Childhood sexual experiences and adult health sequelae among gay and bisexual men: defining childhood sexual abuse

Sonya Arreola [*], Torsten Neilands, Lance Pollack, Jay Paul, Joseph Catania [**]

[*] HIV Research Section, San Francisco Department of Public Health [**] Center for AIDS Prevention Studies, University of California, San Francisco

Journal of Sex Research, July-Sept, 2008 

[Introduction]  
Methods   
- Sampling  
- Measures    
  - Childhood sexual experience    
  - Mental health - Depression    
  - Well-being     
  - Substance use - Heavy drinking    
  - Poly drug use    
  - High frequency of drug use    
  - HIV transmission risk - High-risk sex    
  - HIV positive status  
- Statistical analysis  
Results  
- Overall  
- Tables 1, 2 & 3  
- Mental Health  
- Substance Use  
- Transmission Risk  
Discussion 

[Introduction] 

In the United States, gay and bisexual men continue to carry the burden of new HIV infections (Centers for Disease Control and Prevention [CDC], 2003). Studies using opportunistic samples have shown that childhood sexual abuse significantly predicts negative health outcomes such as risk for HIV and other sexually transmitted infections, as well as mental health outcomes, including depression, suicidal ideation, and substance abuse among adults 

(Briere, Evans, Runtz, & Wall, 1988; Briere & Zaidi, 1989; Browne & Finkelhor, 1986; Dhaliwal, Gauzas, Antonowicz, & Ross, 1996; Holmes, 1997; Holmes & Slap, 1998; Koenig, Doll, O'Leary, & Peguegnat, 2003; Lenderking et al., 1997; Miller, 1999; Molnar, Berkman, & Buka, 2001; Remafedi, Farrow, & Deisher, 1991). 

Although most childhood sexual abuse research has focused on women, the research on childhood sexual abuse among men shows that the prevalence of childhood sexual abuse is higher among gay and bisexual men than heterosexual men

(Jinich & Slap, 1998; Laumann, Gagnon, Michaels, & Michael, 1993; Paul, Catania, Pollack, & Stall, 2001), 

approximating that of women according to some studies 

(Doll et al., 1992; Finkelhor & Dziuba-Leatherman, 1994; Jinich et al., 1998). 

The recognition of high rates of childhood sexual abuse among gay men and indications that experiences of childhood sexual abuse may vary by gender has led to a growing interest in how childhood sexual abuse affects health outcomes of gay men. Results of research on the consequences of childhood sexual abuse among women often are generalized to men, despite conflicting findings regarding gender differences 

(Dhaliwal et al., 1996; Rind, Tromovitch, & Bauserman, 1998; Stanley, Bartholomew, & Oram, 2004). 

A study of differences based on a symptom checklist among men and women who had experienced childhood sexual abuse found no statistically significant differences between men and women in psychological sypmtomatology, whereas both genders compared with non-abused individuals had significantly more problems with depression and anxiety 

(Briere et al., 1988). 

Some studies that do find differences, however, report that men are more likely to externalize behavior aggression, but women are more likely to internalize behavior depression 

(Briere, 1988; Finkelhor, Hotaling, Lewis, & Smith, 1990). 

Further, a meta-analysis of studies of college students revealed that among those who reported childhood sexual abuse, men reacted much less negatively than women 

(Rind et al., 1998).

The variations in findings between women and men suggest that childhood sexual abuse sequelae may be more homogeneous for women than for men. Whereas women experience overwhelmingly experience negative psychological outcomes resulting from childhood sexual abuse, men show greater variability in their responses to childhood sexual abuse. Discrepant findings may result from the use of inconsistent criteria for what constitutes childhood sexual abuse. An extensive review of the literature on childhood sexual abuse among boys found that prevalence estimates ranged from 4% to 76% depending on the definition used and the population studied 

(Holmes & Slap, 1998). 

The definition of childhood sexual abuse often is based solely on an age discrepancy of 5 or more years between the child (usually under age 18) and the perpetrator, assuming that the power differential between the child and older person inherently constitutes abuse 

(Lenderking et al., 1997; Rind et al., 1998). 

Findings from a convenience sample of gay and bisexual men, however, found that self-esteem and sexual identity development did not differ between boys who reported consensual sex between the ages of 12 and 17 years old with someone older and those who reported no sex during that time 

(Rind, 2001). 

Thus, conceptualizing childhood sexual abuse as sex in childhood with someone older, and not distinguishing consensual from forced sex, may be undermining the ability to differentiate the variety of outcomes seen in adulthood that results from a range of childhood sexual experiences. 

For example, a study from Canada found that the prevalence of childhood sexual abuse was 26% when only the age differential was used but dropped to 12.5% when both the experience of coercion and an age differential of 5 years were required to define childhood sexual abuse 

(Stanley, Bartholomew, & Oram, 2004). 

By differentiating coercive from consensual sex, the authors also found outcome differences between the two subgroups. Those who reported consensual sex with older partners described their experiences as neutral or positive, whereas those who were coerced had greater adjustment problems including difficulties with competitiveness, coldness, expressiveness, and general interpersonal problems.

The wide range of outcomes among men and the high rates of childhood sexual abuse among gay men warrant further examination of childhood sexual abuse specifically among gay and bisexual men.

While the studies cited are not specific to HIV-related health outcomes, the findings suggest that the link between childhood sexual abuse and health outcomes, including HIV, may vary by kind of childhood sexual experience and call for a more refined analysis of childhood sexual experiences among gay and bisexual men and the potential effects of these early experiences. Specifically, we must begin to distinguish forced or coercive childhood sexual experiences from consensual ones and compare the adult health outcomes of these two groups with each other as well as with those of adults who have no childhood sexual experiences.

The current study looked at the differential effects in adulthood of forced, consensual, and no childhood sexual experiences among a probability sample of adult MSM in four large urban centers in the United States 

(New York, Los Angeles, San Francisco, and Chicago; Paul et al., 2001). 

Consistent with the developmental outcomes studies sited above, we hypothesized that the forced group would have more psychological distress, greater substance use problems, and higher risk for STIs including HIV than those who reported consensual or no childhood sexual experiences.

Methods 

Sampling 

The sampling frame construction methods, demographics, and distribution of HIV infection within the sample have been described in detail elsewhere 

(Catania et al., 2001; Stall et al., 2001). 

In summary, geographical areas (operationalized as zip codes) of four large U.S. cities (Los Angeles, San Francisco, Chicago, New York) were identified as being relatively rich in gay and bisexual residents. The identified areas cover more districts and a much higher proportion of the estimated population of MSM in each city than prior studies 

(Binson et al., 1996; Mills et al., 2001). 

Random-digit-dial methods were used to sample households in these areas, with interviews being conducted by telephone from November 15, 1996, through March 1, 1998. Men aged 18 years or older who had had sex with a man since age 14 or who defined themselves as gay or bisexual were eligible for interviewing. If a household contained more than one man who met these inclusion criteria, one of them was selected randomly to be interviewed.

The density of eligible households by zip code ranged from a high of 31% to as little as 1%. Two thousand eight hundred eighty-one interviews (78% of eligible households) were completed, lasting an average of 75 minutes. HIV serostatus was confirmed for a subsample of respondents using OraSure (Epitope: Portland, Oregon) oral HIV testing

(Osmond et al., 2000).

Measures 

Unless otherwise specified, all measures were based on self-reports by respondents.

Childhood sexual experience 

Childhood sexual experience was composed of three categories: 

Mental health - Depression 

Depressive features were measured using the Center for Epidemiological Studies-Depression (CES-D) scale. Although a cut-point of 16 has been the conventional cut-off for depression, for greatest specificity and sensitivity to depression, as compared with distress, a score greater than 22 was used to indicate depression 

(Fava, Pilowsky, Pierfederici, Bernardi, & Pathak, 1982; Herman, Susser, & Struening, 1994; Lyness et al., 1997; Radloff, 1977; Stall et al., 2003; Turk & Okifuji, 1994).

Suicidal ideation. This was measured dichotomously as ever attempted suicide versus never 

(Paul et al., 2002).

Well-being 

Well-being was measured using an abridged version of Ryff's instrument 

(Ryff & Keyes, 1995). 

A single-scale score was calculated based on 16 items from all six of Ryff's original subscales. For this article, the scores from the bottom quartile, or the lowest well-being scores, were compared with the remaining top three quarters.

Substance use - Heavy drinking 

Respondents who reported drinking at least weekly and had on average five drinks each time was designated a "heavy drinker," approximating the National Institute of Drug Abuse (NIDA) definition.

Poly drug use 

Poly drug use was measured as three or more types of recreational drugs from a list composed of marijuana, cocaine, crack cocaine, heroin, hallucinogens, inhalants, amphetamines, methamphetamines, methylenedioxymethamphetamine (MDMA, or "ecstasy"), barbiturates or tranquilizers, and painkillers in the past 6 months

(Stall et al., 2001).

High frequency of drug use

Respondents reported the number of times in the last 6 months they used each of the 11 types of drugs listed above. To avoid the inflating effect of simultaneous multiple drug use, frequency was defined as the maximum of the 11 responses. High frequency of drug use was operationalized as drug use more than one time per week (25 + times in 6 months), versus one or fewer times per week.

HIV transmission risk - High-risk sex  

A respondent was coded as having had high-risk sex if he reported engaging in anal intercourse (either receptive or insertive) without a condom ending in either withdrawal or ejaculation with a nonprimary male partner in the past year.

HIV positive status

Respondents were determined to be HIV positive if they reported being HIV positive or tested positive on the oral HIV test kit. HIV negative respondents were those who tested negative on the oral HIV test kit or who reported being HIV negative but were not part of the oral HIV test kit subsample 

(Osmond et al., 2000).

Statistical Analysis

The primary goal of this analysis was to describe and compare the UMHS participants across three childhood sexual experience groups (none, consensual, and forced) on various sentinel mental health and behavioral risk outcomes. 

To describe the population we computed proportions and 95% confidence intervals of proportions for 

stratified by childhood sexual experience group status. 

We also computed proportions for three demographic variables used in subsequent logistic regression analyses:

Comparisons of the three groups' odds of a particular mental health outcome (e.g., depression) or behavior (e.g., heavy drinking) were conducted via logistic regression in which each binary outcome was regressed onto indicator variables representing 

These logistic regressions were repeated with consensual sex as the referent for childhood sexual experience to permit computation of odds ratios for the forced sex vs. consensual sex comparisons. 

All regression models achieved adequate fit, which was operationalized as p > .20 on the Hosmer-Lemeshow goodness-of-fit test.

Due to the complex nature of the UMHS sampling methodology, all analyses were conducted using Stata Version 8's survey estimation commands (i.e., -svyprop-and -svylogistic-) that properly incorporate case weights, stratification, and clustering based on PSUs.

Results 

Overall, 

Table 1 

displays the distributions of the control variables (race/ethnicity, parent education, and age) among adult MSM reporting no sex, consensual sex only, or any forced sex before age 18. 

Table 2 

shows prevalence of the sentinel mental health, substance use, and transmission risk outcomes stratified by childhood sexual experience. 

Table 3 

reports the adjusted odds ratios comparing consensual and forced sex with no sex before age 18, and consensual to forced sex before age 18.

Mental Health 

Those who had forced sex were significantly more likely to be depressed or have suicidal ideation than those who had consensual sex and those who had no sex before age 18. 

There was no difference between the consensual sex group and those who had no sex before age 18. The level of well-being was significantly higher for the consensual group compared with the 'no sex before 18' group and the forced sex group. The latter two groups did not differ from each other on well-being.

Substance Use 

Both the forced and consensual groups had higher odds of heavy drinking than the no sex group and were not significantly different from each other. 

Prevalence of drug use was significantly higher in the consensual sex group compared with the no sex group and was also significantly higher in the forced sex group compared with the consensual sex group. This same pattern held for frequency of drug use, although with a larger effect than for prevalence of drug use.

Transmission Risk 

Prevalence of high-risk sex was significantly higher for the forced sex group than the consensual sex group, and it was significantly higher for the consensual sex group than the no sex group. 

Prevalence of HIV positive sero status was significantly higher for the forced sex group and the consensual sex group as compared with the no sex group, but those two groups did not differ significantly from each other.

Discussion 

Generally speaking, the forced sex group had the highest levels of psychological distress, substance use, and HIV transmission risk.

Isolating forced from consensual sex revealed that the consensual sex group is no different from the no sex group in rates of depression and suicidal ideation, highlighting the importance of differentiating the two types of childhood sexual experiences from each other.

The findings on substance use and transmission risk outcomes were different. On all five outcome measures both the consensual sex group and the forced sex group had higher rates than the no sex group. For frequent drug use and high risk sex, however, the rates are significantly higher for the forced sex group than they are for the consensual sex group.

The present findings suggest that, for gay and bisexual men, distinguishing forced from consensual childhood sexual experiences is critical to understanding mental health and transmission risk outcomes, and may provide a deeper understanding of the conditions under which substance use places men at risk. 

The operationalization of childhood sexual abuse as sex before age 18 with someone 5 or more years older may be functional for researching childhood sexual abuse among women. It may be that for women most childhood sexual experiences with older men are coercive, making it less necessary to differentiate forced from consensual sex. For men, however, this definition may be too inclusive.

The results obtained in the present study are consistent with the literature's findings; i.e., forced sex was significantly riskier than no sex for mental health and transmission risk outcomes. The finding that forced sex was significantly riskier than consensual sex, however, indicates that we must pay special attention to the unique effects of forced childhood sexual experiences on mental health and transmission risk outcomes among gay and bisexual men. 

For these men, it may be that two different trajectories emerge from childhood sexual experiences, 

The substance use findings underscore the need for further research.

Although both types of childhood sexual experiences, forced and consensual, were associated with higher rates of substance use and heavy drinking, frequent drug use was highest among the forced group. It may be that for the forced group, substance use plays a role in self-medicating psychological consequences of the traumatic nature of the coercion. By contrast, for those who experienced consensual childhood sexual experiences, substance use may become part of a more general risk-taking trajectory.

Interestingly, the forced sex group and the no sex group were statistically indistinguishable in their level of well-being, while the consensual sex group was significantly more likely to have a higher level of well-being than either of the other two groups. 

This suggests that consensual sex before 18 years of age may have a positive effect, perhaps as an adaptive milestone of adolescent sexual development. The emphasis in these data on pathology does not permit further exploration of this possibility.

While prior research has found good reliability and validity for retrospective reports of childhood sexual abuse 

(Whitmire, Harlow, Quina, & Morokoff, 1999) 

and sexual risk behavior 

(Coates et al., 1988; Saltzman, Stoddard, McCusker, Moon, & Mayer, 1987),

interpretation of any temporal or causal relationships are limited by the retrospective nature of the data. 

Further, there is always the risk of mis-categorizing childhood sexual experiences. For example, those who may have been coerced into sexual activity in childhood, but who do not see themselves as having been forced, may have been mis-categorized into the consensual group. This may have resulted in under-reporting of forced sex, and may have inflated the risk outcomes for the consensual group.

Findings from a qualitative study of Latino sexual development provide an example of a Latino gay man who described having had sex repeatedly with someone older beginning at age 6 over 3 years, but did not define his experience as abusive even though he clearly was upset by the memory of the events 

(Arreola, 2006). 

These histories must be taken seriously, and it is in this kind of situation that the complexity of human sexuality is most challenging. Labeling all childhood sexual experiences as childhood sexual abuse independent of the nature and subjectivity of the experience, however, renders useless our ability to more precisely predict the particular adjustment problems unique to each group or with whom health services should intervene and how.

The implications of these findings recommend that research move toward a more mature understanding of childhood sexual experiences that includes the subjective experience of childhood sex. 

In particular, it appears that at least for gay/bisexual men, the subjective experience of being coerced is particularly salient in predicting and differentiating mental health and transmission risk outcomes.

Future research should assess whether those who report consensual sex in childhood are older at the time of the childhood sexual experience than those who report forced sex. This may account for the differences in outcomes between the two groups. It is possible that the earlier the childhood sexual experience, the more likely it is to have been coercive. 

Data from studies using convenience samples suggest forced sex occurs at earlier ages compared with consensual sex 

(Rind, 2001; Stanley, Bartholomew, & Oram, 2004; Steever, Follette, & Naugle, 2001).

The present population-based study was unable to examine this comparison due to the absence of questions asking the respondent to specifically characterize the circumstances of his initial sexual encounter. 

It also is essential to conduct more nuanced evaluations of childhood sexual experience characteristics themselves in order to better understand the nature of these experiences. For example, severity and duration of childhood sexual experiences will help to differentiate those likely to need more intensive interventions, such as individual psychotherapy, from those who may not. 

Finally, it is important to note that both the forced and the consensual groups appeared to be at increased risk compared with the no sex group.

The findings imply, however, the following: 

These findings highlight the importance of clarifying the risk patterns and mechanisms that maintain them for each group, so that we may tailor our interventions to the specific trajectories related to each experience.