Psychological Distress and Its Associated Risks Among People Who Use Drugs in Cambodia

Background: People who use drugs are at a disproportionately higher risk of mental disorders as a consequence of prolonged exposures to social and psychological issues. However, studies on mental health among people who use drugs in resource-constrained countries are scarce. This study sheds lights on the prevalence and correlates of psychological distress among people who use drugs in Cambodia. Methods: This national survey was conducted in 12 provinces in 2017 using the Respondent Driven Sampling method. A structured questionnaire was used for face-to-face interviews. Psychological distress was measured using the General Health Questionnaire (GHQ-12). A total score of GHQ-12 > 2 indicated high psychological distress. A multiple logistic regression analysis was performed to examine factors associated with psychological distress. Results: This study included 1677 people who use drugs who had an average age of 28.6 years (SD = 7.8). Of the total sample, 41.9% had high psychological distress – 49.7% in women and 37.3% in men. In the regression model, the odds of having high psychological distress was signicantly higher among participants who were 25–34 years old (AOR 1.30, 95% CI 1.01–1.70) and 35 years and above (AOR 1.68, 95% CI 1.19–2.35), had been to a drug rehabilitation center (AOR 2.06, 95% CI 1.48–2.86), had been insulted by family members (AOR 2.09, 95% CI 1.62–2.70), and had been sexually harassed/abused by someone (AOR 1.80, 95% CI 1.38–2.36). The odds of having high psychological distress was signicantly lower among participants who were male (AOR 0.53, 95% CI 0.41–0.69), lived in own dwelling (AOR 0.56, 95% CI 0.41–0.77), reported injecting as the mode of rst drug use (AOR 0.56, 95% CI 0.34–0.91), and had someone taking care of when getting sick when they were growing up (AOR 0.68, 95% CI 0.47–0.99).

South Asia (3.4 million), Sub-Saharan Africa (1.8 million), and Latin America and the Caribbean (1.5 million) [1]. At the country level, China and the United States topped the list in 2017 by having the highest loss of DALYs of 5.6 and 5.5 million due to these disorders, respectively, followed by India (2.5 million), Iran (0.8 million), and Russia (0.7 million) [1]. In Southeast Asia, Indonesia and Vietnam experienced the highest DALYs loss at 0.37 and 0.22 million, respectively in the same year followed by the Philippines (0.17 million) and Thailand (0.16 million) [1].
The relationship between mental health and risky drug use, particularly intravenous drug use and unsafe sexual practices has been well documented in previous studies. People who use drugs with poor mental health, for example severe depressive symptoms, are more likely to adopt unsafe injection practices like sharing needles and syringes [3,6,21]. Besides, they also have more sexual partners and engage in frequent unprotected sex [22,23]. These risky behaviors are associated with a higher risk of acquiring human immunode ciency virus (HIV) [24].
Mental health problems of people who use drugs is also a major public health concern as it not only cooccurs with drug use disorders but also mediates other viral infections such as HIV and hepatitis C virus (HCV). Therefore, studies on risk factors associated with mental health among people who use drugs is critical to inform policies of the indirect bene ts from addressing viral infections and drug use disorders through tackling mental health problems among people who use drugs. Previous studies have documented a number of socio-demographics such as gender, age, and education as correlates of psychological distress [24][25][26], while drug use behaviors and experiences such as duration, frequency, overdose, and drug rehabilitation are also associated risk factors [24,25,[27][28][29]. Exposure to violence, e.g. stigma and discrimination and sexual assault in the past six months, and social support such as family intimacy and adaptability have also been shown to be predictors of psychological distress among people who use drugs [29][30][31][32][33][34][35].
To our knowledge, two studies on mental health of people who use drugs have been conducted in Cambodia [24,25]. Yi et al. included only people who use injecting drugs living in the capital city of Phnom Penh, while Heng et al. used data from the national survey conducted in 2012 and focused primarily on comparing mental health outcome of people who use drugs who had been to a drug rehabilitation center with that of people who use drugs who had no history of rehabilitation. In this study, we used data from the most recent national integrated biological and behavioral survey and expanded the focus to examine the association between several factors in different domains and psychological distress among people who use drugs in Cambodia.

Study design, sites, and participants
This national survey was conducted in 2017 in the capital city and 11 provinces. A feasibility assessment was conducted before selection of the 12 sites, which consisted of 21 operational districts with a high burden of HIV and a large population of people who use drugs. People who use drugs are people who have used any types of illicit drugs in the past 12 months as de ned by the Cambodian Law on Control of Drugs [36]. To be eligible for the survey, participants must: (1) be at least 18 years old; (2) have a predetermined study coupon; (3) never participate in this survey earlier; (4) meet the de nition of people who use drugs; and (5) be able and willing to provide written informed consent to participate in the survey. Details of the main survey have been published elsewhere [37][38][39].

Sample size and sampling procedures
In this survey, sample size calculation was done separately for people who use non-injecting drugs and people who inject drugs. Assumptions were made including a 20% drop in the prevalence of HIV between 2012 and 2017, population size of 13000 people who use non-injecting drugs and 1300 people who inject drugs [40], the prevalence of HIV of 4% among people use non-injecting drugs and 25% people who inject drugs [40] with a margin of error of 1.5% and a con dence interval [CI] of 95%, 90% response rate, and 1.5% design effect. The 12 selected sites were used as strata, while 15% of people who use non-injecting drugs and 22% of people who inject drugs were sampled from their respective estimated population in each site acquired from the National Authority for Combatting Drugs and non-governmental organizations (NGOs) working with people who use drugs across the country. The estimated population of people who use non-injecting drugs were predominantly in Phnom Penh, Banteay Meanchey and Battambang, while that of people who inject drugs was concentrated mainly in Phnom Penh, the capital city.
The Respondent Driven Sampling (RDS) method was adopted in this study because of the hard-to-reach nature of the people who use drugs population. The Strengthening the Reporting of Observational Studies in Epidemiology for RDS studies (STROBE-RDS) statement was followed [41]. There were ve steps in the sampling method. Initially, support from local NGOs in 21 data collection locations within the 12 selected sites was sought to obtain four seeds who had good connection with people who use drugs in each data collection site. Then, a personal identi cation number was assigned to each seed after receiving written informed consent from the participant. Next, each seed was provided with three coupons which they could use to refer three other people who use drugs. A seed having successfully referred a person who use drugs received US$2 and was expected to extend to additional three to six people who use drugs in each site. However, new seeds would be selected had the enrolment not continued because of the driedup of the recruitment tree or had initial seed not recruited participants. Finally, participants recruited for the study were invited to become seeds allowing them to recruit other people who use drugs whom they had known.

Data collection training
All members of the data collection teams attended a three-day training on data collection covering topics such as informed consent process, interview techniques, administration of questionnaire, privacy and con dentiality as well as quality check of the data including spot checking and reviewing the lled questionnaires. The training was primarily aimed to ensure that information was properly collected with minimal errors. All data collection team members were also required to sit in the training on data collection tools and study protocol, while training on HIV and HCV tests were also provided to members assigned to collect the information. Daily consultation between team members and team leaders on data collection issues and monitoring of progress of data collection were also performed by team leaders.

Questionnaire development and measures
A structured questionnaire was developed based on standardized tools utilized in previous studies among HIV key populations [24,40,42,43]. The questionnaire also went through a validation workshop participated by representatives from communities, NGOs, development partners, and national programs working on HIV and harm reduction in Cambodia. The questionnaire was piloted with 20 people who use drugs residing in the capital city of Phnom Penh.
The questionnaire collected information related to socio-demographic characteristics, drug use behaviors, sexual behaviors, HIV and other sexually transmitted infections (STIs), HCV, other substance use, adverse childhood experiences, and psychological distress. Demographic information included age, gender, type of community (urban, rural), years of formal schooling attained, average monthly income earned during past six months, employment status, and living arrangement. Information on drug use covered the sorts of illicit drug used in the past three months, duration of drug use, expenditure on drugs, and the experiences of having been to a rehabilitation center, a detaining center or a prison in the past 12 months.
Five questions on adverse childhood experiences (ACEs) were adapted from the brief screening version of the Childhood Traumatic Questionnaire [47]. The questions asked the studied participants about their experiences of emotional abuse, sexual abuse, physical abuse, and physical and emotional neglect at the time when they were growing up. All ve questions were close-ended, where respondents chose between zero (No) and one (Yes) to describe their ACEs.
The General Health Questionnaire (GHQ-12) was used to assess level of psychological distress among people who use drugs [44]. A four-point Likert-type scale, which varied from "0 = less than usual" to "3 = much more than usual," was applied to each of the 12 questions [44]. A dichotomous variable of "1 = high psychological distress" and "0 = low psychological distress" was developed based on the GHQ-12 guide.
Therein, the four-point Likert-like scale was re-coded by using a "0-0-1-1" method to eliminate bias [45]. The median of the sum of the newly derived scores for the whole sample was used as a cut-off to identify participants with high (GHQ-12 > 2) and low (GHQ-12 ≤ 2) psychological distress [44]. Cronbach's alpha was also estimated to assess the reliability of the scale, and it was 0.88 con rming good reliability of the scale, as it was higher than the minimum acceptable level of 0.50 [46].

Data analyses
Stata (StataCorp LP, version 14.2) was used for data analyses in this study. In bivariate analyses, Chisquare test was used to compare socio-demographic characteristics, substance use, sexual behaviors, and ACEs of participants who had low psychological distress (GHQ-12 ≤ 2) with those of participants who had high psychological distress (GHQ-12 > 2). We applied Fisher's exact test when the expected cell value was less than 5, and Mann-Whitney test, rather than Student's t-test, was used when a continuous variable does not follow normal distribution. We conducted a multiple logistic regression analysis to examine factors associated with psychological distress controlling for potential confounders. We included age and gender regardless their level statistical signi cance and other variables that had a pvalue < 0.05 in the bivariate analyses in the model. Adjusted odds ratio (AOR) were obtained and shown with CI and p-values.

Ethical considerations
All individuals participated in the study on a voluntary basis and provided written informed consent. Our teams ensured that participants were well informed of risks that may have occurred and bene ts that they may have received during the study. We removed personal identi ers and kept collected information in a private room to maintain privacy and con dentiality of participants. This study received ethic approval from the National Ethics Committee for Health Research (NECHR) of the Ministry of Health in Cambodia (No. 420 NECHR).

Results
This study included 1677 people who use drugs who had an average age of 28.6 (SD = 7.8) years, average years of schooling completed of 6.0 years (SD = 3.9), and a median monthly income in the past six months of US$100.0 (interquartile range [IQR] = 90). People who inject drugs constituted to around 18.5% of the study sample. Around one-third of the participants were residing in Phnom Penh, while the other 16% were living in Banteay Meanchey and 11% in Battambang province, which border Thailand. Of the total sample, 41.9% had high psychological distress -49.7% in women compared to 37.3% in men.

Socio-demographic characteristics
Socio-demographic characteristics are presented in Table 1. Almost two-thirds (62.8%) of the study participants were male. The majority of the participants (88.5%) were living in an urban setting. About two-thirds (65.6%) were in the age group of 25 and above, and 53.6% reported having primary or no formal education. Almost all participants (94.5%) were in Khmer ethnic group, and 46.2% were never married. While 46.7% of the participants were living with their family or relatives, 10.3% were living on the streets. Approximately one-third (35.6%) of the participants had their main career as a laborer or a farmer, while 18.3% worked as an entertainment worker. The majority (82.6%) reported having an average monthly earning of less than US$200. Psychological distress was measured using the General Health Questionnaire (GHQ-12), and a total score of GHQ-12 > 2 was used to de ne high psychological distress. * Chi-square (or Fisher's exact test when a cell count was smaller than 5) was used.

Substance use
As shown in Table 2, participants reported having used drugs for a median duration of 18 months (IQR = 42). Nearly one in 10 (9.3%) described injecting as the mode of their rst drug use. The proportion of participants who had been to a prison and a drug rehabilitation center in the past 12 months was 11.2% and 15.8%, respectively. More than two-thirds reported having daily alcohol drinking habit, while half had smoked at least 100 cigarettes in their lifetime. The proportions of participants who reported having used any drugs in the past three months (85.2% vs. 75.6%, p < 0.001), used methamphetamine in the past three months (77.8% vs. 69.3%, p < 0.001), and been to a drug rehabilitation center (22.3% vs. 11.2%, p < 0.001) were signi cantly higher among participants who had high psychological distress compared to those among participants who had low psychological distress. Psychological distress was measured using the General Health Questionnaire (GHQ-12), and a total score of GHQ-12 > 2 was used to de ne high psychological distress. * Chi-square (or Fisher's exact test when a cell count was smaller than 5) was used for categorical variables and Mann-Whitney U test for continuous variables. Table 3 shows sexual behaviors among the study participants in the past three months. Of the total, 92.8% reported having had sexual intercourse; of them, 24.9% reported having always used condoms with any partner. Of the 1557 sexually active participants, 39.7% reported having sexual intercourse while a partner was intoxicated. One in ve (20.5%) reported having always used condoms with partners without an exchange for money or gifts. More than one-third (35.8%) reported having sexual intercourse in an exchange for money or gifts, and 39.2% reported always using condoms with the commercial partners.

Sexual behaviors
The proportions of participants who reported having had sexual intercourse while a partner was intoxicated (49.5% vs. 31.9%, p < 0.001), having sex with partners without an exchange for money or gifts (58.6% vs. 49.9%, p = 0.003), and having sexual intercourse in an exchange for money or gifts (41.5% vs. 31.2%, p < 0.001) were signi cantly higher among participants who had high psychological distress compared to those among participants who had low psychological distress. Adverse childhood experiences (ACEs) As shown in Table 4, 54.5% reported having been slapped, kicked or received physical punishment from a family member or a guardian when they were growing up. Of the total, 51.2% described having been insulted by a family member or a guardian, and 21.9% reported having been sexually harassed or abused. The majority of participants reported having been taken care of by someone in the family (89.3%) and having received emotional support and care from a family member (85.4%). Participants who had high psychological distress were signi cantly more likely to report that they had experience with physical punishment from a family member or a guardian (62.7% vs. 48.6%, p < 0.001), insulted by a family member or a guardian (63.5% vs. 42.4%, p < 0.001), and had been sexually harassed/abused by someone (30.1% vs. 16.0%, p < 0.001) when they were growing up. Psychological distress was measured using the General Health Questionnaire (GHQ-12), and a total score of GHQ-12 > 2 was used to de ne high psychological distress. * Chi-square (or Fisher's exact test when a cell count was smaller than 5) was used for categorical variables and Mann-Whitney U test for continuous variables.

Factors associated with psychological distress
Results of multiple logistic regression analysis are presented in Table 5  Psychological distress was measured using the General Health Questionnaire (GHQ-12), and a total score of GHQ-12 > 2 was used to de ne high psychological distress.

Discussion
This study provides evidence of the prevalence and factors associated with psychological distress, measured by GHQ-12, among people who use drugs in a resource-constrained country. It showed that the prevalence of psychological distress among people who use drugs in this study was 41.9%, which was similar to the 42% found in our previous study conducted in 2014 using the same measure of psychological distress among people who inject drugs in the capital city of Phnom Penh [24].
Our ndings suggested that male people who use drugs were less likely to have high psychological distress than their female counterparts, which is consistent with results shown in previous studies [24][25][26]. In Taiwan, female people who use drugs were more likely to have suicidal thoughts than their male counterparts, which resulted in suffering from more psychiatric illnesses than men [26]. In our additional analyses, we further examined this gender difference by separately including the interaction terms between gender and main occupational categories and age groups in a multiple logistic regression model. We found that female participants who were sex workers (AOR = 30.3, p = 0.004), sellers (AOR = 2.8, p = 0.03) and farmers (AOR = 4.0, p = 0.001) experienced signi cantly higher level of psychological distress than their male counterparts. Adjusted odds ratios of the interaction terms between gender and other relevant variables such as ACEs, suicide ideation and drug use duration were not statistically signi cant at any conventional level.
People who use drugs aged 25 and above were found to be signi cantly more likely to have high psychological distress than their younger counterparts aged 18-24, and this corroborates with ndings in a previous study in Cambodia [25]. We observed that older people who use drugs (aged 25+) were more likely to have suicidal thoughts (22.5% vs 12.7%, p < 0.001) and drink alcohol four times or more per week in the past three months (29.3% vs 23.3%, p < 0.001) than their younger counterparts (aged [18][19][20][21][22][23][24]. The co-occurrence of substance use and mental illness among people who use drugs [11][12][13][14] and the association between suicidal thought and mental health [3,24] has also been documented in previous studies. On the living arrangement, we found that people who use drugs living in their own dwelling were less distressed than those living with family. We probed this association further by including the interaction terms between living arrangements and ACEs quintiles generated using principal component analysis in a multiple logistic regression model. The results showed that people who use drugs living in own dwelling and having more ACEs (2nd quintile: AOR = 0.33, p = 0.03; 3rd quintile: AOR = 0.39, p = 0.04; 4th quintile: AOR = 0.49, p = 0.04; 5th quintile: AOR = 0.40, p = 0.02) were less psychological distressed than people who use drugs living with family. ACEs could have been a driver of the choice of living arrangement. And it should be noted that marriage and cohabitation might have mediated the association as 60.4% of the people who use drugs living in own dwelling were married and cohabited. Breaking free from past ACEs with family and forming their independent cohabitation with a spouse, people who use drugs living in own dwelling were less psychological distressed than those living with family.
People who use drugs who had been to a drug rehabilitation center were found to have more psychological distress than people who use drugs who had never been to the center, which con rmed ndings in previous studies in Cambodia [24,25], suggesting that the role of drug rehabilitation center remained counterintuitive ve years later. Poor overall quality of life and health of people who use drugs with a history rehabilitation could be the underlying drivers of level of psychological distress. However, we found no statistically signi cant interactions with having been to a rehabilitation center and quality of life and health. Therefore, adverse experience in rehabilitation centers remained a valid explanation as indicated in another study in 2012 [25]. Similar accounts had also been documented in China [48], Taiwan [26], and Vietnam [49].
Last, on the ACEs, people who use drugs who had been insulted by family members or guardians or had been sexually harassed or abused while growing up exhibited a higher level psychological distress than those who had not. On the other hand, people who use drugs who experienced care provided by someone when getting sick as a child were less psychological distressed than those who did not. These ndings are consistent with those in studies in the United States [50] and Cambodia [24]. In their systematic review, De Venter et al. showed that people who use drugs who had experienced ACEs exhibited symptoms or diagnoses of depressive and anxiety disorders [51]. The pathway of the association likely ran from ACEs to entry to drug use and later to psychological distress. Another study showed that people aged 14 and above in California with higher ACE score were about two to four times more likely to initiate drug use than people with lower ACE score [52].

Limitations of the study
Despite several strengths, limitations of this study should be noted. First, causal inference from our multiple logistic regression analysis could not be made as we did not address endogeneity (e.g. omitted variable bias or reverse causality) of each independent variable, and thus results should be interpreted as the association between the dependent variable and independent variables included in the model. Second, since our measure of psychological distress was constructed based on self-reported responses to the GHQ-12 questionnaire, but not performance-based psychological measure, our results of the association could be biased due to social desirability and recall bias. Third, our results were also likely biased due to selection of people who use drugs into the study sample because the study targeted locations with heavy burden of HIV and drug use in order to obtain as large sample as possible, and participants likely self-selected into the study as they could have been motivated by the incentive (token) provided through the RDS method.

Conclusions
This study documents a high prevalence of and risk factors associated with psychological distress among people who use drugs in Cambodia. Risk factors found to be signi cantly associated with psychological distress among people who use drugs in this study included gender, age, history of drug rehabilitation, and ACEs. Women showed higher level of psychological distress than men, while people aged 25 and older also exhibited higher level of psychological distress than people aged 18 to 24. Our results on the counterintuitive role of drug rehabilitation and the negative effect of ACEs on psychological distress of people who use drugs also corroborated with ndings in earlier studies. Therefore, interventions that attempt to address mental health among people who use drugs should be gender-and age-sensitive, while they should be tailored to more vulnerable and marginalized subpopulations and individuals with a history of ACEs and drug rehabilitation.

Declarations
Authors' contributions SY, ST, and PM designed the study and developed the study protocol and tools. NC, PC, TS, and PM were responsible for trainings and data collection. CS, KP, PU, and SY analyzed data and wrote the manuscript.
All authors contributed to the conceptualization of the research questions, interpretation of the results, and manuscript writing. All authors read and approved the nal manuscript.