Routine HIV testing in men who have sex with men : from risk to prevention

We conducted a critical review of the literature on recurrent use of HIV testing in men who have sex with men (MSM). We performed a narrative review of the literature in which we analyzed the various conceptions on frequent testing over time, the implications for health programs, and the main social markers that influence the incorporation of HIV testing as routine care. Although it has existed since the 1990s, recurrent testing among MSM was frequently interpreted as increased exposure to HIV due to lack of condom use, and therefore as “unnecessary” testing. Beginning in the 2000s, periodic testing has become a programmatic recommendation and has been interpreted as a goal. Individuals’ perception of their use of the test has rarely been considered in order to characterize such use as routine care. On the social and cultural level, individual aspects associated with recent or routine testing were included in contexts of favorable norms for testing and less AIDS stigma. Differences in generation, schooling, and types of affective-sexual partnerships play an important part in testing. Such differences highlight that the epidemiological category “men who have sex with men” encompasses diverse relations, identities, and practices that result in specific uses of the test as a prevention strategy. Thus, dialogue between programs, health professionals, and the persons most affected by the epidemic is crucial for building responses with real potential to confront the HIV epidemic, based on respect for human rights. AIDS Serodiagnosis; HIV; Male Homosexuality; Human Rights Correspondence B. R. L. Redoschi Faculdade de Saúde Pública, Universidade de São Paulo. Av. Prof. Mello Moraes 1721, bloca A, salas 103/105, Cidade Universitária, São Paulo, SP 05508-030, Brasil. brunarl@usp.br 1 Faculdade de Saúde Pública, Universidade de São Paulo, São Paulo, Brasil. 2 Programa de Pós-graduação Stricto Sensu em Saúde Coletiva, Universidade Católica de Santos, Santos, Brasil. 3 Instituto de Psicologia, Universidade de São Paulo, São Paulo, Brasil. REVISÃO REVIEW doi: 10.1590/0102-311X00014716 Cad. Saúde Pública 2017; 33(4):e00014716 This article is published in Open Access under the Creative Commons Attribution license, which allows use, distribution, and reproduction in any medium, without restrictions, as long as the original work is correctly cited. Redoschi BRL et al. 2 Cad. Saúde Pública 2017; 33(4):e00014716 Introduction The importance of periodic HIV testing as a preventive strategy in the programmatic response to the HIV/AIDS epidemic has been highlighted at the global level, especially in disproportionately affected segments in contexts of concentrated epidemic, notably men who have sex with men (MSM). This is a relatively new tendency, marked by optimism over the possibility of combining different prevention strategies (biomedical, behavioral, and structural) 1 and supported by technological advances that have facilitated the expansion of testing. An example of this trend is the 90-90-90 Target proposed by the Joint United Nations Programme on HIV/AIDS (UNAIDS) in 2014 2, which consists of eradicating the HIV epidemic by 2030 through diagnosis of 90% of the persons infected with HIV, maintenance of 90% of the infected persons in antiretroviral therapy (ART), and viral suppression in 90% of these persons. Another expression of this trend is the recommendation by the World Health Organization (WHO) that persons at risk of HIV infection (defined as MSM, sex workers, injecting drug users, and persons with a seropositive stable partner) have an HIV test at least once a year 3. Still, despite the higher prevalence of testing among MSM when compared to other affected segments, in recent years Brazil and other countries have witnessed a rising incidence of HIV among MSM, with worrisome rates of late diagnosis 4,5. In this context, in which access to treatment and viral load suppression are the principal strategies for ending the HIV/AIDS epidemic, early serological diagnosis is fundamental for public health strategies. However, this central role of HIV testing in building programmatic responses to the epidemic is relatively new. Although MSM had been using HIV testing as a prevention strategy since the 1990s 6, this use of the test was not immediately acknowledged by those responsible for the production of public policies for prevention. On the one hand, we have the development of the testing technologies themselves. Until the late 1990s, having an HIV test was not simple. ELISA (enzyme-linked immunosorbent assay) testing requires laboratory equipment, water and electricity, and experienced technicians that know how to operate the equipment, prepare reagents, and pipette correctly 7. Although the first rapid tests date to 1992 8, they were not incorporated immediately into the prevention programs, but were used initially in developing countries lacking the infrastructure for performing ELISA 9. In the United States, the Centers for Disease Control and Prevention (CDC) only began to recommend the rapid test in 1998 (for high-risk populations), and the test was not effectively incorporated until 2003 7. In Brazil, the Ministry of Health, working in partnership with the CDC, also incorporated the rapid test at this time 10. Australia, for example, only released use of the rapid test in 2012 11. Meanwhile, there have been advances in treatment. At the beginning of the epidemic the principal program strategy for prevention was counseling 12. The benefits of serological diagnosis for persons with HIV were debatable. Besides the fact that it was a fatal disease for which no treatment existed, AIDS could lead to the isolation and stigmatization of groups and individuals 12. The prevailing international position was that testing should be confidential, voluntary, and accompanied by counseling, meaning that it could only be performed with the individual’s free and informed consent 12,13. Brazil’s experience with anonymous testing dates to this time. However, with advances in treatment, including the availability of zidovudine (AZT) in 1987 and highly active antiretroviral therapy (HAART) in 1996, another place for the test began to be built at the programmatic level. Not only could the test bring individual benefits in places where HAART was available, but it could also potentially interfere in the HIV transmission chain 14, as observed in 1994 with the role of AZT in the mother-to-child transmission of HIV 12. In the early 2000s, a new global discourse emerged on testing, with more flexibility towards counseling and encouragement for the expansion of testing 12,15. This movement reached one of its high points in 2009 with publication of the article by Granich et al. 16 in Lancet, promoting treatment as prevention (TasP) as a strategy to eradicate HIV 1,12. This required periodic, universal testing and access to HAART for all persons infected with HIV 1,12,16,17. Expectations in relation to TasP provided the basis for strategies underlying the UNAIDS 90-90-90 Target and the Brazilian Ministry of Health’s current clinical protocol and treatment guidelines for management of HIV infection in adults 1. In this sense, in the age of HAART and in the context of the growing HIV epidemic among MSM 4,5, annual repeat testing by MSM was incorporated into the program guidelines 3,18. ROUTINE HIV TESTING IN MEN WHO HAVE SEX WITH MEN 3 Cad. Saúde Pública 2017; 33(4):e00014716 The current study aims to conduct a critical review of the literature on recurrent use of HIV testing among MSM in light of the technological and social developments that have repositioned the test in the program strategies for confronting the HIV/AIDS epidemic. We will specifically examine the views on recurrent HIV testing and the social markers of difference 19 implied in the incorporation of the test as routine care among MSM. Methodological procedures A narrative review of the literature was performed. We selected studies with original research on factors associated with repeat/routine testing among MSM, prioritizing those in which this was the central theme. The principal database used was the PubMed, and we did not limit the searches to any given period, in order to have a historical overview of this literature. Given the exploratory nature of the narrative review, different search terms (not necessarily descriptors) were used as we proceeded with reading the articles, and we also included studies cited in the articles’ reference lists and institutional sources. Our line of analysis focused on the different conceptions and positions on repeat testing/routine testing over time and the principal social markers (affective-sexual partnerships, age, schooling, living place, spaces of socialization, disclosure of homosexual orientation, knowing persons with HIV, symbolisms of AIDS, perception and knowledge of the test) involved in this use of HIV testing in men who have sex with men. Table 1 and Figure 1, respectively, present the articles analyzed and a brief contextualization of these studies according to important references in the development of testing policies.


Introduction
The importance of periodic HIV testing as a preventive strategy in the programmatic response to the HIV/AIDS epidemic has been highlighted at the global level, especially in disproportionately affected segments in contexts of concentrated epidemic, notably men who have sex with men (MSM).This is a relatively new tendency, marked by optimism over the possibility of combining different prevention strategies (biomedical, behavioral, and structural) 1 and supported by technological advances that have facilitated the expansion of testing.
An example of this trend is the 90-90-90 Target proposed by the Joint United Nations Programme on HIV/AIDS (UNAIDS) in 2014 2 , which consists of eradicating the HIV epidemic by 2030 through diagnosis of 90% of the persons infected with HIV, maintenance of 90% of the infected persons in antiretroviral therapy (ART), and viral suppression in 90% of these persons.Another expression of this trend is the recommendation by the World Health Organization (WHO) that persons at risk of HIV infection (defined as MSM, sex workers, injecting drug users, and persons with a seropositive stable partner) have an HIV test at least once a year 3 .Still, despite the higher prevalence of testing among MSM when compared to other affected segments, in recent years Brazil and other countries have witnessed a rising incidence of HIV among MSM, with worrisome rates of late diagnosis 4,5 .In this context, in which access to treatment and viral load suppression are the principal strategies for ending the HIV/AIDS epidemic, early serological diagnosis is fundamental for public health strategies.However, this central role of HIV testing in building programmatic responses to the epidemic is relatively new.Although MSM had been using HIV testing as a prevention strategy since the 1990s 6 , this use of the test was not immediately acknowledged by those responsible for the production of public policies for prevention.
On the one hand, we have the development of the testing technologies themselves.Until the late 1990s, having an HIV test was not simple.ELISA (enzyme-linked immunosorbent assay) testing requires laboratory equipment, water and electricity, and experienced technicians that know how to operate the equipment, prepare reagents, and pipette correctly 7 .Although the first rapid tests date to 1992 8 , they were not incorporated immediately into the prevention programs, but were used initially in developing countries lacking the infrastructure for performing ELISA 9 .In the United States, the Centers for Disease Control and Prevention (CDC) only began to recommend the rapid test in 1998 (for high-risk populations), and the test was not effectively incorporated until 2003 7 .In Brazil, the Ministry of Health, working in partnership with the CDC, also incorporated the rapid test at this time 10 .Australia, for example, only released use of the rapid test in 2012 11 .
Meanwhile, there have been advances in treatment.At the beginning of the epidemic the principal program strategy for prevention was counseling 12 .The benefits of serological diagnosis for persons with HIV were debatable.Besides the fact that it was a fatal disease for which no treatment existed, AIDS could lead to the isolation and stigmatization of groups and individuals 12 .The prevailing international position was that testing should be confidential, voluntary, and accompanied by counseling, meaning that it could only be performed with the individual's free and informed consent 12,13 .Brazil's experience with anonymous testing dates to this time.
However, with advances in treatment, including the availability of zidovudine (AZT) in 1987 and highly active antiretroviral therapy (HAART) in 1996, another place for the test began to be built at the programmatic level.Not only could the test bring individual benefits in places where HAART was available, but it could also potentially interfere in the HIV transmission chain 14 , as observed in 1994 with the role of AZT in the mother-to-child transmission of HIV 12 .In the early 2000s, a new global discourse emerged on testing, with more flexibility towards counseling and encouragement for the expansion of testing 12,15 .This movement reached one of its high points in 2009 with publication of the article by Granich et al. 16 in Lancet, promoting treatment as prevention (TasP) as a strategy to eradicate HIV 1,12 .This required periodic, universal testing and access to HAART for all persons infected with HIV 1,12,16,17 .Expectations in relation to TasP provided the basis for strategies underlying the UNAIDS 90-90-90 Target and the Brazilian Ministry of Health's current clinical protocol and treatment guidelines for management of HIV infection in adults 1 .In this sense, in the age of HAART and in the context of the growing HIV epidemic among MSM 4,5 , annual repeat testing by MSM was incorporated into the program guidelines 3,18   The current study aims to conduct a critical review of the literature on recurrent use of HIV testing among MSM in light of the technological and social developments that have repositioned the test in the program strategies for confronting the HIV/AIDS epidemic.We will specifically examine the views on recurrent HIV testing and the social markers of difference 19 implied in the incorporation of the test as routine care among MSM.

Methodological procedures
A narrative review of the literature was performed.We selected studies with original research on factors associated with repeat/routine testing among MSM, prioritizing those in which this was the central theme.The principal database used was the PubMed, and we did not limit the searches to any given period, in order to have a historical overview of this literature.
Given the exploratory nature of the narrative review, different search terms (not necessarily descriptors) were used as we proceeded with reading the articles, and we also included studies cited in the articles' reference lists and institutional sources.Our line of analysis focused on the different conceptions and positions on repeat testing/routine testing over time and the principal social markers (affective-sexual partnerships, age, schooling, living place, spaces of socialization, disclosure of homosexual orientation, knowing persons with HIV, symbolisms of AIDS, perception and knowledge of the test) involved in this use of HIV testing in men who have sex with men.
Table 1 and Figure 1, respectively, present the articles analyzed and a brief contextualization of these studies according to important references in the development of testing policies.

Different views of repeat testing
Based on a view of prevention until the mid-1990s, centered on counseling for abstinence, fidelity, and condom use in 100% of sexual relations 6 , the news that many people had tested more than once in life was received with concern by health researchers, professionals, and policymakers.There was the notion that repeated negative HIV tests would lead persons to underestimate their exposure to HIV and fail to adopt protection in their sexual relations.A review in 2002 on repeat HIV testing showed that persons who tested several times tended to resist counseling 20 .
The clash over whether to repeat the test and the possible disadvantages of this practice were based mainly on the discussion of the "risk behaviors" of persons that repeated the test.Risk served as a measure of both the need to re-test and as the basis for discussions on the effectiveness of counseling.We emphasize that "risk" was defined in various ways.In some studies it was synonymous with belonging to a risk exposure category 21 or having had sex with other men 20 .Such understanding of recurrent testing is sustained by the notions of risk groups and behaviors that marked the epidemic in various countries until the early 1990s.
The lack of understanding on what led persons to test several times is reflected in the criteria used to define the so-called "repeat testers".Some studies defined a "repeat tester" as someone who had tested more than once 22,23 , while others included in this category men who had had three or more tests in their lives 24,25,26,27,28,29 .Since the 2000s, the term "repeat testers" has been replaced by other notions of the frequency of repeat testing, more adequate for the new guidelines on frequent testing for MSM.
Some studies had already raised the issue of regular testing 24,25,28 , but in some cases this regularity was still defined according to the number of lifetime tests 24,28 .Beginning in the early 2000s, the parameter was the time elapsed since the last test, in keeping with the expectation that MSM should test periodically 30,31,32,33,34,35,36,37 .

Repeat HIV testing as reiteration of exposure
Phillips et al. 24 conducted one of the first studies on the theme, still in the pre-HAART era and prior to more widespread use of the rapid test.According to the authors, one-fourth of HIV tests in the United States in 1992 were done in persons that had already been tested; the CDC proposed that an effort be made to decrease "unnecessary" testing.At the time, the test did not have a preventive role, but a diagnostic one.It was thought that repeat testing was being used as a substitute for safe sex practices  or that it could lead to a perception of invulnerability in the case of repeated negative results 24 .Some authors at the time questioned the effectiveness of counseling to reduce the risk of infection and repeat testing by persons with little risk of infection and who wanted to confirm a negative result 27 .
According to some studies, MSM that tested repeatedly were also the ones with the highest risk of HIV infection 22,24,26,27 .According to this line of reasoning, while some authors justified repeat testing based on this high risk of infection, others were concerned about the possible negative effects of this practice.It was argued that successive negative results could reinforce risk behaviors for HIV infection 26,27 .This concern over repeat testing resulted in proposals for interventions and specific counseling strategies for those who had tested more often (more than once lifetime or three or more times), aimed at reducing the risk to these persons through more intense interventions 27 .
In some of the studies that emphasized the relationship between higher-risk practices and repeat testing, the criteria used to define risky sexual relations were limited only to individual aspects.While acknowledging that the factors associated with repeat testing were multifaceted and that such repetition could be part of routine health precautions 22 , repetition of HIV testing was often associated with and interpreted according to risk behaviors, i.e. unprotected anal and oral sex and higher number of sex partners 25,30,31,32,34,37,38 .Such studies also highlighted that having more lifetime partners or more partners in a given period was associated with repeat 22 , recent, or regular testing 25,30,31,32,34,37,38 among white and black MSM, as well as in other ethnic groups like Hispanics 28 .
Cad. Saúde Pública 2017; 33(4):e00014716   Although backed by different justifications, authors that analyzed repeat testing from a more biomedical perspective and those who considered the importance of community prevention strategies opposed programs that discouraged repeat HIV testing 22,24,25,26,27 .

Repeat HIV testing as a prevention strategy
In the context of expanded testing and optimism regarding the substantial drop in AIDS mortality following the introduction of HAART in 1996 and thus the greater possibility of controlling new infections, beginning in 2002 we witnessed a new scenario of publications on routine repeat testing.Another view of HIV testing began to mark the analyses on repetition or frequency of the test.HIV testing began to be identified more clearly as a highly recommendable prevention strategy from the point of view both of governments and programs 30,39 .
This notion of repeat testing as part of individuals' routine prevention was already present in some studies published before 2002.In 1995, Phillips et al. 24 defended the preventive role of the test in the sense of reinforcing safe practices and increasing safety for gays and bisexual men deeply impacted by the epidemic.For these authors, testing repeated times could be a rational solution for the uncertainties confronted by these men, e.g., in relation to the low probability of infection via oral sex, in addition to being fundamental for men in serodiscordant stable relationships and for those who wanted to treat early in case of having been infected.In 2000, Leaity et al. 26 reached similar conclusions, indicating that the test was incorporated as part of a personal risk reduction strategy.Along     the same line, in 1997 Kalichman et al. 25 had stated that recurrent use of HIV testing was associated with higher rates of condom use and positive attitudes towards one's own health, so that it could function as a prevention strategy for some MSM.According to the authors, the test had been incorporated by MSM as a self-care resource, like condoms.However, the studies published since 2002 show concerted efforts at establishing program guidelines on routine repeat testing.The issue was no longer whether repeat testing should be discouraged, but rather to identify who should be encouraged to test 23,31 .Although some authors suggested different counseling strategies for different groups of MSM 23 , no mention was made of restricting access to HIV testing.

Recent test as an indicator of current individual care
In this context, in 2002 the CDC began to recommend that sexually active MSM be tested for HIV at least once a year 39 , a policy that was maintained in subsequent publications 40,41 .Adherence to this recommendation also became the object of analysis 31,35,36,42 .While repeat testing had initially raised concern, in this new context the "problem" became those who had not tested in the previous year (e.g., Katz et al. 38 , Knussen et al. 35 ).
At this moment, based on both a reading that acknowledged the importance of community HIV prevention strategies and another that prioritized a biomedical perspective, many studies began to detect the preventive potential of periodic HIV testing.According to community strategies, repeat testing was seen as resulting from a rational and informed decision, especially by men in the gay community 30 .The studies also highlighted the importance of knowledge of serological status in establishing sexual agreements with partners 37 .From a biomedical perspective, some studies contextualized the importance of routine testing to avoid late detection of HIV, the resulting high costs, and greater risk of HIV transmission 31,34,36,37 .Others claimed TasP as the justification for routine use of HIV testing in prevention 33,35 .
Test regularity became central in these analyses.Some authors began to establish approximations in this sense, constructing categories of "regular" or "repeat testers" as a function of the lifetime number of tests 28 .Other studies aimed to understand the differences between MSM who had never tested or had tested longer ago versus those who had tested recently 31,32,37 .
Although annual testing became a recommendation, the debates continued on the risk of HIV infection among those that tested repeatedly.Unprotected oral sex was no longer mentioned in these studies or was relegated to a secondary level as a lower-risk practice 23 .The practice of anal sex continued to interest researchers.Some authors sought to contextualize unprotected sex, identifying the type of partnership and/or partner's seroconcordance (e.g., Jin et al. 30 ).Other studies adopted a biomedical perspective whereby sexual risk was evaluated according to seroconversion rates 23 .The same was done with risk perception: MacKellar et al. 31 found that seroconverted MSM who did not disclose episodes of exposure made an inadequate assessment of their risk.
Along this line of inquiry, some authors continued to treat protected or unprotected anal sex independently of the partner, making it difficult to understand whether the risk mentioned in their studies was "actual risk" from the men's point of view, or risk reduction 28,36 .Although some of these studies acknowledged this limitation and even the multiplicity of different practices that were potentially being treated as the same phenomenon (e.g., Rendina et al. 36 ), nearly all of them concluded that persons that tested repeatedly or recently were exposed to greater risk of HIV infection 28,36 .
The potential danger of successive negative results, frequently mentioned in the first studies on repeat HIV testing, began to be cited less frequently.The theme of disinhibition/relaxation of sexual practices became less debated.However, some authors still interpreted the association between recent testing and unprotected anal sex as influenced by a feeling of invulnerability resulting from successive negative results 36 .A qualitative study on this theme revealed a broad mosaic of possible situations in the presence of repeated negative results 29 .The authors concluded that their data were worrisome, since for some individuals the negative results signaled immunity, while for others they reinforced the maintenance of practices that were not "100% safe"; meanwhile, in light of the risk reduction strategies, the interviewees' discourse gained another dimension.HIV testing appeared to be grasped Cad.Saúde Pública 2017; 33(4):e00014716 according to each individual's needs and histories, and there was more than one way of using the test, with a considerable share of these ways consisting of creating sustainable risk reduction practices or seeking care in case of exposure to risk of infection.
By including community prevention strategies in their analyses, other authors felt it was expected that men who incorporated the test into their lives would have higher odds of reporting unprotected anal sex than men who had never tested 32 .Test-seeking was thus not viewed as the result of failure in prevention or increased risk, but as a conscious strategy by MSM.
Some authors observed that men who had tested recently reported fewer practices considered risky, e.g., unprotected anal sex with a partner with unknown serology 35,38 .Unprotected anal sex was not associated with recent HIV testing without taking into account knowledge of the partner's serology and type of partnership 35 .Men tested recently appeared to adopt condom use in anal sex or other risk-reduction strategies 35 .
In addition to the debate on the association between unprotected sex and repeat testing, the literature points to other factors associated with testing more than once that reinforce the use of the test as a preventive strategy and self-care.Routine testing was frequently associated with greater concern and/or care with one's own health 22,24,25,26 .More recently, Katz et al. 34,38 showed that shorter interval between HIV tests was associated with having regular access to health services and taking routine care for one's health.According to this logic, Marcus et al. 37 found that "routine" was the most frequent reason for testing in the previous year among MSM.
Lorenc et al. 43 , in a review of qualitative studies in England, United States, Scotland, and Canada found that habit could lead to seeking the test, regardless of any specific trigger.Routine testing was associated with a sense of responsibility for one's own health and commitment to prevention, an aspect not always emphasized in studies on HIV testing.
Along this line, some studies showed that perceiving the importance of receiving information on HIV/AIDS prevention from a health professional and discussing the test with a professional were predictors of recent testing (≤ 1 year) 31 .Holt et al. 32 observed that it was more likely for men who had tested recently to seek information/counseling from a health professional when compared to those who had tested more than a year previously, and that they had sought help more frequently than MSM who had never tested.
Other studies found an association between repeat testing and having acquired or thinking that one had acquired an STI at least once in life 22,28 .Fernández et al. 28 interpreted that believing that one had acquired an STI increased the perception of risk of HIV infection.This association was also observed in MSM that had tested recently (time since last test ≤ 1 year) 31 .

Social markers of difference in testing -how to interpret the epidemiology of risk beyond individual behavior?
As highlighted by Aggleton & Parker 44 , when we use the term "MSM" we are including very different experiences, subjectivities, and positions within a single category.Thus, it is not possible to affirm a hegemonic practice or a collective MSM identity that be fully encompassed by prevention programs.It is thus essential to understand how the dimension of individual practices and behaviors discussed above is inscribed in diverse intersubjective contexts informed by symbolic relations and social norms that operate as modulators of repeat testing.Along this line, we will present social markers of difference that act as facilitators or barriers for MSM to incorporate the test into their routine health care.Analyzing social markers of difference allows understanding how the varied cultural and discursive normative productions on difference (type of partnership, age, socioeconomic status, skin color, and others) can result in inequality and oppression, as well as diversity and agency 19 .

• Affective-sexual partnerships
Affective-sexual partnerships and relationships are central to understanding how MSM incorporate the test into their preventive practices for managing their sexual risk.Responsibility for their own health and that of others can be decisive for an individual to seek the test 43 .The test can also play a part in the relationships as proof of serological status for a potential partner, for example, to be able to Cad.Saúde Pública 2017; 33(4):e00014716 have sex without condoms or to reduce risks in sexual interactions 26,43 .Meanwhile, in relationships involving trust and commitment with the partner, the test will not necessarily be incorporated, since the partners do not feel exposed to the risk of infection 43 .
Monogamous relationships or ones that have lasted at least two years were associated with a decrease in recent testing in Australia 30 .In Germany, MSM not in stable partnerships showed higher odds of recent testing (versus having been tested longer ago) when compared to those in open relationships 37 .
Having a seropositive partner was also associated with having tested recently, which means a rational use of HIV testing 30 .In addition, disclosing one's serology to the partner was more common among MSM that had tested recently 31,32 .Being unaware of the partner's serology was more common among young MSM who were testing for the first time, according to MacKellar et al. 27 .Holt et al. 32 showed that the expectation that HIV-negative partners would disclose their serology before having sex was greater among MSM who had tested recently and MSM who had never tested than among those who had tested longer ago.Living up to their expectations, men who had tested recently discussed their serological status more frequently with casual partners 32 .Meanwhile, men never tested were perhaps counting on disclosure from tested partners, since they themselves had never tested 32 .

• Age
Age was frequently associated with repeat or regular testing.However, the effect of age on inclusion of the test in the strategies to deal with HIV is not homogeneous and depends on the contexts in which the analyses are performed.Some studies showed that younger MSM tested more often than older MSM 25 or had tested more recently 35 .In Australia, at two different moments, diverse associations were observed between recent testing and age.In 2002, recent testing was more frequent among younger men 32 .In 2012, recently tested MSM were older than those who had tested longer ago, who in turn were older than those who had never tested 30 .
According to other authors, the most frequent age bracket of men who had tested more than once was 25 to 34 years 23 , which coincided with the age bracket with the highest HIV incidence in San Francisco (USA).In Germany, this was also the age bracket with the highest recent test rate 37 .

• Schooling and place of residence
In relation to schooling, Holt et al. 32 observed some differences, e.g., MSM with more schooling were more likely to have tested more than a year previously (compared to those who had never tested), but there was no difference in relation to those who had tested recently.
Place of residence was associated with testing in some studies 32,37 , and repeat testing was more likely in residents of State capitals and other large cities 32,37 .Myers et al. 45 pointed out that it is common for MSM to be attracted to large cities.Historically, large cities provide a more favorable setting both for anonymity and for social interaction and the development of a gay culture 46 .
The relationship between these markers and the incorporation of HIV testing into routine care for one's health should be analyzed in its specific context, based on a reading that includes networks of sociability, access to information, and prevention services.

• Spaces for socialization and disclosure of homosexual orientation
Some contextual conditions were considered more favorable for repetition or regularity of HIV testing.Socialization in the gay community, coming out as gay, and exposure of one's sexual orientation to others were analyzed as factors that favored test-seeking.These differences may mark greater socialization in cultures of prevention incorporating HIV testing as a strategy.
Belonging to the gay community was positively associated with recent testing in Australia in 2002 30 .Ten years later this association was observed again: having more gay friends was associated with recent testing (≤ 1 year) 32 .As observed in Australia, in Germany, MSM who visited spaces of MSM sociability showed higher odds of having tested recently as compared to never having tested or having tested longer ago 37 .
Cad. Saúde Pública 2017; 33(4):e00014716 As for online socialization, Holt et al. 32 showed that MSM that had never tested spent more time on social networks on the internet, a space suggested by the authors for circulating HIV prevention campaigns 32 .The culture of using the test as a preventive strategy may not have taken hold in this virtual space of MSM sociability.
Belonging to a gay community can favor sharing and experiencing community prevention practices, and facilitate access to culturally adequate information on prevention.MSM that had tested recently did not differ from those who had tested more than a year before, but they had higher odds of having sought information in the gay community through HIV/AIDS organizations, compared to those who had never tested 32 .
With a similar effect to the occupation of spaces of MSM sociability, the acceptance of one's own sexual orientation appears to act as a facilitator for the incorporation of HIV testing into HIV prevention practices.Marcus et al. 37 stated that it was more likely for MSM with lower levels of internalized homophobia to have tested recently, compared to never having tested.In addition, MacKellar et al. 31 showed that MSM who had disclosed their sexual orientation to others had higher odds of recent testing (compared to those who had never tested or had tested longer ago).In the case of friends and coworkers, according to Marcus et al. 37 , MSM who had tested recently revealed their sexual orientation more to coworkers than those who had never tested, but they revealed their orientation less than those who had tested longer ago.Having disclosed one's sexual orientation to a health professional was more frequent among MSM tested recently than among those never tested and those tested longer ago 37 , which could be an important marker of access to services and adequate information on prevention according to sexual orientation.
Another relevant element for prevention is drug use in spaces of MSM sociability.An association can be seen between recent HIV testing and illegal drug use in the previous six months 31 , as well as regular testing and the use of inhalable nitrites in the previous year 38 .

• Knowing persons with HIV and symbolisms of AIDS
Further in relation to the context favoring test-seeking among MSM, we highlight the association between knowing seropositive persons and repeat testing 25,26 .Some of these studies suggest that repeat testing can be encouraged by experiences with seropositive friends and relatives, e.g., experiences with illness and deaths.
More favorable attitudes and norms towards HIV/AIDS and HIV testing also helped build a more favorable context for frequent testing.When compared to MSM that had never tested, those tested in the previous year had fewer negative attitudes towards sex with seropositive partners 33 and were less prone to stigmatize them 37 .MSM that had tested recently (≤ 1 year) were less likely to report fear of the test result than MSM who had never tested 33,35 and MSM tested longer ago 35 .Meanwhile, in relation to fear of the result, MSM that had tested more than a year previously did not differ from those who had never tested 35 .

• Perception and knowledge of the test
Social norms favoring HIV testing, communication of the test result, and perception of benefits in obtaining medical care in case of infection were associated with repeat testing in the study by Phillips et al. 24 .More recently, Knussen et al. 35 also observed this association: MSM tested in the previous year were more likely to perceive HIV testing among gays as a norm ("almost all of my gay friends have tested") when compared to MSM tested longer ago or never tested.No difference was seen between MSM that had tested more than a year previously and those never tested 35 .Flowers et al. 33 showed that perceiving the test as a benefit was associated with having tested recently (compared to never having tested).Along this line, one study showed that knowing of the existence of HAART was also associated with recent testing 31 .According to Lorenc et al. 43 , in the presence of more negative social norms towards HIV, it may be easier to deal with uncertainty than with a potentially positive test result.Not knowing one's own serology is as if the person continues on the same level as all the untested persons, whereas having the test would require taking stances and making changes 43 .

Final remarks
In the studies analyzed here, the criteria used to classify frequency of HIV testing in the literature were the number of tests or time elapsed since the last test, except for two studies 25,38 in which the categorization was based on the individuals' perception of how they relate to HIV testing.In the semantic field, "repetition" of the test, a characteristic of the literature in the 1990s, appears to represent the "amount" of tests and is linked to behaviors interpreted predominantly on the individual level.Meanwhile, "routine" testing appears to emphasize the incorporation of a measure of care with the individual's own health, and in this sense the social markers that influence individual behavior allow a broader understanding of the different testing contexts for men who have sex with men.
In this sense, it is essential to integrate epidemiological knowledge with other approaches (e.g., qualitative) in order to take the men's own perspective into account on routine use of HIV testing.In order for the programmatic response to be adequate and effective, it is crucial to have a better understanding of the motivations and contexts of those that have adopted the test as routine.
On the social and cultural levels, individual aspects associated with repeat or recent testing are part of a context marked primarily by favorable norms for the test and less stigmatization of persons living with HIV/AIDS.This context may have been built in spaces of homosexual sociability in which individuals find acceptance and support for disclosing their sexual orientation.Thus, the intersection of these different social markers allows glimpsing test-seeking trajectories as routine care in which MSM gradually emancipated themselves as subjects of the right to health, by confronting processes of stigmatization and discrimination that impacted this population and substantially shaped the social response to the AIDS epidemic.It is also possible that generational differences constitute an important element in the incorporation of routine testing.Such differences are not homogeneous and need to be carefully contextualized in order to understand how they affect the incorporation of HIV testing in the lives of younger or older MSM.
As an expression of the right to health, information on HIV testing should be spread widely, and access should be free of financial, geographic, or discriminatory constraints 47 .From the point of view of prevention that includes persons as protagonists of their own daily lives, as subjects of the right to health 48 , it is essential that the spontaneous search for the test be incorporated into the routine of persons potentially exposed to HIV, as best adjusted to their wishes and needs for care.It is thus necessary to remember that the epidemiological category "men who have sex with men" always includes a wide diversity of settings, identities, and practices that result in specific uses and strategies of the test as prevention for each person and in each territory.
Equally important is to qualify and expand the supply of HIV testing among MSM as part of a routine that includes other forms of care, as recommended by the WHO since 2007 49 .Although testing initiated by health professionals has led to an important increase in testing among women in prenatal care in various countries 50 , such testing is frequently associated with late diagnosis of HIV 51 .Therefore, overcoming organizational barriers in services to prioritize care for persons with greater vulnerability to HIV infection, organizing testing in the intense flow of care, and training professionals not to feel embarrassed when addressing questions on sexual practices with users 52 are central tasks for more active routine testing in health services.
We also emphasize the late response by public health when compared to the community response.Long before the programmatic recommendation of routine testing for MSM, in the mid-1990s gay men were already using repeat HIV testing as a strategy for dealing with the epidemic 6 .As we have seen, this use of the test was not always interpreted the same way by health professionals, policymakers, and researchers, but as a possible failure of counseling programs, lack of knowledge, or even resistance to the incorporation of safe sex practices.A history of routine use of HIV testing shows that dialogue between programs, health professionals, and the persons most affected by the epidemic, who have accumulated knowledge from managing their sexual risk on a daily basis, is central to building responses with real potential to deal with the HIV epidemic and ensure respect for human rights.

Contributors
The four authors participated equally in elaborating the article.

Table 1
Original studies analyzed, listed by order of publication.

Table 1 (continued)
Timeline: articles analyzed, contextualized according to advances in treatment, technologies, and testing policies.