HIV and associated factors among female sex workers in Tajikistan: Analysis from a national bio-behavioral survey

Abstract

Tajikistan, a country of approximately nine million people, has a relatively small but quickly growing HIV epidemic. No peer-reviewed study has assessed factors associated with HIV, or associated risk factors, among female sex workers (FSWs) in Tajikistan. The purpose of the current study is to elucidate the factors associated with HIV status and risk factors in the Tajikistani context and add to the scant literature on risk factors among FSWs in Tajikistan and Central Asia. We used cross-sectional data from an HIV bio-behavioral survey (BBS) conducted among FSWs in the Republic of Tajikistan (n = 2174) in 2017. Using Respondent Driven Sampling Analysis Tool software, we calculated the prevalence of HIV, diagnosed cases, linkage to antiretroviral therapy (ART), and the prevalence of syphilis for FSWs in Tajikistan. Prevalence data were adjusted for network size and any clustering effects in the network. Further, using univariate and multivariable logistic regression, we determined correlates of HIV-positive status. Results were as follows: Of all FSWs in Tajikistan, 2.6% (95% CI: 1.7–3.8%) are HIV positive, 2.3% (95% CI: 1.4–3.5%) are diagnosed and aware of their status, and 2.0% (95% CI: 1.2–3.1%) are on ART. About 5.7% (95% CI: 4.5–7.4%) of FSWs in Tajikistan have ever had syphilis, and 0.8% (95% CI: 0.4–1.3%) have active syphilis infections. The epidemic of injection drug use was found to be strongly synergistic with HIV infection as having had sex with a person who injects drugs was shown to be strongly associated with HIV-positive status (OR: 5.2; 95% CI: 2.6–10.2) in the multivariable model. While this study estimates that HIV prevalence among Tajikistani FSWs is relatively low, it is likely an underestimated due to selection and social desirability biases. To curb the small, but potentially volatile, HIV epidemic among FSWs, the government should consider targeted testing and linkage-to-care efforts for FSWs who inject drugs or who have people who inject drugs partners. Services should also be prioritized in Gorno-Badakhshan, which has a higher number of FSWs per capita relative to other regions. Additionally, the link between HIV and experiences of stigma, violence, and discrimination against FSWs should motivate advocates to protect Tajikistani FSWs from these experiences.

Introduction

Sex workers and their partners are at high risk for HIV infection, making up more than half of all new HIV infections globally.¹ There are significant barriers to linking female sex workers (FSWs) with HIV testing, care, and antiviral therapy,^2,3 making the HIV epidemic in this population a global priority.

Eastern Europe and Central Asia (EECA) is the only UNAIDS-defined region in the world where the annual rate of HIV incidence continues to rise.⁴ Tajikistan, a Central Asian country of approximately nine million people, has a relatively small, but quickly growing, HIV epidemic that parallels HIV trends in the broader EECA region.²

In Tajikistan, a low-income country where the annual GDP per capita is about $826,⁵ sex work has increasingly become a way for women to maintain a stable income.⁶ The UNAIDS estimates that there are 17,500 sex workers currently living in Tajikistan.⁷ While the purchase of sex work is legal, the provision of sexual services is a minor misdemeanor crime.⁸ As of 2015, a new amendment allows the arrest of “repeat” sex workers and penalties of up to two weeks under house arrest.⁹ Beyond punitive laws,⁹ FSWs face high levels of violence and stigma in Tajikistan, with reports of physical abuse inflicted by commercial partners and instances of police violating the privacy of FSWs.¹⁰

Only one systematic review has addressed the state of the literature on factors associated with HIV infection among FSWs in Europe and Central Asia.¹¹ The 2013 review, which focused on studies conducted in Europe and Central Asia, found high heterogeneity between studies. The review showed that HIV prevalence in post-Soviet countries of Eastern Europe and Central Asia varied significantly, and injecting drug use was consistently found to be a predictor of HIV-positive status. For FSWs who inject drugs, selling sex for drugs, injecting daily, and injecting homemade drugs were found to be predictive of HIV status. One study in nearby Uzbekistan¹² found that drug use, and particularly injection drug use, was the strongest associated factor with HIV-positive status. Other individual risk factors included: unprotected sex, number of commercial partners, having noncommercial sexual partners, and having sex with persons living with HIV. Structural risk factors, like the type of sex work venue, the existence of outreach teams, and migration, were inconsistently associated with HIV status. A recent systematic review has determined that coercive policing practices are associated with HIV risk among FSWs globally.¹³ Another systematic review¹⁴ found that social support for FSWs is consistently associated with behavior protective against HIV.

A Tajikistani FSW is about 20 times more likely to be HIV positive than a member of the general population. Female sex workers make up only 0.16% of the Tajikistan’s population, but 3.04% of known HIV cases presented in the previous literature.¹⁵ Despite the abundance of literature discussing HIV risks among FSWs globally, no peer-reviewed study has assessed this problem in Tajikistan. The purpose of the current study is to elucidate the factors associated with HIV status and risk factors in the Tajikistani context and add to scant literature on risk factors among FSWs in Tajikistan and Central Asia.

To accomplish this aim, we analyzed cross-sectional data from a 2017 bio-behavioral survey (BBS) conducted among FSWs in Tajikistan.

Methods

This study presents an analysis of cross-sectional data from an HIV BBS conducted among FSWs in the Republic of Tajikistan (n = 2174) in 2017. The data were collected using face-to-face interviewer-administered questionnaires and blood samples, which were screened for HIV first, using the HIV 1/2 STAT-PAK^® DIPSTICK rapid test and then confirmed by the Murex HIV Ag/Ab combination ELISA test. Screening for syphilis was conducted using the CHEMBIO DPP^® screen and was confirmed using a rapid test kit. The WHO has approved both tests for epidemiological surveillance.¹⁰ The BBS was conducted in eight cities of Tajikistan, and respondents were recruited using respondent-driven sampling (RDS). These eight cities are Dushanbe, Vahdat, Rudaki, Panjikent, Kurgan-Tebe, Kulob, Khujand, and Khorog. Dushanbe is the country’s capital, while Rudaki and Vahdat are near the capital. Kurgan-Tebe and Kulob are cities in the south of Tajikistan, while Panjikent and Khujand are located in the relatively economically prosperous northern region. Khorog is a city in Gorno-Badakhshan, a large, sparsely populated, mountainous area in the east of Tajikistan which shares a long border with Afghanistan. Gorno-Badakhshan is the poorest region in Tajikistan and is populated by Tajikistan’s Ismaili Shia ethnic and religious minority.

Recruitment

All FSWs were recruited with the help of local NGOs. Two initial respondents (seeds) were recruited from each of the eight sites. Each seed was given three coupons to distribute to members of their peer networks. NGO workers assisted the seed participants in recruiting peers along the following diversity criteria: (1) “type of FSWs (locations where commercial partners are solicited and locations of providing sex),” (2) “geography (residing in different parts of the city/district), (3) “different marital statuses, education levels, ages, and ethnicities,” (4) “length of time participating in sex work,” (5) “high and low risk (using drugs vs not using drugs, HIV positive vs not HIV positive, unprotected sex vs protected sex, etc.),” and (6) “degree of engagement with NGO services.” Each recruited participant was given an additional three coupons to recruit more peer FSWs in the same manner.

Inclusion criteria

To be included in the study, an individual had to be a woman aged 18 years or older, who had exchanged vaginal and/or anal sex for money, goods, or services within the six months preceding their participation in the survey, and spoke Tajik or Russian well enough to participate in the survey. Participants had to provide verbal consent for all study procedures to be included in the study.

Respondent-driven sampling

Respondent-driven sampling “is a probability-based sampling method which includes an analysis process to adjust for social network sizes (number of peers known to participants) and differential recruitment.”¹¹ Unlike non-probabilistic sampling methods, RDS enables inferences to be made about the networks of a given population. As RDS requires that a population group be well-networked, preliminary qualitative interviews were conducted to assess FSW network size. The preliminary data showed that FSWs have sufficiently large network sizes for the proposed sampling method.

Power calculations

The average sample size was 275 FSWs per interview site, with a total of eight sites and 2174 FSWs. The sample sizes in any given location ranged between 200 and 500. Sample size and power calculations were based on the original objective of the survey, which was to estimate the prevalence of HIV among FSWs. The following formula was used:

$$n=\frac{\left\{DEFF\times \frac{z_{1-a/2}^2}{d^2}\right\}}{1-NR}$$

where n is the minimum sample size required and DEFF is the design effect that accounts for the increased variation of the estimates under RDS. A design effect of 2.0 was used in the calculation based on evidence in the literature for RDS surveys;¹⁶ Z_1−α/2 is the z-score corresponding to a 90% confidence interval (i.e. z = 1.645), p is the expected prevalence in the sample based on previous studies, q is the population not living with HIV (i.e. 1-p), d is the margin of error of the confidence interval, and NR is the nonresponse rate (which is zero, based on the nature of the study design).

Measures

Drawing on the results of a systematic review of studies on HIV and other STIs among FSWs in Europe and Central Asia,¹¹ the following factors were included as independent variables of HIV risk: injection drug use, selling sex for drugs, having a sexual partner who injects drugs, unprotected sex with commercial partners, number of commercial partners, number of casual or permanent partners, condom use, street versus off-street venues, sex with someone living with HIV, and experiences of violence.

Injection drug use was assessed by the question: “Have you ever injected (intravenous) drugs?” to which participants could respond “yes,” “no,” “do not know,” or refuse to answer.

Selling sex for drugs was assessed by the question: “Have you exchanged sex in order to obtain a drug dose in the last 12 months?” to which participants could respond “yes,” “no,” “do not know,” or refuse to answer.

Having a sexual partner who injects drugs was assessed by the question: “Have you ever had sex with a person who uses drugs?” to which participants could respond “yes,” “no,” “do not know,” or refuse to answer.

Unprotected sex with commercial partners was assessed by the question: “How often did you use condoms with commercial sex partners in the last 30 days?” to which participants could respond “always,” “almost always,” “half of the time,” “sometimes,” “never,” “I do not have such a partner,” “I do not know,” or refuse to answer.

Number of commercial partners was assessed by the question: “How many commercial partners did you have in the last seven days?” to which participants could give a numeric response, say “I do not know” or refuse to answer.

Number of casual or permanent partners was assessed with the questions: “Of your sex partners from the last 30 days, how many are casual partners?” and “Of your sex partners from the last 30 days, how many are permanent partners?” to which participants could give a numeric answer, say “I do not know” or refuse to answer.

Condom use was assessed using three questions: “How often did you use condoms in the last 30 days with regular or casual or permanent partners?” to which participants could respond “always,” “almost always,” “half of the time,” “sometimes,” “never,” “I do not have such a partner,” “I do not know,” or refuse to answer. The same question was repeated three times, each time replacing the type of partners (regular/casual/permanent).

Working street venues was assessed with the question: “How did you usually find new commercial partners in the last 12 months?” to which participants gave open-ended answers. Interviewers then coded their responses as any combination of “in a sauna, “on the street,” “by telephone,” “through a pimp/mother,” “through other commercial partners,” “via the Internet,” “other,” “do not know,” or “no answer.” This variable was recoded as binary depending on whether respondents reported finding commercial partners “on the street” or whether they did not report this.

Sex with someone living with HIV was assessed with the question: “Do you know the HIV status of your regular partner?” to which participants could respond: “yes, he said he was negative,” “yes, he said he was positive,” “have not discussed this with my partner,” “I don’t have regular partner,” “do not know,” or refuse to answer.

Experiences of violence were assessed with the question: “In the last 12 months, how often have you experienced violence, stigma, and discrimination because you exchange sex?” to which participants could respond “never,” “sometimes,” “often,” “always,” “do not know,” or refuse to answer. The term “violence” was not defined in the question, rather it was left open-ended and the respondent of the survey interpreted the term themselves.

Where participants reported not knowing the correct answer or where they refused to answer, the response was recoded as missing.

Data analysis

RDSAT software version 7¹⁷ was used to estimate the prevalence of HIV and potential risk factors, adjusting for respondents’ network size and network characteristics.

Two logistic models were constructed per outcome variable (i.e. positive HIV status), one univariate (unadjusted), and one multivariate (adjusted) model per outcome variable. P-values less than 0.05, where 95% confidence intervals did not cross 1, were considered statistically significant. Data were analyzed using RStudio.¹⁸

Results

Two thousand one hundred and seventy-four women were interviewed and given an HIV and syphilis test. We estimated that the prevalence of HIV among FSWs in Tajikistan is 2.6% (95% CI: 1.7–3.8%). Further, 2.3% (95% CI: 1.4–3.5%) of FSWs are HIV positive and aware of their status, and 2.0% (95% CI: 1.2–3.1%) are HIV positive and are linked to antiretroviral therapy (Figure 1). 5.7% (95% CI: 4.5–7.4%) of FSWs in Tajikistan had ever had syphilis, and 0.8% (95% CI: 0.4–1.3%) had active syphilis infections.

Figure 1.

HIV cascade of care for female sex workers in Tajikistan. Estimates were derived using RDSAT software. Error bars represent 95% confidence intervals.

51.8% (95% CI: 45.1–60) of FSWs come from Kulob, and 6.7% (95% CI: 4.5–9.4) come from Kurgan-Tebe, both cities in the Khatlon region of Tajikistan, which is also the most populous. 16.1% (10.1–19.6) of the FSWs in Tajikistan are estimated to come from Khorog, Gorno-Badakhshan, a region with less than 3% of the country’s total population.

We estimate that 1.3% (95% CI: 0.6–2.3%) of FSWs in Tajikistan had ever injected drugs, and 12.1% (CI: 10.4–15.1%) had ever had sex with a person who injects drugs. Among sex workers living with HIV in our sample, 40.6% had ever had sex with a person who injects drugs, and 6.3% had ever injected drugs themselves (Tables 1 to 3). Almost half of FSWs reported that they had experienced violence, stigma or discrimination in the preceding 12 months.

Table 1.

Demographic characteristics of the FSW population in Tajikistan; adjusted for network characteristics using RDSAT software.

	Count in IBBS sample	RDSAT-adjusted estimate
City
Dushanbe	500	5.2% (4–7.2)
Khorog	160	16.1% (10.1–19.6)
Khujand	200	8.4% (6.1–11.7)
Kulob	350	51.8% (45.1–60)
Kurgan-Tebe	350	6.7% (4.5–9.4)
Panjikent	214	3.6% (2.9–4.6)
Rudaki	200	5.6% (3.7–7.4)
Vahdat	200	2.5% (1.6–3.8)
Ethnicity
Tajik	1728	83.9% (81.3–85.7)
Uzbek	298	10.3% (9–12.4)
Russian	88	3.1% (2.3–4.2)
Other	60	2.7% (1.4–4.1)
Experiences of violence and discrimination in last 12 months
Never	37	45.2% (42.3–49.4)
Sometimes	330	43.6% (39.5–46.4)
Almost always	950	9% (7.5–11.1)
Always	839	0.9% (0.6–1.5)
Marital status
Single	293	16.3% (13.2–19)
Married	133	11.1% (8.2–13.6)
Divorced	1409	56.5% (53.2–60.8)
Widowed	338	16.1% (13.7–18.6)
HIV-status
Positive	64	2.6% (1.7–3.7)
Negative	2110	97.4% (96.3–98.3)
HIV diagnosed	57	2.3% (1.4–3.5)
Linked to ART	49	2% (1.2–3.2)
Syphilis (ever)	138	5.7% (4.6–7.3)
Syphilis (active)	18	0.8% (0.4–1.3)
Ever had sex with PWID	273	12.1% (10.3–15.1)
Ever injected drug	28	1.3% (0.6–2.3)

FSW: female sex workers; ART: antiretroviral therapy; PWID: people who inject drugs; IBBS: Integrated Bio-Behavioral Survey.

Table 3.

HIV cascade of care by city. Adjusted for network characteristics using RDSAT software.

Region	Proportion HIV positive	Proportion diagnosed	Proportion on ART
Dushanbe	1.4% (0.4–2.7)	1.1% (0.2–2.2)	0.7% (0–1.8%)
Khorog	1.9% (0.0–5.1)	1.2% (0.0–4.4)	1.3% (0.0–4.5)
Khujand	5.7% (2.0–11.4)	4.0% (0.8–8.7)	4.3% (1.1–9.9)
Kulob	1.3% (0.2–2.6)	1.1% (0.1–2.4)	1.2% (0.2–2.6)
Kurgan-Tebe	4.2% (2.0–7.4)	3.8% (1.6–6.7)	3.0% (1.1–5.7)
Panjikent	2.1% (0.0–4.6)	2.2% (0.0–6.1)	2.0% (0.0–5.6)
Rudaki	2.3% (0.2–4.1)	1.6% (0.0–3.2)	2.0% (0.1–3.9%)
Vahdat	3.3% (1.1–6.7)	3.4% (1.2–6.7)	2.4% (0.3–5.7)

ART: antiretroviral therapy.

Among factors associated with HIV infection, having a partner living with HIV was found to be the most strongly associated with HIV in the multivariate model (OR: 42.3; 95% CI: 12.4–152.1). In the multivariate model, the following factors were also associated with HIV: having had a sex partner who was a people who inject drugs (PWID) (OR: 5.2; 95% CI: 2.6–10.2) and the number of casual partners in the last 30 days (OR: 1.4; 95% CI: 1.0–1.8). There were also factors that were statistically significant in the univariate model but not in the multivariate model. These were: having ever injected drugs (OR: 5.8; 95% CI: 1.7–15.5) and having recruited commercial partners via street-based venues (OR 0.5; 95% CI: 0.3–0.8).

Discussion

This study estimates that the prevalence of HIV among FSWs in Tajikistan is 2.6%, and almost half of FSWs in Tajikistan have experienced violence, stigma, or discrimination in the last 12 months. Sex workers are roughly proportionally distributed throughout the country. One outlier is Khorog. The Gorno-Badakhshan region, where Khorog is based, is sparsely populated. Despite having less than 3% of Tajikistan’s total population, it is estimated to be home to over 16% of all Tajikistani sex workers in this study. As the poorest of Tajikistan’s regions and home to a nationally stigmatized ethnic and religious population, it is likely that social and behavioral factors are responsible for this difference.

Injection drug use was found to be highly synergistic with HIV among FSWs in Tajikistan. Among Tajikistani sex workers living with HIV in our sample (n = 64), 40.6% had ever had sex with a person who injects drugs and 6.3% had ever injected drugs themselves. Conversely, among HIV-negative FSWs, 11.7% had ever had sex with a person who injects drugs and only 1.3% had ever injected drugs themselves. Among factors associated with HIV-positive status, having PWID sexual partners had one of the strongest associations with HIV, as an FSW who has had a PWID sexual partner is over five times as likely to have HIV as one who has not (Figure 2). Interestingly, number of partners, or condom use with any type of partners (regular, commercial, or casual) in the past 30 days were not associated with HIV seropositivity.

Figure 2.

Venn diagram of HIV-positive status, history of injection drug use, and history of sex with at least one person who injects drugs. Among Tajikistani sex workers living with HIV in our sample (n = 64), 40.6% had ever had sex with a person who injects drugs and 6.3% had ever injected drugs themselves. Conversely, among HIV-negative female sex workers, only 11.7% had ever had sex with a person who injects drugs and only 1.3% had ever injected drugs themselves. This overlap implies a synergistic relationship between injection drug use and female sex work on the HIV epidemic in Tajikistan.

Compared to other post-Soviet countries, Tajikistan has a relatively low HIV prevalence among FSWs. One smaller study in nearby Uzbekistan suggested an HIV prevalence rate as high as 10% among FSWs.¹² The prevalence of HIV among Tajikistani PWID is also much lower than in Eastern European post-Soviet countries like Ukraine, which had a FSW HIV prevalence of 7.1% in 2014.¹⁹ The prevalence of injecting drug use among Tajikistani FSWs is also much lower than in most other post-Soviet countries (Russia = ~35%; Kazakhstan = ~12%; Kyrgyzstan = ~5%; Uzbekistan = ~7%).¹¹ This may explain the relatively low prevalence of HIV among FSWs as injecting drug use is the main driver of the HIV epidemic in post-Soviet space. It is also likely, however, that our sample was skewed by selection and social desirability biases, as discussed in the limitations section below.

While this study found that HIV prevalence among Tajikistani FSWs is relatively low, there is room to improve HIV testing uptake. Specifically, the finding that stigma, violence, and discrimination against FSWs acts as a barrier to HIV testing uptake should propel the Tajik government to increase legal protection for FSWs, prosecute people who harass FSWs, and increase advocacy and education efforts to destigmatize sex work. Over 60% of FSWs reported at least some instances of stigma, violence, or discrimination; therefore, steps to counter these events should be prioritized to support the health and safety of FSWs and improve HIV outcomes in the country.

Limitations

There are several limitations that reduce the generalizability of this study. First, the gender identity of FSWs in our sample was not critically considered. Transgender women who partake in sex work are at heightened risk for HIV infection.^{20, 21} While all FSWs share many common risks for HIV infection, evidence has shown that transgender women who engage in sex work are at heightened risk of violence, stigma, and discrimination, have a heightened risk of contracting HIV and benefit from specialized interventions ²². Future BBS should recruit both cisgender and transgender women, to make sure that data are collected equitably and that interventions benefit all FSWs, regardless of their gender assignment at birth.

Further, migration has been a defining feature of Tajikistani society since the collapse of the Soviet Union as nearly one million Tajiks live in the Russian Federation at any given time. Previous research has shown that migrant FSWs from Central Asia are at an increased risk of contracting HIV and have worse HIV outcomes.^23,24 As this BBS did not inquire about migration status, we were not able to control for migration in our analysis. Future iterations of the BBS in Tajikistan should consider adding a question regarding migration.

All of the survey data were gathered through self-reporting, leading to a high probability of social desirability bias. This is especially true of questions regarding injection drug use; as heroin is illegal in Tajikistan, it is likely that we significantly underestimate the prevalence of injection drug use among FSWs. Future studies should incorporate hepatitis C testing, which is often a better estimate of injection drug use prevalence than self-reported data.

As cross-sectional survey research cannot explore the diverse and multifaceted experiences of FSWs fully, future qualitative studies²⁵ should be conducted with this population in Tajikistan. Not only are qualitative studies more holistic, but they are also able to counteract some of the biases produced by high levels of mistrust that might have led to low responses in the present study. Examining policy documents relating to FSWs in Tajikistan, as has been done elsewhere in states of the former Soviet Union,²⁶ can also help shed light on the sociopolitical influences of HIV risk for FSWs in the region.

Selection bias may have also resulted in a relatively low HIV prevalence estimate among FSWs in Tajikistan. Given the high level of stigma and mistrust of government authorities, including the common belief that they may be persecuted in case of positive HIV diagnosis, FSWs engaging in high-risk behavior may have been reluctant to participate in the study. As the government conducted the BBS, FSWs at high risk for HIV infection may be especially wary of participating. Additionally, the original survey did not record how many FSWs refused to participate; therefore, we cannot determine the extent to which selection bias influenced the results of this study.

Finally, according to the 2014 sentinel surveillance, the HIV prevalence among sex workers in Tajikistan was 4.7%. The results reported in this study show that the prevalence among sex workers was significantly higher at approximately 10%. The divergent estimates in HIV prevalence may be explained by the findings of the 2019 Morozova et al. study that found the difference in design and sampling methodologies as well as differing methods of recruitment that lead to the recruitment of lower risk individuals into one of the studies which may explain the divergence of the two estimates.

Conclusion

While this study estimates that the HIV prevalence among Tajikistani FSWs is relatively low, it is likely an underestimate due to selection and social desirability biases. To curb the small, but potentially volatile, HIV epidemic among FSWs, the government should consider targeted testing and linkage-to-care efforts to FSWs who inject drugs or have PWID partners as HIV among FSWs in Tajikistan is highly synergistic with injection drug use. At the same time, the prevalence and HIV-related consequences of stigma, violence, and discrimination against FSWs should motivate campaigns to protect FSWs from these experiences.

Table 2.

Univariate and multivariable associations with HIV-positive status (n = 2174).

		Negative	Positive	OR (univariable)	OR (multivariable)
Ever had sex with PWID	No	1729 (98.1)	34 (1.9)	—	—
	Yes	247 (90.5)	26 (9.5)	5.4 (3.1–9.1, p < 0.001)	5.2 (2.6–10.2, p < 0.001)
	Missing	134 (97.1)	4 (2.9)
Ever injected drugs	No	2081 (97.2)	60 (2.8)	—	—
	Yes	24 (85.7)	4 (14.3)	5.8 (1.7–15.5, p = 0.002)	0.79 (0.1–3.5, p = 0.769)
	Missing	5 (100.0)
Recruits commercial partners via street-based venues in the last 12 months	No	896 (95.9)	38 (4.1)	—	—
	Yes	1214 (97.9)	26 (2.1)	0.5 (0.3–0.8, p = 0.008)	0.6 (0.3–1.1, p = 0.108)
Number of commercial partners in the past 7 days	Mean (SD)	4.3 (3.8)	3.7 (2.9)	0.9 (0.8–1.0, p = 0.156)	0.9 (0.7–1.0, p = 0.131)
Number of regular partners in last 30 days	Mean (SD)	0.9 (0.9)	0.9 (1.1)	1.0 (0.8–1.3, p = 0.839)	1.1 (0.8–1.6, p = 0.475)
Number of casual partners in last 30 days	Mean (SD)	0.5 (1.2)	0.8 (2.0)	1.2 (1.0–1.3, p = 0.049)	1.4 (1.0–1.8, p = 0.019)
Number of commercial partners in last 30 days	Mean (SD)	11.0 (8.9)	9.0 (6.7)	1.0 (0.9–1.0, p = 0.078)	1.0 (1.0–1.1, p = 0.835)
Frequency of condom use with regular partners in last 30 days	Almost always	37 (94.9)	2 (5.1)	—	—
	Always	342 (95.0)	18 (5.0)	1.0 (0.3–6.3, p = 0.972)	1.3 (0.2–15.4, p = 0.780)
	Half of the time	168 (97.7)	4 (2.3)	0.4 (0.1–3.3, p = 0.354)	1.2 (0.2–16.0, p = 0.861)
	Never	542 (98.0)	11 (2.0)	0.4 (0.1–2.5, p = 0.213)	0.7 (0.1–8.1, p = 0.732)
	No such partner	686 (96.9)	22 (3.1)	0.6 (0.2–3.8, p = 0.491)	2.2 (0.3–26.5, p = 0.476)
	Sometimes	296 (98.3)	5 (1.7)	0.3 (0.1–2.2, p = 0.173)	0.7 (0.1–8.8, p = 0.745)
	Missing	39 (95.1)	2 (4.9)
Frequency of condom use with casual partners in last 30 days	Almost always	55 (94.8)	3 (5.2)	—	—
	Always	218 (97.3)	6 (2.7)	0.5 (0.1–2.5, p = 0.344)	0.3 (0.0–3.2, p = 0.283)
	Half of the time	73 (98.6)	1 (1.4)	0.3 (0.0–2.0, p = 0.237)	0.3 (0.0–4.6, p = 0.385)
	Never	16 (94.1)	1 (5.9)	1.2 (0.1–9.7, p = 0.909)	3.7 (0.1–78.4, p = 0.402)
	No such partner	1582 (97.2)	46 (2.8)	0.5 (0.2–2.2, p = 0.304)	0.8 (0.1–9.7, p = 0.861)
	Sometimes	127 (96.2)	5 (3.8)	0.7 (0.2–3.6, p = 0.663)	1.2 (0.1–13.7, p = 0.900)
	Missing	39 (95.1)	2 (4.9)	0.9 (0.1–5.9, p = 0.947)	0.00 (NA-Inf, p = 0.997)
Frequency of condom use with commercial partners in last 30 days	Almost always	126 (96.9)	4 (3.1)	—	—
	Always	1155 (96.3)	44 (3.7)	1.2 (0.5–4.0, p = 0.731)	2.2 (0.6–11.2, p = 0.275)
	Half of the time	254 (98.1)	5 (1.9)	0.6 (0.2–2.5, p = 0.482)	0.8 (0.1–5.2, p = 0.829)
	Never	70 (97.2)	2 (2.8)	0.9 (0.1–4.7, p = 0.905)	1.7 (0.2–13.7, p = 0.628)
	No such partner	18 (100.0)		0.0 (0.0–11786.5, p = 0.981)	0.0 (0.0–524041528.0, p = 0.988)
	Sometimes	448 (98.5)	7 (1.5)	0.5 (0.2–1.9, p = 0.264)	0.7 (0.1–4.5, p = 0.719)
	Missing	39 (95.1)	2 (4.9)	1.6 (0.2–8.6, p = 0.588)	0.00 (NA-Inf, p = 0.997)
HIV-positive regular partner	No	637 (98.2)	12 (1.8)
	Yes	11 (47.8)	12 (52.2)	57.9 (21.6–161.6, p < 0.001)	42.3 (12.4–152.1, p < 0.001)
	Unknown	1462 (97.3)	40 (2.7)	1.5 (0.8–2.9, p = 0.262)	2.00 (0.9–4.7, p = 0.095)
Experience of violence in the last 12 months	Always	35 (94.6)	2 (5.4)	—	—
	Never	807 (96.2)	32 (3.8)	0.7 (0.2–4.4, p = 0.626)	0.9 (0.2–8.7, p = 0.925)
	Often	326 (98.8)	4 (1.2)	0.2 (0.0–1.6, p = 0.082)	0.3 (0.0–3.1, p = 0.227)
	Sometimes	925 (97.4)	25 (2.6)	0.5 (0.1–3.0, p = 0.321)	0.8 (0.2–7.7, p = 0.815)
	Missing	17 (94.4)	1 (5.6)	1.0 (0.1–11.5, p = 0.982)	2.4 (0.1–37.9, p = 0.529)

Bolding indicates statistical significance (p <0.05)

List of Abbreviations

BBS

Bio-behavioral survey

FSW

female sex worker

MSM

men who have sex with men

PWID

People who inject drugs

RDS:

respondent-driven sampling

Data accessibility statement

The data that support the findings of this study are the property of the Ministry of Health of Tajikistan and are not publicly available.