HIV Risks, Testing, and Treatment in the Former Soviet Union: Challenges and Future Directions in Research and Methodology

Background The dissolution of the USSR resulted in independence for constituent republics but left them battling an unstable economic environment and healthcare. Increases in injection drug use, prostitution, and migration were all widespread responses to this transition and have contributed to the emergence of an HIV epidemic in the countries of former Soviet Union. Researchers have begun to identify the risks of HIV infection as well as the barriers to HIV testing and treatment in the former Soviet Union. Significant methodological challenges have arisen and need to be addressed. The objective of this review is to determine common threads in HIV research in the former Soviet Union and provide useful recommendations for future research studies. Methods In this systematic review of the literature, Pubmed was searched for English-language studies using the key search terms “HIV”, “AIDS”, “human immunodeficiency virus”, “acquired immune deficiency syndrome”, “Central Asia”, “Kazakhstan”, “Kyrgyzstan”, “Uzbekistan”, “Tajikistan”, “Turkmenistan”, “Russia”, “Ukraine”, “Armenia”, “Azerbaijan”, and “Georgia”. Studies were evaluated against eligibility criteria for inclusion. Results Thirty-nine studies were identified across the two main topic areas of HIV risk and barriers to testing and treatment, themes subsequently referred to as “risk” and “barriers”. Study design was predominantly cross-sectional. The most frequently used sampling methods were peer-to-peer and non-probabilistic sampling. The most frequently reported risks were condom misuse, risky intercourse, and unsafe practices among injection drug users. Common barriers to testing included that testing was inconvenient, and that results would not remain confidential. Frequent barriers to treatment were based on a distrust in the treatment system. Conclusion The findings of this review reveal methodological limitations that span the existing studies. Small sample size, cross-sectional design, and non-probabilistic sampling methods were frequently reported limitations. Future work is needed to examine barriers to testing and treatment as well as longitudinal studies on HIV risk over time in most-at-risk populations.


Driving forces behind the epidemic
Efficacy of HIV testing and access to treatment have been evaluated at both the individual and societal levels in many regions of the world, but studies of how vulnerable populations-especially IDU-access these services in the FSU have been limited. 9 In recent years, approaches to HIV/AIDS worldwide have broadened to focus not only on individual risk-taking behavior, but also on the environmental and societal factors that influence risky behavior and use of health services. 10,11 Most-at-risk populations-IDU, migrant workers, and commercial sex workers (CSW)-are particularly vulnerable without access to HIV testing, treatment, and prevention resources. They also are among the FSU's least studied groups. 10 Initiating public health research around highly stigmatized populations, however, has proven to be especially challenging in the FSU. 12 In the most extreme case, this stigma has resulted in little to no research on HIV in Turkmenistan, where it is unlawful to diagnose or report a patient with HIV. 13,14 This review makes little reference to the HIV situation in Turkmenistan, where there is limited national data.
Furthermore, the body of literature on HIV in the FSU is just beginning to take shape. However, the eventual goal of building a substantive body of literature around the causes of and barriers to reduction of HIV is to identify why at-risk populations are more vulnerable to HIV, as well as the barriers underlying suboptimal access to testing and treatment. 15 Knowing these barriers, programs can be redirected and new initiatives prepared. Furthermore, in order to conduct more effective research, challenges and limitations of past studies must be discussed. The primary aim of this review is to systematically evaluate the literature and provide a concise review of research and methodological challenges to-date on the HIV epidemic in the FSU. A second goal of this analysis is to provide guiding factors for the planning and implementation of future studies for the design of more effective testing and treatment programs in the region.

Search methods
Search terms used on PubMed included "HIV", "AIDS", "human immunodeficiency virus", "acquired immune deficiency syndrome", "Central Asia", "Kazakhstan", "Kyrgyzstan" "Uzbekistan", "Tajikistan", "Turkmenistan", "Russia", "Ukraine", "Armenia", "Azerbaijan", and "Georgia" (Countries were chosen on the basis of availability of research literature). Bibliographies of relevant articles and reviews were scanned for further studies. primary outcomes. These overarching themes are subsequently referred to as "risk" and "barriers". Studies highlighting risks and barriers were chosen for review because HIV prevention and treatment efforts cannot be realized unless the underlying risks are understood. 16 The flow of study selection is illustrated in Figure 1. The search engine PubMed was used in the collection of studies for this review. Studies deemed irrelevant were either only tangentially related to HIV in the former Soviet Union or contained the key words but were not answering a research question that contributed to the aims of the review.  Table 1 shows the main features of each study; Tables 2a, 2b, and 2c summarize the most frequently reported risks, barriers to testing, and barriers to treatment, respectively. Table 1. Summary of studies examining risks of HIV infection, barriers to HIV treatment and challenges in HIV prevention.
Pertinent characteristics include location, study aim, design, population, sampling method(s), and sample size

Study design
The majority of studies employed a crosssectional design (n=36). Of these studies, 10 also obtained biological samples to determine HIV status of participants. The only longitudinal study was from Georgia where investigators and implemented both HIV testing and biobehavioral surveys at two different time points, three years apart from each other. 23 However, the study reported that the sample size was insufficient to power a comparison between the two time points. Therefore, small sample size among a marginalized population was a key limiting factor.

Study population
IDU were the target population in 17 studies (44%) and were conducted in each country except Turkmenistan. Other major populations were migrant workers (n=8), female and CSW (n=7).

Study limitations
The most frequently reported methodological challenges were cross-sectional study design (n=12), inability to obtain a representative sample (n=11), use of self-report (n=11), sub-optimal participant recruiting procedures (n=8), and/or a small sample size (n=6). Further methodological limitations included: data were found not to be generalizable outside of the country in which the research was conducted (n=5), specifically having used purposive or snowball sampling to recruit participants (n=5), low participation rates (including not This journal is published by the University Library System of the University of Pittsburgh as part of its D-Scribe Digital Publishing Program and is cosponsored by the University of Pittsburgh Press. having obtained data on specific groups that declined to participate) (n=4), translation issues and cultural misunderstanding of qualitative data (n=4), and likely underreporting of risky, illegal, and/or stigmatized behaviors in surveys and interviews (n=4).

Risk factors for HIV infection
The most frequently reported categories of risks were condom misuse (n=9), risky intercourse (n=9), unsafe injection practices among IDU (n=8), and spread of infection through people who inject drugs (n=8) ( Table  2). Additional groups of risk factors included migration challenges (n=6), low HIV/AIDS knowledge (n=4), and a history of STI (n=4).

Barriers to HIV testing
The most prominent barriers to testing for HIV status included the perception that it was shameful to test for HIV (n=2), that testing was inconvenient (n=2), and that test results would not be held confidential (n=2) ( Table 3). Table 3. Barriers to HIV testing as determined by the reviewed studies detailed aspects of each category and supporting studies

Barriers to HIV treatment
The most frequently reported barriers to obtaining treatment for HIV were based on a distrust in the treatment system and experience with the lack of efficiency in the structure of the treatment system (Table  4). Specifically, the barriers included a fear of disclosure of treatment status (n=6), an inefficient and ineffective treatment structure (n=6), difficulty in registering for and/or being accepted into a treatment facility (n=5), and difficulty in accessing treatment facilities (n=5). Table 4. Barriers to HIV treatment as determined by the reviewed studies with detailed aspects of each and a list of the supporting studies for each category

Discussion
This is, to our knowledge, the first systematic review of the current body of research spanning the stages of HIV infection in the FSU, from risk/infection through testing and treatment. The results reveal several important areas in which the current state of research and knowledge is incomplete due to methodological limitations of many studies. As shown in Tables 3 and 4, barriers to testing and treatment remain strong among atrisk populations (IDU, CSW, and migrant workers) but remain inadequately researched when compared to the number of studies examining risk factors for HIV infection. Among the reasons for this include sample recruitment challenges, and other methodological challenges, which are further discussed below.

Most-at-risk populations
CSW, men who have sex with men (MSM), IDU, and migrant workers have been found to be key players in the spread of HIV. 1,4,7 However, they are the most stigmatized and marginalized groups and have very little access to HIV treatment. 55 Further, economic, social, and institutional factors in the region can be linked to the spread of HIV among these groups.

Migrant workers
Migration between Russia, the Caucuses, and Central Asia has been observed as a driver of the epidemic. 1,11,56 Many Tajik and Kyrgyz migrants travel through Kazakhstan and into Russia to find work. As might be expected, extensive travel often puts them at risk. 56 Often, financially compromised and separated from family, migrants have been shown to engage in behaviors that increase the risk of HIV transmission. 57 Their financial and legal status in the host country make it extremely difficult to access medical care should HIV be suspected or treatment needed. 9

Injection drug users
Throughout the FSU, the IDU population has been growing and is associated with harmful drug use and co-infections of TB and HCV. 4 The example of Central Asia demonstrates both behavioral, economic, and geographical factors, among others, at play in the concentration of HIV among IDU. [58][59][60] The trafficking of opium out of Afghanistan results in large amounts of the drug being transported through Kazakhstan, fueling rapid growth of the nation's population of IDU. Additionally, Kazakhstan and other Central Asian nations' location at the centers of labor migration routes compound the effect of drug trafficking: when migratory patterns considerably overlapped with drug trafficking routes, the number of cases among IDU increased five-fold in the 13 years following Kazakhstan's independence. 61

Commercial sex workers
This group consists of both men and women who engage in sex work for compensation and suffer tremendously from the stigma that accompanies their work. Among this group, female sex workers are more stigmatized than their male counterparts. Those who also inject drugs experience a form of double jeopardy. 62

Methodological challenges
One, studies have been conducted in many, but not all, of the countries of the FSU. Research in the field of HIV/AIDS is particularly limited in Ukraine and Georgia, along with being severely limited in Turkmenistan. While the number of studies from Russia and Kazakhstan, for instance, are relatively numerous, they cannot necessarily be generalized to other FSU countries. This could be explained by the simultaneous similar-and-different nature of the countries: on one hand, they shared some common elements of their political, social, and economic history for most of the 20 th century. On the other hand, each country has its own history and ethno-cultural fabric, which is likely to uniquely affect the mentality and psychology of its people. Therefore, studying the risks and barriers within every FSU country is necessary in order to help each one best prepare and implement an approach to ameliorate the HIV/AIDS epidemic. Two, many of the studies used a cross-sectional design in collecting their data. It is difficult to make statements of causation from such designs. More longitudinal designs are needed to study the range of factors for any given at-risk group. For example, migrant workers may need to be studied throughout the migration process to evaluate the stage of the migration experience that introduces the most vulnerability to exhibit HIV risk behaviors.
Three, many of the studies employed nonprobabilistic sampling. It is difficult to know the probability with which the target population has been represented in the sample when using a non-probabilistic method of sampling. Such methods that have been employed in this review's studies include convenience sampling, purposive sampling, and snowball sampling. The studies required participation by individuals who exhibit illicit, illegal, or stigmatized behaviors and are, therefore, socially marginalized. Convenience sampling, purposive sampling, and snowball sampling were used to gain access to such populations. Although migrant workers, IDU and CSW are understandably difficult to access and representatively sample, studies could be designed in a more rigorous way that takes into account these limitations of working with hard-to-sample populations that are hidden and lack most formal forms of rosters or lists of documentation, from which probabilistic samples could be obtained. other parts of the world, including Thailand and Australia, it was interesting to find that many of the studies did not demonstrate a need for purposive or respondent-driven sampling techniques, for instance. Instead, the researchers often approached CSW, in many cases, in testing or treatment facilities. 63,64 This may likely indicate a difference in the difficulties inherent in recruiting marginalized populations in the FSU, when compared to the same task in other parts of the world. 65 Four, many of the studies reported that it was likely that subjects may have underreported stigmatized, illegal, or risky behaviors in self-report questionnaires and interviews. Although not verified, it was an observation made by researchers who were likely aware of the stigmatized nature of most-at-risk populations. For this reason, it was thought to be likely, given that study participants may have distrusted the researchers and feared that the interview results deemed confidential would be released to the police. 66 Furthermore, data obtained from self-report can be subject to the "socialdesirability bias," by which a participant may answer questions in a certain manner in order to portray themselves as lawful and socially acceptable. While this insight is helpful in interpreting the data, it sill reveals that the data acquired are not thoroughly accounting for the range and prevalence of behaviors that put subjects at risk for HIV infection or pose barriers to testing and treatment.

Limitations
There are several limitations of the review that must be noted. First, while most of the studies were generally accessible in English, several studies were available only in Russian (n=8) or full-text was inaccessible. Without the ability to identify and retrieve all relevant studies, the review's scope may diminish from the ideal. Second, the study of HIV is a relatively new area of epidemiologic and public health focus in the FSU, and thus the number of relevant studies is limited, resulting in 39 eligible studies for review. Third, the methodological limitations that were extracted from the study manuscripts and tabulated above were based on what the authors had listed in their own evaluation of study limitations and/or from what was available in the methodological descriptions of the studies. For instance, purposive sampling-as a methodological limitation in a study-was determined from the methodological descriptions and/or from the discussion of limitations provided by the study's authors. This means that certain methodological limitations deemed infrequent in this review-especially "distrust of researchers" (3% of studies reviewed)-should not be viewed as certainly infrequent. Many of the studies may have suffered from participant distrust, which may have either gone unnoticed and had an effect on data or sampling outcomes or have been noticed but not reported in the manuscript. Participant distrust of the researchers can have an effect on many parts of a study that involve mostat-risk populations: participation rate, sample size, and underreporting can all be affected, but it is important to note in FSU-based studies when distrust occurs in order to help determine ways to improve the relationship between most-at-risk populations and researchers.

HIV research in the former Soviet Union
Today, the epidemic grows as risky behavior continues, and prevention and treatment programs face difficulties gaining a foothold in the still-transitioning atmosphere of the FSU. However, the findings of this review reveal a few particular ways in which the current state of knowledge is incomplete as a result of methodological limitations of many of the existing studies. For instance, while 20 studies reported risk factors for HIV infection, only seven of the 39 studies reported barriers to HIV testing, and 9 studies reported barriers to treatment. These numbers illustrate where the bulk of the research has been conducted in the HIV infection pathway (risks factors and infection, testing, and treatment) in the FSU. One reason for this finding may be that there are many more individuals at risk for or infected with HIV than there are individuals who have sought testing and/or treatment. Given that small sample size has been a limitation and concern among many of the reviewed studies, it may be a reason for the relative scarcity of studies on barriers to testing and treatment when compared to studies on the risk factors for HIV infection.

Recommendations for future studies
Going forward, energy and resources would be best spent on research to study the barriers to getting tested and treated for HIV. Of the studies included, the number of which that looked barriers to testing and treatment was minimal compared to what was aimed at studying the risks of HIV infection. Future research would include a combination of studies that address the described methodological challenges and one or more attempts at meta-analysis of the data from thematically aligned studies. Efforts to apply the results from the above-mentioned research would assist in improving existing HIV programs and advising the development of new ones in the FSU. 67