Interpersonal violence victimisation, HIV-related behaviours and STIs among adult, urban Indigenous and non-Indigenous gay, bisexual and other men who have sex with men in Guatemala

Abstract

Gay, bisexual and other men who have sex with men are disproportionately affected by HIV in Guatemala; interpersonal violence may increase behaviours that augment the risk of HIV in this group. Although 44% of Guatemalans identify as Indigenous, little information exists on the experiences of Indigenous sexual minority individuals. In this study, we sought to compare different forms of violence and HIV-related behaviours by Indigenous identity among gay, bisexual and other men who have sex with men; and determine if associations between violence and HIV-related behaviours differed by Indigenous identity. We used cross-sectional survey data from 716 Spanish-speaking, adult men recruited from urban centres to examine the prevalence of and relationship between different forms of interpersonal violence and HIV-related behaviours using logistic regression analyses, including the moderating effect of Indigenous identity. In general, fewer Indigenous participants reported interpersonal violence victimisation and HIV-related behaviours compared to their non-Indigenous counterparts. In weighted multivariable analyses, non-Indigenous participants who reported physical and/or sexual violence were over five times as likely to report transactional sex (OR=5.17, 95% CI 2.11-12.68, p<0.001), but the relationship was not significant for Indigenous participants. Findings suggest that Indigenous sexual minority men have unique contexts and that additional strengths-based research is needed to ensure that actions and efforts to promote violence and HIV prevention meet their needs.

Introduction

Cisgender sexual minority men (gay, bisexual and other men who have sex with men) are disproportionately affected by HIV and other sexually transmitted infections (STI) in Guatemala. Among such men, HIV prevalence is 10.5% and active syphilis prevalence is 11.0%, compared to less than 1% for both among the general population (Guardado Escobar et al. 2017). Although 44% of Guatemalans identify as Indigenous, there exists very little information on the experiences of Indigenous sexual minority (non-heterosexual) individuals, including HIV prevalence for this group (Instituto Nacional de Estadística de Guatemala 2019).

One factor that may increase vulnerability to HIV and STI is interpersonal violence, which refers to violence between individuals. This includes family and partner violence (child maltreatment, elder abuse and violence from an intimate partner) as well as community violence (acquaintance and stranger violence outside of the home, including in workplaces and other institutions) (World Health Organization 2014). Being the victim of violence can have a direct effect on vulnerability to HIV through rape or assault, as well as an indirect effect through asymmetry of power within a relationship and condom negotiation (Cagney 2014; Jewkes and Morrell 2010). Among gay and bisexual men, studies have found that different forms of interpersonal violence are associated with several HIV-related behaviours, including condomless anal sex, substance use, transactional sex, and sex with multiple concurrent partners (Beyrer et al. 2012; Chellan et al. 2011; Houston and McKirnan 2007; Wheeler et al. 2014); and HIV and STI transmission and acquisition (Marks et al. 2005; Mustanski et al. 2020).

In Guatemala, sexual minorities experience frequent stigma, discrimination and violence because Guatemalan society endorses rigid sexual and gender norms (Duarte et al. 2020). Previous work by members of our team found that a combined sample of sexual minority men and transgender women experienced a high prevalence of interpersonal violence (psychological: 29.8%, verbal: 38.2%, physical: 20.2%, sexual: 6.4%) (Wheeler et al. 2014). Additionally, among a sample of sexual minority men living with HIV, we found that 28.6% reported ever having experienced intimate partner violence (IPV) (Davis et al. 2020). Beyond interpersonal violence, sexual minority Guatemalans also report experiencing various forms of structural violence which has lasting impacts on their mental and physical wellbeing (Duarte et al. 2020; Rhodes et al. 2015). Structural violence goes beyond interpersonal violence “to explain the mechanisms through which social forces such as poverty, racism and gender inequity become embodied as individual experiences and health outcomes” (Shannon et al. 2017).

Sexual minority men who also identify as Indigenous likely experience intersecting forms of stigma which may increase their vulnerability to violence victimisation and HIV, as Indigenous Guatemalans have also historically faced violence and marginalisation (Lawton 2015). Guatemala is a pluricultural society in which 44% of the population identifies as Indigenous and 30% speak an Indigenous language as their first language (Instituto Nacional de Estadística de Guatemala 2019). Three Indigenous groups are officially recognised by the Guatemalan government: the Maya (comprised of 21 sociolinguistic subgroups and representing 95.6% of Indigenous Guatemalans), the Xinka (representing 4.1% of Indigenous Guatemalans), and the Garifuna (representing 0.3% of Indigenous Guatemalans) (Instituto Nacional de Estadística de Guatemala 2019). Since the arrival of the Spanish over 500 years ago, a variety of systems of oppression have evolved to ensure non-Indigenous groups (ladinos and mestizos, those of mixed European, Indigenous and African ancestry) maintain power and wealth in the country (Paredes 2017). As a recent example, a report published after the 36-year civil war ending in 1996 found that 83% of the more than 200,000 persons killed were Indigenous, attributing 93% of the war’s acts of violence to the army and state security forces (Comisión para el Esclarecimiento Histórico 1999). After the civil war, many former soldiers joined expanding criminal organisations that engage in arms trafficking, money laundering, extortion, human smuggling, and the drug trade, all of which disrupts social fabric of Guatemalan society (Grann 2011; Isaacs 2010).

The legacy of centuries of oppression is visible today as Indigenous Guatemalans are more likely to live in extreme poverty, have lower rates of educational attainment, and experience poorer health outcomes compared to their non-Indigenous counterparts (Lawton 2015; United Nations Human Rights Council 2018). This is partly due to the lack of access to health services for Indigenous Guatemalans, as 71% of medical providers are based in and around Guatemala City; and in the limited health services that exist in rural areas with higher proportions of Indigenous populations, few healthcare providers speak Indigenous languages (Ávila et al. 2015). These challenging conditions have forced many Indigenous sexual and gender minority persons in Guatemala to flee the country in search of a better life (Human Rights Watch 2020). In a recent study of US asylum seekers between 2012 and 2017, Guatemalan applicants represented 8.4% of the 11,400 LGBT applicants who claimed persecution for their sexual and/or gender minority status (Shaw et al. 2021).

While there is some anecdotal evidence that HIV and other STIs are a growing problem in Indigenous communities in Guatemala, epidemiological surveillance is lacking and suffers from underreporting (Minichiello, Rahman, and Hussain 2013; Orrego Dunleavy, Chudnovskaya and Simmons 2018; Russell et al. 2019; USAID 2010). In a study of three departments in Western Guatemala, Indigenous sexual minority men had roughly the same HIV prevalence as their non-Indigenous counterparts (5.1% v. 4.9%) (Mendizabal Burastero and Yancor 2017). In qualitative research with Indigenous gay and bisexual men, researchers report structural factors, such as stigma and discrimination; historical factors, such as broken family ties from the civil war; and environmental factors, such as drug trafficking through regions where a higher proportion of Indigenous Guatemalans live, as increasing Indigenous gay and bisexual men’s vulnerability to HIV and other STIs (Orellana, Alva and Yac 2014). In the one known study to compare HIV-related sexual behaviours between Indigenous and non-Indigenous sexual minority men and transgender women in Western Guatemala, researchers found that Indigenous participants were more likely to experience barriers to HIV testing and treatment than non-indigenous participants (86.3% vs 67.7%, p=0.004) (Ikeda et al. 2018).

To better understand the unique health experiences of Indigenous sexual minority men and the relationship between violence and HIV-related behaviours in this group, in this study we sought to compare different forms of violence, HIV-related behaviours, and STI symptoms by Indigenous identity among sexual minority men; and determine if associations between different forms of violence and HIV-related behaviours and STI symptoms differ by Indigenous identity among sexual minority men in Guatemala.

Methods

Parent Study

In 2015, Population Services International’s (PSI) regional affiliate, the Pan American Social Marketing Organization (PASMO), conducted a population-based survey in Guatemala to evaluate and improve HIV prevention efforts among sexual minority men. The survey was a follow-up to a 2011 study conducted by PASMO that aimed to study violence among key populations affected by HIV (Wheeler et al. 2014). Ethical approval was granted by PSI’s Research Ethics Board (REB), the Guatemala Ministry of Health’s Ethics Committee for the Protection of Human Subjects, and the Institutional Review Board at the University of North Carolina at Chapel Hill.

Recruitment and Procedures

The study team used respondent-driven sampling in the two largest cities in Guatemala, Quetzaltenango and Guatemala City, to recruit for the study. Respondent-driven sampling is a methodology widely used to sample ‘hard-to-reach’ populations, such as sexual minority individuals in a context such as this, while also attaining a representative sample by using statistical weights to control for peer and social network recruitment (Montealegre et al. 2013). The team recruited a sample of initial individuals, called seeds, in both cities based on the following criteria representing subgroups of sexual minority men (gay vs. bisexual vs. heterosexual); socioeconomic strata (low-, middle-, and high-income); and by sex work practices. The parent study was not designed to intentionally recruit Indigenous sexual minority men.

These seeds then began recruitment using chain referral methodology (Johnston et al. 2010). RDS uses a dual incentive system by having seeds recruit peers from within their social network. Participants were paid US$5.00 upon being interviewed and given US$5.00 for recruiting a maximum of three eligible peers to the study. Eligibility criteria included being a cisgender man between the ages of 18 and 40; living in the geographic area where the interview took place; having had anal sex with another man in previous three months; knowing by name the person that referred him; having seen in the previous month the person that referred him; being in possession of participant voucher.

In total, 716 participants were enrolled in the study and completed the survey. All participants provided verbal informed consent. The survey was implemented by CID Gallup and conducted in Spanish at the study site or a location indicated by the participant between May and July 2016.

Measures

Predictor Variables

Participants responded yes or no to the four measures of interpersonal violence collected as part of the study, which were drawn from the World Health Organization’s (WHO) guidelines on violence and health (WHO 2002):

• Psychological violence: having felt threatened, in fear or put in danger by somebody in the past 12 months.

• Verbal violence: having been yelled at, insulted, humiliated, or made felt inadequate by somebody in the past 12 months.

• Physical violence: having been slapped, punched, hit, or harmed physically by somebody in the past 12 months.

• Sexual violence: having been forced or coerced to have sexual relations against their will in the past 12 months.

Outcome Variables

The five dichotomous (yes/no) HIV-related factors below were used as outcome variables. These variables were chosen given evidence of their association with violence predictors and to be able to compare with earlier studies the team had conducted (Wheeler et al. 2014).

• Transactional sex: reporting partners with whom the participant received payment for sex in the past 30 days.

• Multiple partners: reporting more than one sex partner in the past 30 days.

• Sex under the influence of alcohol or drugs: reporting sex under the influence of alcohol or drugs with any type of partner in the past 30 days.

• Condom use at last sex: reporting the use of a condom at the last sexual encounter.

• STI diagnosis or symptoms: reporting abnormal or excessive genital discharge, an ulcer, sore, pustule, or excessive genital itching; and/or having an STI diagnosis in the past 12 months.

Moderator Variable

Indigenous identity:

participants were asked how they identified and given the options of selecting Afro-descendent or Garifuna, Indigenous (Maya or Xinka), Mestizo or Ladino, white, Asian, or other. Individuals who self-identified as Indigenous were categorised as such, while all others were categorised as non-Indigenous for the purposes of our analysis. While individuals who identified as Garifuna experience some similar structural barriers to health as the Maya and Xinka in Guatemala, their context is unique, and we therefore chose to only focus on participants who identified as Maya or Xinka.

Control Variables

The following four control variables were used in the multivariable analyses given evidence of their association with violence and HIV-related factors and to be able to compare with earlier studies the team conducted (Wheeler et al. 2014):

• Age: age of participant in years at time of data collection.

• Sexual Orientation: self-reported sexual orientation gay or homosexual, bisexual, or heterosexual (reference).

• Education: the highest reported educational attainment reported by the participant. These were collapsed into: some/completed primary, some/completed secondary, some/completed university, or some/completed high school (reference)

• Salary: self-reported monthly income (low=less than Q1,500/US$200, low-mid=Q1,501-3,900/US$201-500 (reference), upper-mid=Q3,901-8,500/US$501-1,100, high=Q8,501/US$1,101 or more).

• City: the site of data collection, Guatemala City or Quetzaltenango.

Data Analysis

Unweighted demographic characteristics of study participants and covariates are presented in Table 1. Individualised RDS estimator weights for each outcome variable were calculated using RDSAT 7.1 and exported to SAS 9.4, where the remainder of analyses were conducted.

Table 1.

Sample sociodemographic characteristics (N=716)

		n (%)
Age (mean & range)		25.39 (18-40)
Ethnicity
	Mestizo or Ladino	496 (69.27)
	Indigenous (Maya or Xinka)	135 (18.85)
	Garifuna or Black	22 (3.07)
	White	13 (1.82)
	Asian	6 (0.84)
	Other	1 (0.14)
Sexual Orientation
	Gay	387 (54.05)
	Bisexual	267 (37.29)
	Heterosexual	62 (8.66)
Civil Status
	Single	620 (86.59)
	Married	59 (8.24)
	Separated	24 (3.35)
	Divorced/widowed	12 (1.68)
Highest Education
	No education	4 (0.56)
	Some/completed Primary	81 (11.31)
	Some/completed Secondary	168 (23.46)
	Some/completed High School	293 (40.93)
	Some/completed University	157 (23.33)
	Postgraduate	3 (0.42)
Monthly Salary (GTQ)
	Low (Q1,500/US$200 or less)	227 (31.70)
	Low-mid (Q1,501-3,900/US$102-500)	292 (40.78)
	High-mid (Q3,901-8,400/US$501-1,100)	141 (19.70)
	High (Q8,501/US$1,101 or more)	16 (2.24)
City of Data Collection
	Guatemala City	544 (75.98)
	Quetzaltenango	172 (24.02)

To evaluate our first objective, we estimated the weighted prevalence of the different forms of violence (psychological, verbal, physical, sexual), HIV-related behaviours, and STI symptoms. Additionally, we conducted chi-square tests to assess if the weighted proportions of our variables of interest were significantly different (p<0.05) between Indigenous and non-Indigenous participants.

For our second objective, we used weighted logistic regression to examine associations between different forms of violence and HIV-related behaviours and STI symptoms. Due to their high correlation and conceptual overlap, we combined measures of psychological and verbal violence into a single variable. Due to the low prevalence of physical and sexual violence, we also combined these measures into a single variable.

Additionally, we assessed moderation of the direct effects of violence on sexual behaviours and STI symptoms by Indigenous identity using Hayes’ steps for testing moderation (Hayes 2018). Moderation analysis is a common method utilised in racial disparities research to examine the statistical significance and magnitude of an interaction term (or moderator) between ethnicity and predictor variables of interest (Knol and VanderWeele 2012). First, we ran a full regression model including all independent variables and the interaction between violence variables and Indigenous identity one at a time. If the interaction term was statistically significant, we probed the interaction to determine the effect of violence on sexual behaviours and STI symptoms by Indigenous identity. In all models, we controlled for age, sexual orientation, education, salary and city of data collection.

Results

Sample Characteristics

Table 1 reports sociodemographic characteristics of the total sample of 716 men. Their mean age was 25.4 years (range 18-40); the majority identified as gay (54.1%; n=387) and ladino or mestizo (69.3%; n=496). 40.9% (n=293) of men reported some high school as their highest level of education and 40.8% (n=292) earned between GTQ 1,501-3,900 monthly (USD $201-500). Nearly a quarter of participants completed the survey in Quetzaltenango (24.0%; n=172).

Prevalence of Violence Victimisation, Sexual Behaviours, and STI Symptoms

Weighted prevalence of different forms of violence victimisation, HIV-related behaviours and STI symptoms are presented in Table 2. Among the entire sample, verbal violence was the most commonly reported form of violence experienced in the past 12 months (21.3%, n=159), followed by psychological violence (15.1%, n=113), physical violence (10.3%, n=77) and sexual violence (3.5%, n=27). Indigenous men reported less psychological, verbal, physical and sexual violence compared to non-Indigenous participants, however these differences were not statistically significant.

Table 2.

Weighted prevalence of violence, sexual behaviours and STI symptoms among study participants (N=716)

		Total	Indigenous†	Non-Indigenous†	chi-square p-value
n (%)		716 (100.0)	170 (24.8)	489 (74.2)
Violence
	Psychological	113 (15.1)	21 (12.4)	83 (17.0)	0.33
	Verbal	159 (21.3)	24 (14.1)	118 (24.1)	0.06
	Physical	77 (10.3)	15 (8.8)	52 (10.6)	0.72
	Sexual	27 (3.5)	5 (2.9)	17 (3.5)	0.88
Sexual Behaviours and STIs
	Transactional sex	134 (17.9)	15 (8.8)	99 (20.2)	0.02*
	Sex under influence drugs or alcohol	196 (26.2)	30 (17.6)	142 (29.0)	0.06
	Multiple partners	483 (64.5)	99 (58.2)	329 (67.2)	0.29
	Condom use	637 (85.3)	146 (85.9)	422 (86.7)	0.80
	Reported STI diagnosis/symptoms	66 (7.5)	11 (6.5)	35 (7.2)	0.86

^†Proportions shown in Indigenous and non-Indigenous columns only include participants who reported their ethnicity (N=693)

^*p-value ≤ 0.05 and considered statistically significant

Nearly one fifth of all participants reported transactional sex (17.9%, n=134) and a quarter reported sex under the influence of alcohol or drugs (26.2%, n=196). The majority of participants reported having more than one sexual partner in the previous 30 days (64.5%, n=483) while 85.3% (n=637) reported using a condom during their last sexual encounter. 7.5% (n=66) of participants reported an STI diagnosis or symptom. Indigenous men reported significantly less transactional sex (8.8% vs. 20.2%, p=0.02) compared to non-Indigenous men. Fewer Indigenous participants reported sex under the influence of drugs or alcohol and having multiple partners compared to non-Indigenous participants, however these differences were not statistically significant.

Multivariable Analyses

In weighted multivariable analyses with the entire sample, participants who experienced psychological and/or verbal violence had significantly higher odds of reporting an STI diagnosis or symptoms (aOR: 4.66; 95% CI 2.03-10.68; p<0.001) compared to those who had not, after adjusting for covariates (Table 3). Psychological and/or verbal violence was not significantly associated with reporting transactional sex, multiple partners, sex under the influence of drugs or alcohol or condom use. Participants who experienced physical and/or sexual violence had significantly higher odds of reporting transactional sex (aOR: 3.03; 95% CI 1.34-6.85; p<0.001) compared to those who had not; but there was no significant association with multiple partners, sex under the influence of drugs or alcohol, condom use, or STI diagnosis or symptoms.

Table 3.

Weighted multivariate results of experiences of violence on sexual behaviours (N=702)

	Transactional sex		Multiple partners		Sex under influence drugs or alcohol		Condom use		Reported STI diagnosis/symptoms
	aOR	95% CI	aOR	95% CI	aOR	95% CI	aOR	95% CI	aOR	95% CI
Psychological/Verbal violence†	1.05	0.57-1.94	1.25	0.67-2.35	1.66	0.91-3.04	0.67	0.36-1.26	4.66	2.03-10.68*
Physical/sexual violence†	3.03	1.34-6.85*	0.52	0.22-1.19	2.10	0.97-4.54	0.75	0.33-1.72	1.25	0.47-3.36

^†Models controlled for transactional sex^††, multiple partners^††, sex under the influence of drugs or alcohol^††, age, sexual orientation, education, salary, city of data collection

^††These sexual behaviours were not included as control variables in the models for which they were the outcome variables

^*95% confidence intervals considered statistically significant because they do not cross 1

Moderation Analyses

In moderation analyses (Table 4), we found that there was a statistically significant interaction between physical and/or sexual violence and Indigenous identity in predicting transactional sex (b=−3.52, p<0.001). Non-Indigenous participants who had experienced physical and/or sexual violence had over five times the odds of reporting transactional sex (aOR=5.17, 95% CI 2.11-12.68, p<0.001), compared to those who did not report physical and/or sexual violence. Participants who identified as Indigenous and reported physical and/or sexual violence had less odds of reporting transactional sex, although this finding was only marginally significant (aOR=0.15, 95% CI 0.022-1.04, p=0.06) (Figure 1).

Table 4.

Moderating effect of Indigenous identity on relationship between violence and sexual behaviour (N=693)^†

	Transactional sex				Multiple partners		Sex under influence drugs or alcohol		Condom use		Reported STI diagnosis/symptoms
	Estimate	p-value	aOR	95% CI	Estimate	p- value	Estimate	p- value	Estimate	p- value	Estimate	p- value	aOR	95% CI
Psychological/Verbal violence††	0.32	0.37	-	-	0.35	0.38	0.37	0.30	−0.77	0.05	1.71	<0.001	-	-
Indigenous Identity	−0.14	0.76	-	-	−0.22	0.60	−0.38	0.35	−0.53	0.27	0.86	0.12	-	-
Psy/Verb violence X Indigenous Identity	−1.96	0.04*	-	-	−0.52	0.47	0.39	0.57	1.69	0.06	−1.07	0.28	-	-
Indigenous	−1.64	0.06	0.19	0.034-1.10	-	-	-	-	-	-	-	-	-	-
Non-Indigenous	0.32	0.37	1.38	0.68-2.80	-	-	-	-	-	-	-	-	-	-
Physical/sexual violence††	1.64	<0.001	-	-	−0.51	0.29	0.25	0.59	−0.56	0.29	0.48	0.41	-	-
Indigenous Identity	−0.16	0.68	-	-	−0.35	0.37	−0.44	0.25	−0.31	0.46	0.82	0.09	-	-
Phys/Sexual violence X Indigenous Identity	−3.52	<0.001*	-	-	0.07	0.94	1.06	0.23	1.53	0.13	−3.10	0.03*	-	-
Indigenous	−1.87	0.06	0.15	0.022-1.04	-	-	-	-	-	-	−2.62	0.05*	0.07	0.0056-0.95
Non-Indigenous	1.64	<0.001*	5.17	2.11-12.68	-	-	-	-	-	-	0.48	0.41	1.61	0.52-4.96

^†Table includes only participants who reported their ethnicity (N=693)

^††Models controlled for transactional sex^†††, multiple partners^†††, sex under the influence of drugs or alcohol^†††, age, sexual orientation, education, salary, city of data collection

^†††These sexual behaviours were not included as control variables in the models for which they were the outcome variables

^*p-value ≤ 0.05 and considered statistically significant

Figure 1.

Moderating effect of Indigenous identity on the relationship between physical and/or sexual violence and transactional sex among (N=693)

There was also a statistically significant interaction between psychological and/or verbal violence and Indigenous identity in predicting transactional sex (b=−1.96, p=0.04). However, upon probing the interaction neither estimate provided statistically significant (Figure 2). We also found a statistically significant interaction between physical and/or sexual violence and Indigenous identity in predicting STI diagnosis or symptoms (b=−3.10, p=0.03). Indigenous participants who experienced physical and/or sexual violence had a 93% decrease in the odds of reporting an STI diagnosis or symptom compared to Indigenous participants who did not experience physical and/or sexual violence (aOR=0.07, 95% CI 0.006-0.95, p=0.05). Among non-Indigenous participants, exposure to physical and/or sexual violence had no effect on STI diagnosis or symptoms (p=0.41) (Figure 3). No other interaction terms between violence variables and Indigenous identity were significant in predicting other sexual behaviours or STI symptoms.

Figure 2.

Moderating effect of Indigenous identity on the relationship between psychological and/or verbal violence and transactional sex (N=693)

Figure 3.

Moderating effect of Indigenous identity on the relationship between physical and/or sexual violence and STI diagnoses or symptoms (N=693)

Discussion

Gay, bisexual and other men who have sex with men in Guatemala experience a high burden of interpersonal violence and report behaviours that are associated with HIV and STI transmission and acquisition. To our knowledge, this is the first study to quantitatively examine differences in prevalence of different forms of interpersonal violence between Indigenous and non-Indigenous sexual minority individuals in Guatemala. We found that Indigenous participants reported less interpersonal violence than their non-Indigenous counterparts, although most of these differences were not statistically significant. While racism and xenophobia lead to high levels of stigma among Indigenous Guatemalans, this stigma is also likely the result of structural violence, and therefore may not manifest in the different forms of interpersonal violence we measured. Future studies among Indigenous sexual minorities should consider employing a structural violence framework to explore the mechanisms through which social forces such as poverty, racism, homophobia and gender inequity affect HIV-related behaviours and other health outcomes (Farmer 2004; Ho 2007). Other researchers have successfully used such a framework to identify facilitators of both social wellbeing and oppression that influence individual health and disease outcomes in Latin America, including in Indigenous communities (Duarte et al. 2020; Shannon et al. 2017; Orellana, Alva and Yac 2014). Shannon and colleagues, for example, found that using a structural violence framework elucidated how systems of oppression contributed to high rates of interpersonal violence among Indigenous groups in Peru (Shannon et al. 2017). For groups with multiple marginalised identities, such as Indigenous sexual minority men, future research should also consider using measures of violence that capture perceived reasons for violence (i.e. sexual orientation or ethnicity) for a more complete understanding of experiences of interpersonal violence.

In general, our findings that fewer Indigenous participants reported sexual behaviours associated with HIV and STI transmission and acquisition may be because Indigenous sexual minority men in Guatemala have higher perceived risk to HIV and other STIs. This is supported by a recent study from Ikeda and colleagues (2018) who found that compared to their non-Indigenous counterparts, Indigenous sexual minority men in Guatemala perceived themselves to be at a higher risk for HIV acquisition (87.7% vs. 51.6%, p=0.001), endorsed the benefits of HIV testing and treatment at a higher rate (79.5% vs. 63.1%, p=0.04), and were more likely to have had an HIV test in the previous 12 months (97.3% vs. 85.4%, p=0.008). Additional research is needed however to understand why Indigenous sexual minority men have a higher perceived HIV risk compared to their non-Indigenous counterparts, including the contribution of HIV prevention efforts in Western Guatemala. These findings are contrary to several studies in Guatemala that have shown less perceived HIV risk and HIV knowledge among heterosexual Indigenous individuals compared to their non-Indigenous counterparts; however many of these latter studies were conducted in more rural settings and so may not be comparable to Indigenous sexual minority men in an urban environment (Dunleavy, Phillips and Chudnovskaya 2019; Marín Marroquín, Moreira and Monzón 2013; Taylor, Hembling and Bertrand 2015; Blasco Hernandez et al. 2012). It is also possible that Indigenous participants from our sample, which were recruited from urban centres, may experience interpersonal violence and have HIV-related behaviours different from Indigenous sexual minority men in rural settings.

Although previous studies in Guatemala show that gay and bisexual men often engage in sex work because they have limited employment opportunities and are rejected by their families (Miller et al. 2019; Munson, Davis and Barrington 2020), a possible explanation of our finding that fewer Indigenous participants engaged in transactional sex may be that Indigenous sexual minority men experience more support from their families or communities. Evidence exists which supports the notion that some Indigenous communities in the region may have been more accepting of sexual and gender fluidity compared to Western cultures prior to the Spanish conquest (Picq 2020; Olivier 2010). However, to our knowledge this has not been explored among the present-day Maya of Guatemala. Future ethnographic research on sexual and gender fluidity among the Maya in Guatemala is needed to explore this topic in more detail. It is also possible that this finding is spurious and may be the result of social desirability bias. For example, Indigenous participants may have been less comfortable responding to questions about sexual practices and other behaviours considered taboo. Additionally, traditional mores related to sex work may differ between Indigenous and non-Indigenous communities. Future qualitative research should centre Indigenous sexual minority men and explore topics of sexuality and HIV-related behaviours to determine the best way to ask about HIV-related behaviours quantitatively.

When compared to findings from an earlier study by Wheeler and colleagues (2014) among 3,829 sexual minority men and transgender women from five Central American countries, our sample from Guatemala found a stronger effect of certain forms of violence on some HIV-related behaviours. For example, participants in our study who reported physical and/or sexual violence were more likely to report STI diagnoses compared to the larger Central American sample (aOR: 4.66 vs. 1.64, respectively) and those in our sample who reported psychological and/or verbal violence were more likely to report transactional sex compared to the larger sample (aOR: 3.03 vs.1.74, respectively) (Wheeler et al. 2014). This may speak to the unique context of Guatemala and its high levels of generalised violence compared to other countries in the Central American region (Wheeler et al. 2014). Our examination of the moderating effect of Indigenous identity on these relationships strengthens the assertation that experiences of violence and their impact on HIV-related behaviours may be different for Indigenous sexual minority individuals. For example, we found that a significant relationship exists between physical and/or sexual violence and transactional sex only among non-Indigenous participants. Further qualitative research among Indigenous and non-Indigenous sexual minority men who have experienced violence and participated in transactional sex and other behaviours associated with HIV is needed to understand what unique aspects of Indigenous identity, if any, may play a role in this relationship, including potentially different perceptions of what constitutes violence. While the cross-sectional nature of our data does not allow us to assess causality, these findings suggest that health services that aim to decrease HIV and STI prevalence among sexual and gender minority individuals in Guatemala should incorporate interpersonal violence prevention and treatment efforts into their programming.

In order to better understand the complex relationship between violence, sexual behaviours, and HIV and STI transmission in Guatemala, future national surveys must ask about ethnicity, sexual orientation, and gender identity, as well as recruit samples from both urban and rural settings. Noting the scarce data on HIV and STIs among Indigenous populations in Latin America, García and colleagues (2014) identified monitoring HIV/STI prevalence trends and developing culturally appropriate sexual health programmes and prevention strategies for Indigenous communities as a regional priority. Additionally, qualitative research with Indigenous sexual and gender minority individuals in Guatemala is needed to explore in more detail their experiences with violence, both structural and interpersonal, and how these experiences interplay with sexual behaviour. While public health research, especially that which focuses on LGBTQ and Indigenous communities, is often deficit-based, inspired by our findings, we challenge ourselves and other researchers to employ a strengths-based approach in future research with Indigenous sexual and gender minority Guatemalans; this will help bolster existing capacities in these communities to address violence and STIs, rather than focusing on community limitations which are often the result of centuries of oppression and intergenerational trauma (Hyett et al. 2019).

Limitations

There are several limitations to this study. As mentioned earlier, the cross-sectional nature of our data does not allow us to assess causation or temporality of these associations, and low power may have impacted our ability to see effects, especially for physical and/or sexual violence. Additionally, all independent and dependent variables were assessed using self-reported measures, which may lead to recall bias.

We dichotomised Indigenous identity and therefore were unable to assess differences between Indigenous groups. As there are 23 officially recognised Indigenous groups in Guatemala, each with its own culture, history and traditions, future research with Indigenous sexual minority individuals in Guatemala should engage with these differences and collect data on Indigenous group accordingly.

Furthermore, our sample was not specifically designed to capture Indigenous sexual minority men and participants were recruited in the two largest cities in Guatemala, Quetzaltenango and Guatemala City. While it is likely that some participants resided outside if these urban centres, our findings are not generalisable to sexual minority men in other cities or rural settings. It is important that future research among Indigenous sexual minority individuals actively recruit participants from both rural and urban settings to gain a more complete understanding of their experiences.

Additionally, surveys were only offered in Spanish, which may have prohibited the participation of Indigenous sexual minority men whose first language was not Spanish. Future research with Indigenous sexual minority individuals should ensure questionnaires can be administered in the first language of the participants. Non-Spanish-speaking Indigenous gay and bisexual men living in rural areas likely experience additional obstacles to accessing health services and may suffer other forms of stigma and violence.

Conclusion

Interpersonal violence is widespread among gay, bisexual and other men who have sex with men in Guatemala. In this study, however, fewer Indigenous sexual minority men reported interpersonal violence victimisation, HIV-related sexual risk behaviours and STIs compared to their non-Indigenous counterparts. Additional strengths-based research is needed to explore the unique features of Indigenous sexual and gender minority individuals and their lifestyles and experiences to ensure that violence, HIV, and STI prevention and treatment services engage with relevant protective factors and meet their future needs.