Determinants of sexual exposure to HIV in Portuguese and Brazilian adolescents: a path analysis

Abstract Objective: to analyze the direct and indirect effects of determinants of sexual exposure to the human immunodeficiency virus among male adolescents who have sex with men and the implications for nursing care. Method: cross-sectional study carried out with 578 Portuguese and Brazilian adolescents aged 18 and 19. Interrelationships of conjugal status, use of dating apps, practice of chemsex, unawareness, partner credibility, challenging sexual practices and ineffective forms of protection against sexual exposure to the human immunodeficiency virus were evaluated using the Path Analysis technique. Results: significant direct effect on sexual exposure to the human immunodeficiency virus: conjugal status (β=-0.16), use of apps (β=-0.30), challenging sexual practices (β=0.48) and ineffective forms of protection (β=0.35). Indirect paths: partner credibility influenced ineffective forms of protection (β=0.77); having a steady/polyamorous relationship influenced the use of dating apps (β=-0.46); chemsex, mediated by challenging sexual practices (β=0.67), determined greater sexual exposure. Conclusion: adolescent sexual behaviors and forms of amorous/sexual relationship must be considered in nursing care planning to reduce sexual exposure to the human immunodeficiency virus.

Even more serious is the fact that half of the people aged 15 to 19 living with HIV in the world were born in South Africa, Nigeria, Kenya, India, Mozambique and Tanzania (3) .
There are many factors why adolescents are at a high risk of acquiring HIV. Adolescence and early adulthood are a critical period of development, with significant physical, cognitive and emotional changes. In addition, in this period people experience increasing personal autonomy and responsibility for their own health. The transition from childhood to adulthood is also seen as a time to explore and deal with peer relationships, gender norms, sexuality and economic responsibility (4)(5) .
Given this context, there are still gaps in the literature addressing adolescents and their relationship with/ exposure to HIV. Studies with limited samples and lack of adequate disaggregation for this age group restrict the available evidence, hindering the implementation of new public policies for this group (2,6) , as evidenced by the noninclusion of adolescents in the HIV strategic plans of several countries (7) , including Brazil (8) and Portugal (9) .
Studies (10)(11) indicate that unprotected sex is the most common cause of HIV infection; however, the factors that determine such exposure are poorly understood. Although the literature points to lack of knowledge about HIV and the means of preventing it as the main associated factor, occasional unprotected sex and drug use are also commonly mentioned (12) . Adolescents from key populations are generally less aware of HIV risks or less able to mitigate them compared to their more experienced peers (13)(14) .
In this context, nursing professionals are strategically positioned and can contribute clinical and behavioral measures for the production of care, both locally and globally, besides health education/ literacy with a focus on coping with HIV, with consequent qualification of care and encouragement of teamwork through interprofessional cooperation (15) .
However, there are still many challenges to be faced, which is why this study focuses on a subject matter related to science and nursing practice, which reinforces its rationale, usefulness and social-scientific relevance.
Therefore, the objective was to analyze the direct and indirect effects of determinants of sexual exposure to HIV among male adolescents who have sex with men and the implications for nursing care.

Population, sample and eligibility criteria
A simple sample calculation for the proportion was performed using G*Power software (version 3.1.9.7), considering the male population aged 18 and 19 in both countries, with a presumed prevalence of 50% (in order to maximize the sample and considering that this is a phenomenon for which there were no prevalence data), a tolerable standard error of 3% and a confidence level of 95%, with the final sample of 578 MSM. There were no lost or discarded survey forms.

Data collection
The participants were adolescents aged 18 and 19 born in one of the nine countries of the Community of Portuguese Language Countries (CPLP). Participants who were already using Pre-Exposure Prophylaxis (PrEP) and those who were HIV positive were excluded.
The adolescents who were part of the sample were recruited online through two strategies. The first was snowball sampling, adapted to the virtual environment and consolidated by other studies (16)(17)(18)(19)(20) . Through this method, the participants themselves are responsible for recruiting other individuals in similar situations through their social and contact networks. Following the method's criteria, we selected 30 MSM with different characteristics, namely: region or district of residence in each country; race/skin color (white/non-white); income and educational level (primary, secondary or tertiary). These were the first participants, who were called seeds.
To identify the seeds, two experienced researchers, properly trained and calibrated, created a public profile on two geolocation-based dating apps (Grindr and Hornet),   and via direct chat with online users sent each participant   a link to the survey, instructing them to invite other MSM   from their social network until the required sample was obtained -dissemination strategy. At this stage, the study approached the first individuals who were available online in each of the two apps and met the inclusion criteria, as recommended by previous studies (14,(16)(17) .
Concurrently, as a second recruitment strategy, the researchers also promoted the research on two social media sites, Facebook and Instagram, targeting the MSM population aged 18 and 19. The social media were used as an additional resource due to their ability to reach people outside metropolitan areas, which is absolutely necessary in the case of a continental-sized country like Brazil (18)(19)(20) .
Only individuals who self-identified as men (cisgender or transgender) and aged 18 and 19 were included. Tourists and non-Portuguese speakers were excluded.
In this research, adolescents were considered to be youngsters aged 18 and 19, based on the definitions of the World Health Organization (WHO) (21) , respecting the minimum age criteria for using the investigated apps.

Data collection instrument
The survey form was hosted on the SurveyMonkey platform for data collection in two versions, with security features that allow only one response per Internet Protocol (IP). Since there are important linguistic differences between the two countries studied, the form was made available in two versions, Brazilian Portuguese and European Portuguese. It was previously validated (face-content validity) by 10 expert judges on the subject, five from each country. There was also a pre-test with five participants from each country.

Study variables
The study's dependent variable was recent sexual exposure to HIV among adolescents. Two variables were used for the construction of this indicator: a) sexual position of preference and b) recurrent practice of bareback sex (with response options 0-never, 1-sometimes and 2-always − this variable was dichotomized into 0-no for never and 1-yes for sometimes and always). partner is known; partner reports having no sexually transmitted infections (STIs); partner reports having been tested recently. The answers to these items were 0-yes and 1-no. For the construction of the indicator, those who answered no to all items were considered to have a reliable partner (0) − less exposed to the outcome; those who answered yes to at least one item were treated as having an unreliable/poorly reliable partner (1) − exposure category in the analyses. f) Challenging sexual practices: these were evaluated through positive answers to the variables "practices cruising," "practices double penetration" and "practices fisting." Adolescents who answered no to the three items were considered 0-no (less exposed) and those who practiced one and/or all three sexual practices were considered 1-yes (exposure category). 0-does not use this form of protection and 1-uses this form of protection. Those who answered no to all ineffective forms of protection were treated as 0-no (less exposed); those who answered yes to at least one form were treated as 1-yes (exposure category).
The variable "ineffective forms of protection" was considered the main exposure for the occurrence of the outcome, that is, sexual exposure to HIV.

Conceptual framework and study hypotheses
Considering the conceptual framework of sexual exposure to HIV, a hypothetical model was developed ( Figure 1) of the determinants in male adolescents who have sex with men. This conceptual framework, containing the study hypotheses, was represented by means of a Directed Acyclic Graph (DAG) (22)(23) , in which the variables, treated in this study as directly observed, are represented by rectangles. In this DAG, the relationship between the variables (vertices) is indicated by arrows (edges) that connect them in order to establish causal relationships that are direct (when an arrow goes from one variable to another) or indirect (when there is one or more mediating variables in the causal path between a variable and the outcome) (22)(23) .

Data treatment and analysis
The data were initially stored in the Statistical Package for Social Science (SPSS) software database, version 24.
First, descriptive analyses of the variables of interest were performed, estimating absolute and relative frequencies, with respective 95% confidence intervals, since all variables were treated as categorical. Next, the database was exported to a format compatible with Mplus software, version 8.7, in which the Path Analysis technique was processed.
Path analysis is an extension of multiple regression, which enables the analysis of complex models and causal relationships. The paths between the variables, whether direct or indirect, are quantitatively expressed through standardized path coefficients (regression coefficients) and aim to determine the strengths of the paths hypothesized in the path diagram (24) . As this is modeling with categorical data, the Weighted Least Squares Means and Variance Adjusted (WLSMV) estimator was used. The Modification Indexes (MI≥10) and the Expected Parameter Changes (EPC≥0.25) were jointly evaluated to identify the need to respecify the modeling paths (25) . Goodness of fit was evaluated using the Root Mean Square Error of Approximation (RMSEA), which must have an index below 0.06 (accepting up to 0.08) with the respective 90% confidence interval below 0.08 (25) ; Comparative Fit Index (CFI) and Tucker-Lewis Index (TLI) above or equal to 0.95 (26) ; and Chi-Square normalization (X 2 /degrees of freedom) below 2, as an indication of excellent fit (27) .    was associated with greater sexual exposure (Table 3).  Portuguese adolescent MSM, which are still related to strategic contexts for public health, such as immigration (33) .

Ethical and legal aspects
Analyzing the global contexts of adolescent health may contribute in a comprehensive manner to overcoming disadvantages between different countries, enhancing the dissemination of strategies and good practices and facilitating the socialization of knowledge.
Although they are countries with striking social, cultural and political differences, previous studies (18)(19)(20)28) have single (34)(35)(36) , but all of them are likely to be involved in casual sexual encounters. In addition, it is important to note that single adolescents tend to have more sexual partners than people over 20 years of age (35)(36)(37) .
The specifics of establishing relationships and partners seem to be key to understanding the determinants of HIV exposure in adolescents. Having some degree of involvement with the sexual partner, regardless of their HIV status, has led MSM to have unprotected sex (14,(38)(39)(40) . Our findings show that the greater the partner credibility, the greater the rate of ineffective forms of protection, the greater the frequency of chemsex and challenging sexual practices and, consequently, the greater the sexual exposure to HIV among adolescents. This behavior is already partly reported in the literature (41)(42) and can be explained by a mistaken notion of safety afforded by a stable or even repeated and known partnership. there is a perception that the risk is lower for those involved, who tend to reduce or dispense with the use of condoms when having sex with a partner (37,41,(43)(44) .
Moreover, they may have little knowledge and poor discernment about the efficacy and effectiveness of prevention methods, assuming a high risk of sexual exposure to HIV (13,45) , as corroborated in this study.  (46) . We identified that the frequency of chemsex, mediated by challenging sexual practices, has a moderate effect on sexual exposure among adolescents. In addition, frequency of challenging sexual practices was an important mediator of the relationship of use of dating apps and partner credibility with sexual exposure among adolescents.
HIV prevention in countries with relatively stable prevalence or where the epidemic is concentrated in specific populations, such as Brazil and Portugal, is especially challenging, as the spread of the virus is fueled by high-risk and often stigmatized behaviors (47)(48) .
So-called unconventional sexual practices are rooted in different aspects of sexuality and socialization, and some of them may be connected with chemsex. These factors present an inherent dualism (pain and pleasure, sobriety and disinhibition) that is not necessarily exclusionary and may coexist during sex, albeit in a conflicting way (49) .
The search for pushing sexual boundaries, whether literally or figuratively, developed within homosexual identities and practices in a queer perspective, shows that such practices are reflections of the non-belonging of sex between men in society, thus being execrated in its concept (50) .
Adolescence is understood as the typically healthy period of life; however, the HIV/STI diagnosis makes this phase more complex and challenging, considering the insufficient psychosocial maturity of teens to deal "alone" with the onset of diseases and conditions, especially infectious and chronic ones such as HIV (51)(52)(53) .