, rationalization of risk, and inaction to protect oneself from HIV despite high-risk perception has been reported among heterosexual women in Kenya and South Africa [
30]. Therefore, it is not entirely implausible that a range of barriers prevented PWUD who thought that they were at high risk from adopting safer behaviors. However, the validity of this speculation will need to be explored specifically among this sub-population who perceive higher risk of HIV compared to the general population. Previous studies conducted in Cambodia have highlighted that the lack of an enabling environment perpetuated by legal and social factors impedes PWUD’s access to health service including HIV prevention and care, and thus threat efforts to control HIV among PWUD [
14,
32]. Future research should seek to understand how perceived high HIV risk is actioned or otherwise rationalized, leading to inaction, among PWUD in our study setting. This is particularly important given that more than one-third (35.2%) of the HIV-positive cases were not aware of their status prior to the survey.
An important finding from this study is that PWUD who lived on the streets were more likely to be infected with HIV. Previous studies have shown the impact of lack of housing on risky substance use and sexual behaviors among PWUD, including those who used amphetamine and ecstasy [
33‐
35]. Structural disadvantages, such as homelessness, are known to encourage high-risk behaviors by providing environments for experimentation among street-dwellers [
36]. This finding is important as it provides an additional and practically useful way in which to stratify profiles of risk among PWUD, while suggesting that structural shelter accommodation could be useful. At the same time, programs should place a higher emphasis on homeless PWUD, ensuring that outreach workers provide intensive education and support to this population.
Finally, an expected finding from this study was the observation that lower education predicted higher risk of HIV infection. Previous studies in Cambodia have shown that both substance use and HIV infection are associated with lower levels of education [
25,
37]. This finding adds additional impetus for reaching PWUD with information and education regarding harms of drug use, with relevant and easy-to-access materials, or through peer-led approaches, prioritizing those that have low literacy and education levels.
Limitations of the study
In interpreting the findings from this survey, several limitations should be borne in mind. First, this was a cross-sectional study which provides a snapshot of the prevalence of HIV and related risk behaviors, and does not capture the dynamic way in which profiles and modes of drug use change in the life course [
38]. Second, this survey was conducted in 12 provinces with a high burden of HIV and drug use, leaving out some 13 other provinces with lower prevalence of both HIV and drug use. This may affect the generalizability of the study findings at the national level. Third, self-reporting measures were used to collect a number of correlates, which are likely to be affected by social desirability bias known to affect reporting of sensitive drug use and sexual behaviors [
39]. Finally, this survey used incentives in the RDS sampling to recruit the participants, which may have affected PWUD’s genuine motivation to participate. However, we anticipate that the resultant motivational bias had minimal impact on our overall findings.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.