Returning for HIV Test Results: A Systematic Review of Barriers and Facilitators

This systematic review aims to identify factors that facilitate or hinder the return for HIV test results. Four electronic databases were searched. Two independent reviewers selected eligible publications based on inclusion/exclusion criteria. Quantitative studies published since 1985 were included. Thirty-six studies were included in the final review. Individual level barriers included sociodemographic characteristics, such as being a male, of young age and low education level, risk behaviours such as injecting drugs, having multiple sexual partners, and psychosocial factors. Older age, higher education level, being a woman, having high self-esteem, having coping skills, and holding insurance coverage were identified as facilitators. Interpersonal barriers and facilitators were linked to risk behaviours of sexual partners. Contextual barriers included essentially the HIV testing center and its characteristics. This review identified the most important factors that need to be addressed to ensure that people return for their HIV test results.


Introduction
HIV testing and counselling (HTC) is the cornerstone of treatment, care, and prevention [1,2]. It is particularly through HTC that the ambitious goal of 90, 90, 90 (90% of all people living with HIV will know their HIV status; 90% of all people with diagnosed HIV infection will receive sustained antiretroviral therapy; 90% of all people receiving antiretroviral therapy will have viral suppression), by 2020 [3] can be reached. Previously, most efforts were focused on voluntary counselling and testing (VCT) as the primary means of providing testing and encouraging people to become aware of their HIV status [2]. However, coverage remains low and many infected persons in both developed and developing countries remain undiagnosed. Despite the availability of rapid test with the possibility to have the results approximately in 20 to 30 minutes, in some contexts particularly, there are many who get tested but fail to return for their results [4][5][6]. For example, in the USA, data from HIV testing performed at publicly funded counselling and testing sites using conventional HIV enzyme immunoassay (EIA) testing from 1999 through 2002 found that 19% to 22% of people with positive preliminary HIV tests did not return for their test results [6]. In 2009, a survey conducted in 12 Sub-Saharan Africa countries with high HIV prevalence showed that only 10% of women and 12% of men were tested and received their test results [7]. In an evaluation of five years of routine program data in Vietnam, Hong et al. found a failure to return (FTR) rate of 3.5% [8]. In a study of female sex workers in China, Xu et al. found a FTR rate of 47.1% [9]. More recently, in France, Laanani et al. (2015) and Pahlavan et al. (2015), respectively, found a FTR rate of 6.5% in a study 2 International Scholarly Research Notices conducted in a free and anonymous screening center [10] and 14.5% in an HIV-positive population [11].
Identifying and targeting these people may improve the return rate for VCT and the proportion of individuals who are aware of their status. Therefore, the objective of this systematic review is to identify the factors that prevent people who are tested for HIV from returning for their results or facilitate their doing so.

Methods
This systematic review examines barriers and facilitators associated with returning for HIV test results in various types of populations and settings. The outcome variable, returning for HIV testing results, is dichotomous. From this point of view, some studies concerned factors associated with returning for HIV test results, while others focused on failure to return (FTR). This review was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines [44].

Search Strategy.
A comprehensive search strategy was developed to identify studies published between January 1985 (introduction of HIV tests) and June 2015. Four electronic databases were searched (PubMed/Biomed Central/Medline, Embase, PsycINFO, and Web of Science), combining terms related to HIV, counselling/testing, and return/failure to return. Retrieved references were imported into Endnote X7, and then duplicates were removed. The detailed search strategy is available upon request.

Study Selection.
All identified records ( = 3,019) were initially screened by two independent investigators and verified by a third researcher. Eligible studies had to meet the following criteria: (1) be an original research study; (2) be written in English or in French; (3) report adolescents, adults, pregnant women, men who have sex with men (MSM), injecting drug users (IDUs), or female sex workers (FSWs); (4) include participants undergoing HIV tests; (5) use a quantitative method to assess return or failure to return for HIV test results; and (6) report a statistical association between a potential predictor/correlate and return or FTR. Study designs of interest were cross-sectional and longitudinal. No geographical restrictions were applied. The reference lists of the relevant articles were also reviewed for additional publications.
A short list of records was prepared and the full text reviewed independently by two authors. Citations that were clearly irrelevant were excluded. Uncertainties and disagreements about inclusion were resolved through discussion involving both investigators (see Figure 1 for flowchart of systematic review).

Data Extraction.
Two authors independently extracted data from each study that fulfilled the inclusion criteria using a standard form. Study characteristics (name of the first author, year of publication, country in which the study was conducted, study design, sampling approach, participating characteristics, and HIV testing procedure) as well as key findings related to factors associated with return or failure to return for HIV test results were extracted. Any factors analyzed associated with FTR or return for HIV test results were listed, and the results of multivariate statistical tests for association (odds ratio) were noted. For studies where a multivariate statistical test was not done, the results of bivariate analyses were noted. When the result of the measure of association in multivariate analysis was not significant and not reported by authors, the factor was not considered in the synthesis.

Quality
Assessment. The Newcastle-Ottawa Scale (NOS) for cohort studies and an adapted form of the Newcastle-Ottawa cohort scale for cross-sectional studies were used to assess methodological quality. NOS is a tool for assessing the quality of nonrandomized studies to be used in a systematic review [45]. Each study is judged with a "star system" on three points: the selection of study groups, the comparability of the groups, and the ascertaining of exposure or outcome. Studies for which at least five out of nine items on the NOS were deemed satisfactory and in which appropriate statistical analysis (e.g., multivariate controlling for confounders) was conducted were considered to be of sufficient methodological International Scholarly Research Notices 3 quality and included in the review (maximum score of 10 for cross-sectional studies and 9 for cohort studies). At each stage of the quality assessment, the reviewers discussed among themselves until a consensus was reached on which studies to include.

Data Synthesis and Combined Effect Sizes Associated with
Return. Factors associated with either return or failure to return may be arranged into barriers and facilitators inspired by the Socioecological Model (SEM), which is a framework to examine the multiple effects and interrelatedness of environmental, contextual, and social factors on individual behaviour [46,47]. Recognizing that most public health challenges are too complex to be adequately understood and addressed from single level analyses, the SEM includes a more comprehensive approach that integrates multiple levels of influence to impact health behaviour and ultimately health outcomes. These levels of influence include intrapersonal and interpersonal factors, organizational factors, and structural factors. In this review, due to their small number, organizational and structural factors were grouped as contextual factors.
Since we anticipate a potential variability of the methodology (e.g., measures of studied factors) across the included studies, we used a random-effects model based on the inverse variance method to estimate the pooled odds ratio (OR) for each factor potentially associated with returning for HIV test results and its 95% confidence interval (CI) [48,49]. The Higgins's 2 statistic was used to quantify the percentage of the variability in individual effect size estimates which is attributable to the heterogeneity [50,51]. This heterogeneity was tested using a chi-squared test [50,51]. Moreover, we performed sensitivity analysis by removing the included studies from the pooled size estimation one at a time. These analyses allowed us to explore the individual contribution of each study to the heterogeneity in the meta-analysis. When we could not explain the heterogeneity, we have interpreted the pooled effect size estimates with caution because these effect sizes would be explained by other factors, which were not taken into account in our analyses. A value of less than 0.05 was considered statistically significant. Analyses were performed in Review Manager (version 5.3).

Study Selection.
The primary search strategy identified 3,019 potentially relevant citations. After the removal of duplicates and the initial title and abstracts screening, 60 citations were kept for the full-text review. Studies were excluded if they did not report quantitative results ( = 1) or just reported the rate of return or failure to return (FTR) without assessment of associated factors ( = 23). The remaining 36 studies were appraised for their methodological quality and included in the analysis. No study was excluded on the basis of quality assessment. A flow chart illustrating the selection process is shown in Figure 1. Table 1 provides a brief overview of the key characteristics of the included studies. Of the 36 included studies, 10 were longitudinal cohort studies and 26

Study Characteristics.
were cross-sectional studies. Seven of the studies were carried out in Sub-Saharan Africa; seventeen in the USA; three in Australia; six in Asia; and one in Brazil. Populations under study were diverse, including general population ( = 16), pregnant women ( = 5), injecting drug users ( = 3), men who have sex with men ( = 3), high-risk heterosexual individuals ( = 2), HIV-positive individuals ( = 2), factory workers ( = 2), individuals with psychiatric problems ( = 2), adolescents ( = 1), HIV-negative individuals ( = 1), and female sex workers ( = 1). The outcome of interest was dichotomous, with 20 studies focused on failure to return and 16 on return for HIV test results.
Level of education was reported in nine studies. In 4 studies [15,17,26,30], it was reported as a barrier to returning for HIV test results, especially for people with no education or a low level of education. In 3 studies [9, 34, 39], it was     -- * * * * * * *     reported as a facilitator for those with a medium or high level of education.
In studies with a mixed population (women and men) when sexual orientation was reported ( = 4), being heterosexual or bisexual appeared as a barrier to returning for HIV test results [11,36,37].
Marital status was reported in six studies. In these studies, being married or living in a couple [8] and being a widower [8] emerged as facilitators of a return for results.

Risk Behaviours.
Several risk behaviours were positively or negatively associated with a return for results. The number of sexual partners during the last 6 to 12 months was reported in 8 studies. In 5 of these studies [8,10,26,27,37], having more than 5 sexual partners was reported as a barrier to returning for HIV test results. Otherwise, having a single sexual partner during the last 6 to 12 months [22,37] was not significantly associated with a return for results. Using a condom has been reported as both a barrier [30] and as a facilitator [26,41], but in most cases, the association was not significant [14,37,39]. Having a history of STIs was reported in nine studies. However, it was reported equally as a barrier [16,19,29,41] or a facilitator [13,22,25,31] as regards a return for HIV test results.

Perceived Risk.
Perceived risk has been reported in nine studies. In those studies, having low perceived risk [26] ( = 1) and not seeing oneself at risk [10, 27] ( = 2) was reported as a barrier to a returning for test results. However, this result is somewhat controversial because one of these two studies showed that having high-perceived risk [10] was a barrier to a return for results, and having a medium perceived risk was reported at the same time as a barrier in one study [14] and as a facilitator in another [33]. In addition, the association was insignificant for 4 studies [10,35,38,39].

Psychosocial Factors.
The association between the return for HIV test results and psychosocial factors showed divergent results ( = 5). For instance, not believing in selfprevention from HIV [26], believing that HIV can be cured [29], and thinking that a medical follow-up can improve the course of HIV [29] were reported as barriers to a return for test results. However, having high self-esteem [33] and positive coping skills [33] appeared as a facilitator for a return for results. Feeling anxious about HIV was reported as a barrier [15,42] to a return for HIV test results.

Health Coverage.
The association between the existence of health coverage and a return for results was studied in two articles. Having health coverage [24] appeared as a facilitator and not having health coverage as a barrier [31]. , and having a partner who did not test [28] ( = 1) were all reported as a barrier to returning for results. Domestic violence (abused by a partner) [28] and rape [16,29] were reported as barriers to a return for test results in three studies.
The association between a return for results and the availability of a social network has been studied in six studies. On the one hand, it appears that having one or more gay friends [37], having a counsellor [40], or knowing someone infected with HIV [35] are barriers to a return for results. On the other hand, having social support [40] (friends) and lacking a family confidant [38] were reported as facilitators of a return for test results.

Contextual Level.
A negative association was found between the return for HIV test results and having a confidential test in one study [12]. The same negative association was found when the testing was done in facilities such as family planning clinics [27,36] [46] ( = 1), or having tested in a center located in the same city as the treatment center [46] ( = 1) and not having pretest counselling, emerged as facilitators of a return for results.

Combined Effect Sizes of Factors Associated with a Return
for HIV Test Results. The pool estimates and sensitivity analysis of the association of the return for HIV test results with certain factors, including gender (men versus women) and race (black versus white) for studies conducted in the USA, injection drug use (no versus yes), HIV test results (positive versus negative), and HIV testing history (no versus yes), are shown in Table 3. The combined analysis showed that being female [8, 18-20, 22, 30, 34, 35] is significantly associated with a return for results (OR = 0.86, 95% CI = 0.77-0.96) when studies with specific population (MSM, HIV negative, pregnant women) are excluded. In the studies from the USA, black people tend to return less frequently for their results than white people (OR = 0.76, 95% CI = 0.64-0.90) [12, 18, 22, 27, 34-36, 38, 39]. There is no significant association between returning for test results and HIV test results or HIV testing history. Finally, the association between injection drug use and returning for test results was significant (OR = 0.85, 95% CI = 0.75-0.96) [8,12,14,22,41] when only the general population was considered. Thus, being an IUD appears as a barrier for returning for HIV test results.

Discussion
The objective of this review was to report the factors that were statistically associated with the return for HIV test results in different studies, regardless of the target population, the HIV test method used (standard or rapid tests), the waiting time for results, or the country. Despite these different contexts, periods, and populations, the majority of studies considered the same factors. The vast majority of reported factors are found at the individual level (sociodemographic characteristics, risk behaviours, individual risk perception, and test results). Very few studies have reported contextual factors, such as organizational factors, policies, economic factors, or social factors.
The differences in statistical analysis (classification, categorization, and reference group) introduced a great deal of heterogeneity with respect to the studies. Thus, it was not possible to combine effect sizes for all factors. The factors not included in the meta-analysis were grouped into barriers and facilitators based on their statistical association with the dependent variable (return for HIV test results).
Although the factors have been grouped into categories according to the ecological model, it is important to specify, in accordance with the socioecological approach, that the categories are not exclusive but rather influence each other.

Sociodemographic Factors.
Age and level of education acted both as barriers and facilitators. However, the trend indicates firstly that young people and individuals with a low level of education were less likely to return for their results.
Indeed, there is evidence that young people are often less informed about HIV and also exhibited a lower rate of HIV testing than adults [45,46,52]. They are unaware of their risk behaviours and are less likely to return for their test results. Conversely, individuals with higher levels of education can better understand the importance of screening [44,53] and are more likely to return for their results.

Risk Behaviours.
The literature has shown an association between the return for test results and risk behaviours [52,54]. In fact, people who display risk behaviours can also develop fear and anxiety with respect to knowing their test results. In these circumstances, they are less likely to return for their results even if they had the courage to get tested. Thus, in this review, the positive test result, injection drug use, a high number of sexual partners, getting paid to have sex, and having symptoms of sexually transmitted infections (STIs) at a testing visit were reported as barriers to returning for test results.

Perceived Risk.
Studies that have examined the association between perceived risk and a return for test results are sometimes contradictory. Indeed, some studies have shown that people who have a high-perceived risk of contracting HIV were more likely to return for their results [33,55]. Other studies have shown that people with a low perceived risk do not return for their results [10,14,26,27]. This second situation might be explained by the fact that many people at high risk of contracting HIV do not perceive themselves as at risk [56][57][58]. Therefore, they do not see the importance of returning for their results and knowing their status. This is why it is recommended that the education of individuals be intensified in order to foster a high and precise perception of risk.

Interpersonal Factors.
The sexual partner's risk behaviours were the most frequent group of factors influencing a return for test results. Furthermore, having social support, having an HIV infected partner, or being a client of a sex worker have been reported as factors that encourage people to return for their results. In fact, having sex with a highrisk person might increase the perceived risk, which leads the exposed person to learn his or her HIV status [59]. On the other hand, the family and social network provide social support and reinforce social norms [60] that might encourage a return for results. In contrast, being a member of a social network of people at risk, such as a partner of a sex worker, of an alcoholic, or of a drug user, having gay friends, or knowing someone infected by HIV tend to hinder a return for results. These risk groups often experience discrimination and stigmatization [1,61]. Therefore, they are less likely to get tested, to return for their results, to disclose their HIV status to others, to adopt preventive behaviours, or to access treatment services, care, and support. Finally, domestic violence (intimate partner violence) and sexual assault also hinder a return for results. Despite the implementation of strategies that enable women to get tested at opportune moments such as during pregnancy or childbirth, domestic violence remains a barrier for the entire testing process [62][63][64]. First, the female victim of sexual violence is afraid to return for her results and know her HIV status because she is afraid of being rejected by her partner who can blame her for having tested without his consent and for being responsible for his contamination in the case of a positive result [62,64,65]. Second, the feeling of guilt and fear of victimization and stigmatization experienced by a raped woman can hinder her return for results even if she could be tested [62,66,67].

Contextual
Factors. The HIV testing center and its characteristics were the most frequent contextual factors reported in different studies. Getting tested in most of the sites appeared as a barrier to returning for test results. One reason may be the type of screening test offered at these sites. Many of the studies in this review were conducted before the use of rapid tests. Recently, several HIV testing centers in developed countries and in developing countries have reported an increase in the demand for testing, the proportion of people who received posttest counselling, and the knowledge of status following the introduction of rapid tests [68][69][70][71][72]. Other studies also showed that clients prefer the centers where they can receive their results without delay on the same day [73][74][75]. However, it is also reported that when the testing center is linked with the treatment, the pretest counselling is done well, and there is distribution of condoms, this set of factors encourages people to return for their results [1,7,32,53].
This literature review has some limitations. Firstly, the differences in the measurement of factors and the specificity of certain populations (injection drug users, pregnant women, female sex workers, and men who have sex with men) introduce heterogeneity and do not allow meta-analyses for all factors. Secondly, most of the studies were conducted before the advent of rapid testing, but nowadays HIV testing is performed by rapid tests. Therefore, the issue of failure to return for HIV test results is only important in very specific contexts. The use of rapid tests might change the distribution and frequency of certain factors.
A majority of the studies were conducted in the USA The countries of Sub-Saharan Africa, which represent 2/3 of infected people worldwide, do not often publish their results, or only publish their results in local journal articles, which are not indexed in most databases.
Our search strategy was limited to publications in English and French. Only articles published in peer-reviewed journals were considered; grey literature and conference proceedings were not. This may have some implications for the external validity of our results. However, the review included a large number of studies, covering different regions, a broad range of populations, and barriers and facilitators with respect to returning for HIV test results. Furthermore, to the best of our knowledge, this is the first review to focus on factors associated with returning for HIV test results.

Conclusion
Helping more people learn their HIV status requires the strengthening of counselling and testing services. Returning

22
International Scholarly Research Notices for HIV test results is the gateway for knowledge and acceptance of HIV status. Various recently implemented strategies, such as provider-initiated testing and counselling, community-based testing and counselling, home-based testing and counselling, and the use of rapid tests, might not be effective if the people tested are not well advised and do not accept their results.
This review identified important factors that need to be addressed to ensure that people return for their HIV test results. Most barriers and facilitators identified were found at the individual level. These results highlight the fundamental role of counselling. Individuals most likely to fail to return for their results must be identified and targeted by the counsellor and delivered a specific message.