Introduction

Male sex workers (MSWs) comprise a heterogeneous population with different backgrounds and behaviors. Unfortunately, little is known about their psychological characteristics and profiles, probably because of the surreptitiousness of prostitution and taboos that surround it. This has delayed an effective development of HIV/AIDS prevention and management programs for such a population.

According to the Brazilian National Program for STD and AIDS, almost 630,000 individuals aged between 15 and 49 are infected with HIV/AIDS. Although the prevalence of HIV infection in the population at large has remained stable at around 0.6% since 2004, the most-at-risk subgroups of the population, such as illicit drug users, men who have sex with men (MSM), and female sex workers show HIV prevalence rates of about 5.9%, 12.6%, and 4.9%, respectively (Ministério da Saúde, 2010). Despite the difficulties to reach male sex workers, the HIV prevalence rates have been elusively estimated at about 22% and 40% for male hustlers and transgender sex workers, respectively (Grandi, Ueda, Goihman, Santos, & Amorim, 2001). To date, the UNAIDS (2009) declares that only a small percentage of the prevention programs have focused on these most-at-risk groups.

Notwithstanding the heterogeneity of male sex workers, some common characteristics are frequently detected among them, such as poor education, disrupted upbringings, low socioeconomic level, large number of sexual partners, unsafe sexual relations, alcohol and drug misuse, and history of criminal activities (Brannigan & Van Brunschot, 1997; Caukins & Coombs, 1976; Sausa, Keatley, & Operario, 2007; Scott et al., 2005; Timpson, Ross, Williams, & Atkinson, 2007). Also, comparing the group of male with female sex workers, the former seems to be more exposed to risky sex behaviors than the latter (Belza, 2005; Dandona et al., 2006).

Although infrequent, some researchers have tried to group male prostitutes according to the main motivation to enter such a profession, their self image, the gender role assumed during the commercial activities, and the types of sexual services offered as a way to better reduce the heterogeneity, evaluate this population, and develop more effective health proposals (Dorais, 2005; West, 1993). In fact, gender performance can be an important variable to understand high-risk behaviors between MSWs and their clients, since clients might look for specific types of sex workers. While male-to-female transvestites or transsexuals look like women when they engage in sex work, male hustlers emphasize their masculine appeal. Thus, transgender sex workers may be more commonly serving clients who consider themselves heterosexual and hustlers may be having commercial sex with self-identified homosexual or bisexual clients. In fact, both groups of MSWs partake in high-risk sexual activities and have a great number of sexual partners; however, they can attract clients from different sexual orientations, preferences, and sex-related attitudes, which may imply different patterns of HIV spread within the community (Grandi, Goihman, Ueda, & Rutherford, 2000).

Also, the gender performance can reflect the way that the sexual work is self-conceived. Because of their physical appearance, transgender sex workers face more intense discrimination, even when they are not sex-working. This certainly contributes to making a “transvestite identity” a viable option, not only for earning money but also for getting a supportive and protective group. On the contrary, many male hustlers state that prostitution must remain circumstantial for them, that it must be completely separated from the rest of their lives. Thus, many male hustlers may not regard themselves as prostitutes, but as temporary professionals waiting for other job opportunities (Dorais, 2005; Perlongher, 2008).

Since no specific and evidence-based HIV prevention programs have been developed for male sex workers, there is an urgent need to understand intrinsic and extrinsic factors underlying their sexual behaviors. Once understood, more suitable interventions could be used to meet the unique prevention needs of this population (Baltieri, Prado Cortez, & de Andrade, 2009). The term “prevention” must involve something more than simply distributing condoms or even teaching how to use them correctly; an adequate prevention strategy aims to engage more people in preserving and improving their own health and well-being as a whole. This is only possible if the population we are focusing on is carefully studied, well known, and their needs met.

Despite the critical importance of assessing intrinsic factors between transgender sex workers and male hustlers, such as personality-related aspects and impulsiveness, virtually no research has focused on this yet. For instance, to date, few studies have evaluated the psychological characteristics of transgender people that resort to prostitution and this contributes to the application of powerless strategies, in terms of health public policy.

Despite the lack of scientific literature investigating psychometric features between transgender and male hustlers, we have hypothesized that both groups of hustlers diverge in terms of personality aspects, depressive symptoms, and drug misuse. As trans-people have reported suffering more violence and prejudice than male hustlers even before starting to work as prostitutes (Baltieri & de Andrade, 2009; Baltieri et al., 2009; Johnson, Brett, Roberts, & Wassersug, 2007; Khan et al., 2009) and maintaining strong bonds with peers based on a sharing identity (Collumbien et al., 2009), we predicted that they would show more depressive symptoms and score less on novelty seeking (which is strongly correlated with a search for thrill) and more on harm avoidance (which is directly correlated with fear and behavioral inhibition) than male hustlers. On the contrary, since some studies on male hustlers have described many of them as heavy drug users and frequently involved in criminal activities (Boyer, 1989; Earls & David, 1989; Falcon, 2007), we predicted that these sex workers would show more of these problems than the group of transgender sex workers.

We recognized that there are some theoretical problems when one categorically tries to define “transgender” because this word has represented an “umbrella term” that encompasses anyone who bends the common societal constructions of gender. Thus, cross-dressers, transvestites, transsexuals, drag-queens, and a host of other terms can be included as “transgender”. However, for the purpose of this study, we used the term “transgender” to designate those male sex workers that emphasized a strong appearance while sex working.

Method

Participants

The inclusion criteria for the selection of participants for this study were males aged 18 years or older, currently working as street-based sex workers, and offering their services to other men in Santo André, São Paulo, Brazil. Male-to-female transgender sex workers were defined as men who wear female clothing and use hormonal therapies and/or have breast implants to acquire a feminine appearance; male hustlers were defined as men who dressed as men, emphasizing their male appeal. “Sex work” is a term for any activity involving sex primarily for remuneration. Given that street prostitution is legal in Brazil, but maintaining brothels and encouraging or taking advantage of prostitutes is a crime, we planned our study with street sex workers only. Despite this, all the main types of MSWs are found in Brazil: independent MSWs, MSWs who work in or through agencies, and street MSWs.

About 28 MSWs refused to participate in this study, mainly because they would not be paid for their time. According to the Brazilian legal determinations related to ethics in research, no remuneration is given for the participation in any type of research. Between February 2008 and April 2010, 86 MSWs were interviewed. Of these, 45 (52.33%) cross-dressed and 41 (47.67%) wore men’s clothing full time. Overall, the mean age was 26.60 (SD = 7.44) years, 32.6% were white, and 84.9% were single.

Procedure

This investigation was a quantitative cross-sectional study performed by the staff of the Ambulatory for the Treatment of Sexual Disorders of ABC Medical School (ABSex), in São Paulo. Three professionally instructed interviewers, who were trained in health promotion on the commercial sex scene, carried out fieldwork.

Our recruitment of participants concentrated on men who currently engaged in sex work on a part- or full-time basis. The planning of the recruitment strategy commenced with estimating the size of the pool of potential participants by assessing how many street sex workers were working in Santo André. First, 12 MSWs were recruited through referral from the Harm Reduction Program of Santo André City Hall. More participants were then obtained through application of the “snowball” sampling procedure. This procedure is often used for studies on hidden or hard-to-reach groups. This means that each participant was used as a resource for the identification of one or more male sex workers in the same region.

Our subjects participated in face-to-face interviews that addressed topics concerning their work and private lives in relation to their reported safer sex practices. All interviews were conducted inside a special van, well equipped for this purpose (commonly used by the Harm Reduction Program staff), at the same place where our participants were working at night. Each interview lasted about 100 min and apart from the researchers and the interviewee, no one else was allowed in the van during the interviews to avoid confidentiality violation.

Participants were assured of confidentiality and anonymity and were told of their right to end the interview if they felt any psychological discomfort. A written consent was obtained from participants before all procedures. This study was approved by the Ethics Committees of the Clinical Hospital of the University of São Paulo, of ABC Medical School and of the Municipality of Santo André.

Measures

Temperament and Character Inventory (TCI)

The Brazilian version of the TCI was developed in several steps according to established guidelines, including translations, back-translations by native speakers, population testing, and revisions of items to more colloquial language (Fuentes, Tavares, Camargo, & Gorenstein, 2000). TCI scores were converted into T-scores according to published normative data and previous research (Arnau, Mondon, & Santacreu, 2008). T-score has a normal distribution with a mean of 50 and a SD of 10. These dimensions were divided into four scales of Temperament and three of Character.

The temperament scales are: (1) Novelty seeking (NS): tendency towards exploratory activity in response to novelty, lack of inhibition, impulsiveness. This dimension is related to the level of control and excitability and corresponds to the sum of four subscales measuring more specific traits: Exploratory excitability (11 items), Impulsiveness (10 items), Extravagance (9 items), and Disorderliness (10 items); (2) Harm avoidance (HA): tendency to anxiety, shyness, worry and avoidance of punishment. This dimension is assessed by means of four subscales: Anticipatory worry (11 items), Fear of uncertainty (7 items), Shyness (8 items), and Fatigability (9 items); (3) Reward Dependence (RD): attachment and social attachment systems. It is associated with dependence on external approval. This dimension encompasses three subscales: Sentimentality (10 items), Attachment (8 items), and Dependence (6 items); (4) Persistence (P): capacity to maintain behavior in adverse conditions. It is characterized by making demands on self, hard working, striving for excellence. It has a single 8-item scale.

The Character scales are: (1) Self-directedness (SD): related to maturity, strength, and self-sufficiency. Capacity to manage behavior guided by goals chosen voluntarily and individually and not by circumstances, impulses or external stimulus. This dimension was assessed as the sum of five subscales measuring more specific related traits: Responsibility (8 items), Purposefulness (8 items), Resourcefulness (5 items), Self-acceptance (11 items), and Congruent second nature (12 items); (2) Cooperativeness (C): reveals an inclination towards social tolerance, empathy, friendliness, altruism, and respect for others. This dimension includes: Social acceptance (8 items), Empathy (7 items), Helpfulness (8 items), Compassion (10 items), and Pure-heartedness (9 items); (3) Self-transcendence (ST): reflects a tendency towards spirituality, idealism, religious or mystical feelings, and identification with the wider world, as well as the ability to accept ambiguity and uncertainty, and a sense of communion with others. This dimension encompasses three subscales: self-forgetfulness (11 items), transpersonal identification (9 items), and spiritual acceptance (13 items).

Drug Abuse Screening Test (DAST)

It was constructed to provide a quantifiable self-report instrument for use in clinical and non-clinical settings to detect drug misuse pertaining to a range of psychoactive drugs. The original version of the DAST contains 28 yes–no questions that can be administered in a self-report or clinician-interview format. The overall score ranges from 0 to 28 based on the sum of individual items. A cutoff of 6 or more indicates a probable drug-use problem (Skinner, 1982). Cronbach’s alpha ranges from 0.85 to 0.92 in samples of clinical and non-clinical populations (Yudko, Lozhkina, & Fouts, 2007).

Alcohol Use Disorders Identification Test (AUDIT)

It was designed to detect alcohol consumption that has become hazardous or harmful to health in a range of clinical and non-clinical settings. It was constructed to recognize people with early-stage alcohol problems. This questionnaire consists of 10 multiple-choice items, out of which 3 are on the quantity and frequency of alcohol use, 3 on harmful drinking, and 4 on hazardous drinking. All of the questions were scored using a 5-point Likert scale. Its Cronbach’s alpha is about 0.80 (Bohn, Babor, & Kranzler, 1995; Martino, Grilo, & Fehon, 2000).

Barratt Impulsiveness Scale (BIS-11)

It is a self-report scale composed of 30 items with Likert-type questions which provide a total score of impulsivity and three sub-scores: attention, lack of planning, and motor impulsivity. Scores range from 30 to 120 and there is no established cut-off point. Cronbach’s alpha ranges from 0.79 to 0.83 in large samples of undergraduates and clinical and prison populations (Patton, Stanford, & Barratt, 1995; von Diemen, Szobot, Kessler, & Pechansky, 2007).

Beck Depression Inventory (BDI)

It was developed to measure behavioral responses related to depression among adults and adolescents. In this 21-item instrument, each one with a series of four statements, scores above 10 have been associated with depressive syndrome. Scores range from 0 to 63, and for non-psychiatric samples Cronbach’s alpha ranges from 0.73 to 0.90 (Beck, Rial, & Rickels, 1974; Gorenstein & Andrade, 1996).

We also used a structured questionnaire developed by the Ambulatory for the Treatment of Sexual Disorders of ABC Medical School, Santo André, São Paulo, Brazil that focuses on the following topics: sociodemographic data, personal history of alcohol and drug usage during work, family history of alcohol and drug problems and involvement in criminal offences, personal history of being sexually and/or physically abused in childhood, person who committed the sexual and/or physical abuse, employment history, previous convictions or charges, and the age in which alcohol and drugs were initially used. In addition, an anonymous pre-coded questionnaire was applied for indicators of knowledge, attitudes, and practices on sexuality and HIV/AIDS.

Data Analysis

A multivariate analysis of variance (MANOVA) was used because this study evaluated many dependent variables and only a univariate analysis could increase the probability of missing data. To test the assumption that variance–covariance matrices within each cell of the design were sampled from the same population variance–covariance matrix, we used Box’s M test, a sensitive test of homogeneity of variance–covariance matrices. Also, direct logistic regression analysis (Wald’s method) was constructed to investigate the associations between the types of MSWs and the variables whose significance levels were below .10 in univariate analyses. Given that all predictors enter the equation simultaneously (as long as tolerance is not violated), a predictor that is highly correlated with the outcome by itself may show little predictive capability in the presence of other highly correlated predictors. Therefore, we also computed the correlation matrix of the variables included in the logistic regression analysis. Pearson, phi and point-biserial correlation coefficients were used to indicate relationships between two continuous variables, two categorical variables, and between one continuous variable and other categorical one, respectively.

Results

Descriptive Statistics

As shown in Table 1, there were no significant differences between the two groups with regard to age, marital status, race, educational level, religious affiliation, history of being sexually/physically abused in childhood, and consistent condom use with their clients. However, transgender hustlers earned more money than male hustlers, were more frequently hurt by clients, and lived in hostels. Nineteen (46.34%) male hustlers had other jobs besides prostitution, as against only 5 (11.11%) transgender hustlers. Also, male hustlers mentioned more frequently a family history of criminal activities and discipline problems in elementary school than the group of transgender hustlers.

Table 1 Sociodemographic features and prostitution-related factors among transgender and male hustlers
Full size table

Psychometric Measures

A 2 × 4 MANOVA was conducted with the MSW groups as the independent variable and the temperament-related factors total scores as the dependent variables. Table 2 shows means and SD for these variables. The MANOVA was significant, Pillai’s F(4, 81) = 3.08, p = .02, ∂n 2 = .13. An analysis of univariate effects showed a reliable difference only for the HA total score. This MANOVA model demonstrated homogeneity of variance for the individual dependent variables, as verified by Box’M test (M = 13.31, F = 1.26).

Table 2 Alcohol and drugs problems, impulsiveness levels and depressive symptoms between transgender and male hustlers
Full size table

Another 2 × 4 MANOVA was applied with the DAST, AUDIT, BIS-11, and BDI total scores as the dependent variables. Table 3 shows means and SD for these variables. This model was also significant, Pillai’s F(4, 81) = 3.79, p < .01, ∂n 2 = .16. In the analysis of univariate effects, only the BDI demonstrated significant differences between both groups, with the group of trans-people showing higher values. This model demonstrated homogeneity of variance (Box’M test = 11.88, F = 1.13).

Table 3 Temperament and character dimensions (T scores) between transgender and male hustlers
Full size table

A direct logistic regression was performed on group status as outcome and 10 predictors whose significance levels were below .10: harm avoidance, depression symptoms, monthly income, religion, other job, place of residence, report of physical aggression by clients, runaway history, school problems, and family history of criminality. A test of the full model with all predictors against a constant-only model was statistically reliable, χ 2 = 63.25, 10 df, p < .01.The variance in group membership accounted for was marginal, with Nagelkerke R 2 = .69. The overall success rate for prediction was 84.9%. We checked the model fit using Hosmer and Lemeshow’s test, χ 2 = 5.44, 8 df, ns. As shown in Table 4, five variables—other job besides prostitution, harm avoidance, monthly income, place of residence and severe discipline problems in the elementary school—reliably predicted group membership, according to the Wald’s criteria.

Table 4 Effects of harm avoidance, depression symptoms, sociodemographic, and prostitution-related factors on the group of transgender sexual workers (direct logistic regression)
Full size table

Discussion

Overall, the data showed that transgender sex worker and male hustlers differed with regard to some sociodemographic variables, vulnerability experiences, and personality traits. The findings indicated that male-to-female transgender sex workers had fewer conventional job opportunities, fewer school problems, and higher harm avoidance levels than male hustlers. Also, transgender sex workers reported earning more money through sex work and more frequently living in hostels with peers than their counterparts. Confirming our initial hypotheses, harm avoidance set apart transgender from male hustlers.

Many of these findings were consistent with the literature on social and psychological characteristics of transgender and male sex workers. A relevant aspect of this study was the use of validated psychometric instruments to evaluate if personality traits, depressive symptoms, and drug misuse could be related to the behavior of biological male sex workers at least to some extent.

In Brazil, the majority of transgender sex workers are migrants from small and poor cities, where much more conservative moral values prevail (Kulick, 1998). They search for economic opportunities and a certain degree of sexual freedom. However, due to scarce family financial resources and emotional support to facilitate entry into other activities, difficulty of finding work in a large city, and low educational level, they tend to resort to prostitution (Garcia, 2009). In the metropolis, transgender sex workers tend to live in specific communities that are not only stigmatized but also marginalized from general society. Therefore, their gender identity status overlaps with other vulnerabilities, such as poverty, joblessness, high HIV prevalence, and homelessness (Hwahng & Nuttbrock, 2007; Kulick, 1998).

In this research, transgender sex workers have usually lived together in hostels. They share their spaces and live-experiences, organizing a kind of “family,” and creating strong social structures based on a shared identity. On the contrary, although male hustlers can also divide their lives with peers, they commonly migrate from one city or place to another, lest they be recognized by friends or family members (Collumbien et al., 2009). Both groups are stigmatized; however, trans-people are more promptly identified due to their physical appearance. Indeed, most trans-people, if not all, know too well the consequences of diverging from compliance with the definition and appearance of what is considered a “normal” gender expression. Since the heteronormativity is the rule in our society, people feel isolated if the idea of “normal” is not what they are or even practice. As far as those who do not assimilate with the heteronormative view, one’s status is put on the line (Perlongher, 2008).

Certainly, human behavior, emotion, and thought are pervasively influenced by an essential interpersonal motive to obtain acceptance and to avoid rejection. Negative emotions that arise from perceived rejection and consequent social ostracism can lead to a decrease of self-esteem (Smart Richman & Leary, 2009). When people are rejected or labeled, they can look for support and acceptance in certain groups that have also been discriminated against. It is probable that transgender sex workers try to avoid further rejection and its accompanying pain and, as a result, circumvent any possibility of losing the gained peer acceptance. High harm avoidance, as a temperamental trait, correlates positively with negative stress-coping strategies, that is, rumination, escape, resignation, and self-blame. Thus, temperament may influence stress vulnerability or even stress resiliency, not only by affecting coping styles, but also by impairing appraisals of the potentially stressful situations (Ravaja, Keltikangas-Jarvinen, & Kettunen, 2006). Thus, a long-standing stigmatization and consequent rejection can shape perceptions concerning other people’s reactions and lead to cautiousness as regards social interactions in which one might be devalued or rejected (Onoda et al., 2010; Shelton, Richeson, Salvatore, & Trawalter, 2005).

Our study also showed that transgender sex workers more frequently suffered physical violence than their counterparts. It is true that being physically maltreated is also associated with high stress vulnerability and worsening of self-esteem. Although violence is common in the lives of sex workers in general, trans-people seem to be at higher risk of work, customer, and police violence than male and female sex workers (Cohan et al., 2006; Rhodes, Simic, Baros, Platt, & Zikic, 2008). Probably, perpetrators are motivated by hatred or negative attitudes towards trans-people in a heteronormative society. Also, the episodes of physical and sexual violence against transgender sex workers tend to be repetitive, foment their ostracism and avoidance of social interaction, and increase their unsafe feelings when working (Stotzer, 2009). In line with the above, the life circumstances of these transgender sex workers can create a cycle where avoidance is self-perpetuating.

More than one-quarter of the MSWs reported a history of running away from home during adolescence. In fact, the separation from the family can contribute to feelings of loneliness, low self-esteem, worthlessness, and self-blame later in life, especially if the runaways were suffering psychological, physical or sexual abuse by their parents (Liu, Li, & Ge, 2009; Stoltz et al., 2007). Associated with a transgender identity not widely accepted by diverse segments of our society, including one’s own family, a “fertile ground” is settled for the development of serious psychological problems.

Both groups misuse alcohol and drugs and this does not seem to set apart one from the other. Other studies have already given great attention to the link between drug consumption and risk sex behaviors and the need of adequate interventions (Estcourt et al., 2000; Lau, Cai, Tsui, Chen, & Cheng, 2009; Norris, Kitali, & Worby, 2009; Rietmeijer, Wolitski, Fishbein, Corby, & Cohn, 1998).

Following a previous Brazilian study, transgender sex workers commonly reported that their clients tend to be young executives from middle or higher social class and to pay for receptive anal sex (Grandi et al., 2000). Maybe, this can justify why transgender sex workers claim to earn more than their counterparts herein. Besides, during our interviews, we could perceive that many transgender sex workers aimed at saving some money to support their families. In contrast, male hustlers did not manifest this intention and they usually spend all their money on personal excesses.

Similarly to other studies, our male hustlers reported poor performance at school and family history of criminal involvement, even at higher rates than transgender sex workers. This contributes to difficulties in acquiring conventional professional abilities and knowledge and, consequently, getting regular jobs (Dorais, 2005; West, 1993). Combined with their fewer harm avoidance levels, all of these factors can lead to higher exposure to thrilling situations, although our data cannot affirm this beyond a reasonable doubt. At least, more male hustlers reported to search for other professional occupations besides prostitution.

In addition, transgender people in general have reported serious depressive symptoms, high rates of suicidal thoughts, and lifetime suicide attempts (Baltieri et al., 2009). Unfortunately, although these people manifest desire for mental health counseling or psychiatric services, they reveal difficulties in finding an appropriate treatment in different places around the world (Herbst et al., 2008).

Far from being a single group, MSWs are a heterogeneous population and prevention strategies have to face this fact, without attempts to perpetuate generalizations. In our study, both groups were equally exposed to inconsistent condom use. In fact, it has become clear that the situations that lead to the sexual transmission of HIV should not be seen only from the narrow perspective of “risk behavior.” Although unprotected sex is the main cause of HIV transmission, causal models for justifying these behaviors cannot be totally satisfied with the notion of “lack of knowledge,” “inaccurate susceptibility estimations” or “lack of responsibility” (Caceres, 2002). The inability to negotiate safe sex or even non-penetrative sex can be extremely difficult in a number of different situations for many people, depending on their culture, values, perspectives, needs, desires, impulsiveness, and personality.

Street prostitution may appear attractive to vulnerable individuals lacking in education, training, self-discipline, and social or family support. However, the longer this activity is pursued the more difficult it becomes, since most customers demand striking good looks and youthful appearance. Thus, although the transition to conventional employment is not a simple task, especially for men who have never had proper work experience, it can be presumably easier for some types of MSWs. For the cross-dressers, besides the social prejudice, the temperament aspects may contribute to greater difficulties in making this transition or even in combining the sex work with a conventional one. Also, we put forward that some sex workers would do other things if they were given proper access to jobs and education, but others are pursuing a chosen activity in a reasonably contented and effective manner.

Specialized health programs for this population should include rigorous evaluation of characteristics and necessities of each male sex worker. We recommend that counselors working with this population strike a balance between facilitating self-disclosure and establishing more directive interventions. Structural programs for MSWs should address education, training, and skills that enable them to pursue viable, alternate income generating options in addition to social campaigns that direct societal attitudes and discrimination against sex workers. In the cases where certain traits of personality are found or even suspected, the staff must make efforts to promote the adherence to a specific program. A useful technique is to emphasize the fact that sex workers can have an “extra-endowment,” an above-average amount of feelings, and this can hinder some aspects of their lives (Hutton, 2004). In fact, to promote the therapeutic adherence of patients with certain personality traits is not an easy task; despite this, to enable our patients to see how their behavior can be negatively powerful may create a desire to overcome the problem. Also, some data suggest that people are more prepared to succeed in behavioral change when they have social support provided by capable social networks (Herbst et al., 2008). Thus, the identification of the types of social supports that are most likely to produce an environment fostering healthy and worthwhile behaviors is highly desirable.