 |
|
Sexual Abuse of Boys
Definition, Prevalence, Correlates, Sequelae, and Management
William C. Holmes, MD, MSCE;
Gail B. Slap, MD, MS
JAMA. 1998;280:1855-1862.
ABSTRACT
| |
Objective.— To clarify the definition of sexual abuse of boys, update estimates of its prevalence, and explore critically its reported correlates, sequelae, and management.
Data Sources.— Studies from 1985 to 1997 were identified using OVID-MEDLINE and OVID-CINAHL. The search terms used were sexual abuse, sexual victimization, and sexual assault. Constraints applied included English language, human male subjects, age younger than 19 years, and North American samples.
Study Selection.— Publications were included for review if they appeared in peer-reviewed journals; had clear research designs; reported results for at least 20 male subjects; and were not reviews, perspectives, theoretical treatises, editorials, or letters.
Data Extraction.— Study types and sampling methods were categorized using well-established definitions. Preference was given to studies with large samples, with case-control or cohort designs, and/or with adjustment for effect modifiers or confounders.
Data Synthesis.— We identified 166 studies representing 149 sexual abuse samples. Studies were methodologically limited and definitions of sexual abuse varied widely. Prevalence estimates varied widely (by definition used and population studied), ranging from 4% to 76%. Boys at highest risk were younger than 13 years, nonwhite, of low socioeconomic status, and not living with their fathers. Perpetrators tended to be known but unrelated males. Abuse frequently occurred outside the home, involved penetration, and occurred more than once. Sequelae included psychological distress, substance abuse, and sexually related problems. Evaluation of management strategies was limited.
Conclusions.— Sexual abuse of boys appears to be common, underreported, underrecognized, and undertreated. Future study requires clearer definitions of abuse, improved sampling, more rigorous data collection, more sophisticated data analyses, and better assessment of management and treatment strategies. Regardless, health care professionals should be more aware of and sensitive to the possibility of sexual abuse in their male patients.
INTRODUCTION
SEXUAL ABUSE of young and adolescent girls has been well studied, with a consequent keen awareness of the existing and potential abuse of girls. In contrast, a relative silence about the abuse of young and adolescent boys exists. This has fostered a belief, among both health professionals and society at large, that the problem is uncommon and the outcomes are not severe.1-2 Recent events suggest otherwise. In the past year alone, there have been several widely publicized stories of boys whose alleged homicidal acts were propelled, at least in part, by their own sexual abuse histories.3-4 Greater attention to male sexual abuse and its potential outcomes appears warranted.
In a 1984 review, Finkelhor5 estimated that 3% to 4.8% of males in the United States had a history of prepubertal sexual contact with an adult male. Perpetrators tended to be unrelated acquaintances or strangers, and victims tended to have families of origin that were poor, physically abusive, and only 1 parent. When public authorities were contacted about the abuse, reports were made to the police rather than child protective services. Boys were less likely than girls to report sexual abuse because of the fear of retribution, the social stigma against homosexual behavior, the desire to appear self-reliant, and the concern about loss of independence following disclosure. Finkelhor5 postulated that male underrepresentation in commonly studied databanks from child protection agencies reflected both low reporting overall and preferential reporting to less commonly studied police records.
Studies of male sexual abuse have increased since 1984. Their definitions of abuse, methods, and findings vary widely. The objectives of this review are to clarify the definition of male sexual abuse; to estimate its prevalence; and to explore critically the reported victim, perpetrator and event characteristics, sequelae, and management.
LITERATURE SEARCH
Studies on male sexual abuse cited in OVID-MEDLINE and OVID-CINAHL, from 1985 to 1997, were identified using the search terms sexual abuse, sexual victimization, and sexual assault. Search results were limited to English language, human male subjects, and age younger than 19 years. Bibliographies were cross-referenced. Publications were included for review if they appeared in peer-reviewed journals; had clear research designs; reported results for at least 20 male subjects; were not reviews, perspectives, dissertation abstracts, editorials, and letters; and were conducted inside North America. Studies of mixed sex samples that did not report male subsample results were excluded. The sample for review consisted of 166 studies.6-40 41-80 81-100 101-150 151-171 A total of 149 sexual abuse samples were represented in these reviewed studies (a number of publications used the same sample but reported unique findings). Some studies included for review had sampled men. All of these studies, however, reported adult recollections and outcomes of childhood and/or adolescent sexual abuse.
Study designs and sampling methods were categorized using well-established definitions.173 Studies with mixed-sex samples were categorized according to the methods applied to the male subsample. Study characteristics of samples used more than once were counted only once.
LITERATURE QUALITY
Ninety-four studies (63%) were cross-sectional (or repeated cross-sectional), 38 (26%) were case series, 12 (8%) were case-control, 2 (1%) were cohort, 2 (1%) were longitudinal, and 1 (1%) was meta-analytical. Sampling techniques were nonprobability in 103 studies (69%), probability in 27 (18%), and unspecified in 19 (13%).
The median sample size was 159 male subjects. Twenty-two studies (15%) reported results from large male samples (>1000 subjects each), 3 of which included 16,000 to 60,000 subjects. Eight percent of the study samples had 500 to 999 male subjects, 38% had 100 to 499 subjects, 21% had 50 to 99 subjects, and 18% had 20 to 49 subjects. No investigators presented sample-size calculations or power analyses.
Descriptive statistics (eg, proportions, rates) about nonrepeated samples were reported in 137 studies (92%). Measures of association (eg,  2 tests, correlation coefficients) were used in 57 studies (38%). Univariate inferential statistics (eg, t tests, analysis of variance, and nonparametric equivalents) were used in 21 studies (14%), multivariate inferential statistics (eg, multivariate regression, discriminant function, and factor analyses) in 12 (8%), and both types of statistics in 10 (7%). Given the methodological variation and limitations of the studies, a review of the literature by meta-analysis was inappropriate.174
DEFINITION OF SEXUAL ABUSE
The definition of sexual abuse used by investigators varied widely. To explore this variability, we developed the classification system shown in Table 1 and Table 2. Studies first were categorized according to whether and how subjects were asked about their sexual abuse histories (Table 1). Some investigators did not ask subjects at all (eg, they used child abuse registries that used others' explorations of and definitions for sexual abuse) or did not report how they asked subjects. Others used either subjective or objective methods of asking subjects about their abuse experiences. Some investigators then applied additional criteria before defining reported experiences as abusive, such as a minimum age differential between victim and perpetrator (Table 2). This age differential varied across studies, from simply child vs adult, to a fixed number of years, to a graded number of years that increased with victim age. Other investigators required the presence of real or perceived coercion, a negative reaction by the victim, physical contact, and/or penetration.
|
|
|
|
Table 1.—Whether and How Subjects Were Asked About Sexual Abuse Histories
|
|
|
|
|
|
|
Table 2.—Additional Criteria Used by Some Investigators to Define Male Sexual Abuse
|
|
|
More than 30 different combinations of history-elicitation methods and additional criteria requirements were used. Investigators did not question subjects directly or did not report the methods by which subjects were questioned in 72 studies (48%). Subjective questioning was used in 40 studies (27%) and objective questioning in 37 (25%). An age differential was required in 49 studies (33%) (of these, the differential was unspecified in 14 [29%], fixed in 30 [61%], and graded in 5 [10%]). The additional criterion of physical contact was required in 17% of studies, coercion in 11%, authority in 5%, reaction in 3%, and penetration in 2%.
PREVALENCE
In a national telephone survey of men aged 18 years or older in the United States, Finkelhor et al48 found that 16% reported a history of sexual abuse in response to objective questioning. A population-based study from Ontario, also using objective questioning, found 7% of men reported a sexual abuse history.97 The other large-sample prevalence studies (>1000 subjects), most of which questioned younger subjects in schools, reported rates of 4% to 16%.7, 12, 64, 68, 70, 90, 95, 110, 112, 128, 148 Nelson et al112 reported a rate of abuse in the previous week of 2% among 9th- to 12th-grade boys in Oregon. In other large-sample studies using less generalizable samples, 5% of 1296 homeless men, 34% of 1001 men who had sex with men attending a sexually transmitted disease clinic, and 39% of 1574 sexually offending juvenile males reported histories of sexual abuse.8, 37, 89, 141 The only large-scale, nationally representative study of the incidence of sexual abuse in young and adolescent boys was completed for 1986. It found an estimated incidence rate of 1 (SE, 0.31) per 1000 young and adolescent boys.23
In a study of male college students, Fromuth and Burkhart53 found that the prevalence estimates of childhood sexual abuse were directly related to the definition of abuse. Prevalence was 22% when either a graded age differential or coercion was required, 14% when only the former was required, and 10% when both were required. Prevalence fell further to 8% when physical contact was required. Other studies reported that prevalence fell by 24% to 73% when the physical contact criterion was applied.22, 36, 67, 93, 163
Sample characteristics also affected the reported rates of sexual abuse. McCormack et al102 reported rates of 76% for serial rapists.22 Rates of 41% to 43% were reported in other samples of sex offenders.91, 140, 167 Lower rates (22%-33%) were found in non–sex-offending juvenile delinquents and detainees.29-30,140 Rates of 17% to 39% were reported in samples of men who had sex with men.24, 72, 93, 123 Surveys of runaway adolescent boys revealed rates of 21% to 38%, up to 4 times higher than rates found in school-based samples.47, 65, 81 The rates among male psychiatric inpatients and outpatients were similar at 24% to 40%.11, 20, 77, 79, 99, 108 A large sample of clients from substance abuse centers reported a rate of 29%.80 The rate of a Native American sample was 14%.130
Rate estimates were also affected by the methods of data collection. Chart review of psychiatric inpatients yielded a low prevalence rate of 6%, while face-to-face interviews in the same population yielded a rate of 26%.26, 79 Computerized and paper questionnaires completed by male Canadian university students yielded rates of 14% and 8%, respectively, with 90% of subjects reporting more honesty via computer questioning than either paper or face-to-face interview.6 The comparative benefit of telephone interviewing was not estimated.
Rohsenow et al133 studied programmatic effects on the estimated prevalence of sexual abuse in an inpatient chemical dependency rehabilitation site. The method of taking histories evolved from no systematic questioning about abuse to routine inquiry of all patients. Rates increased from 0% to 4% in the first period to 18% to 23% in the final phase. Other investigators have validated these findings.149
VICTIM CHARACTERISTICS
Large sample studies suggest several common characteristics of male sexual abuse victims. The mean and median ages of first sexual abuse were 9.8 years and 10 years, and 58% of boys were younger than 11 years.37, 48, 128 Boys who were nonwhite, lived only with their mothers, or lived with no parents were at increased risk for sexual abuse.37, 68, 95, 110 Siegel et al148 noted that non-Hispanics had a higher sexual abuse rate than Hispanics (6.5% vs 3.2%; P = .04). MacMillan et al97 noted no relationship between sexual abuse and parental education level or community size.
Studies with smaller or less generalizable samples were more varied in their reports of victim characteristics. Reported mean age of first abuse was 5.3 to 8.5 years in studies of children.27, 38, 46, 74, 76, 88, 117, 132, 151 In a study of adolescents, the mean age of reported abuse onset was older (10 years for those abused by males and 11.9 years for those abused by females).84 A study of adults recalling their abuse histories noted a similar age of onset of 9.8 years.128 Other studies indicated that while the range of ages at which abuse began was broad—from infancy to adulthood—most abuse began before puberty.33, 37, 46, 72, 100, 104, 106, 111, 127, 144-145,151, 164 More than 63% of adult subjects from 7 studies reported onset between infancy and 13 years.53, 93, 118, 130, 137, 151, 159
Most small-sample studies indicated that nonwhite males were more likely to be abused than white males.107, 109, 121, 124 One study suggested that black males were less likely to be abused by females than white males.84
Family factors that increased a boy's risk for sexual abuse included living with only 1 or neither parent; parental divorce, separation, or remarriage; parental alcohol abuse; and parental criminal behavior.8, 81, 91, 102, 106, 110, 124, 135, 146, 151, 165 Sexually abused boys were 15 times more likely than nonabused boys to have family members who also had been sexually abused (P = .001).63
Many small-sample studies reported an association between socioeconomic status and male sexual abuse. Resnick and Blum124 found that sexual intercourse prior to age 10 years was associated with low socioeconomic status (P<.001). Violato and Genuis159 reported an association between male sexual abuse and paternal unemployment or unskilled labor (P<.05). Pierce and Pierce117 found that 52% of sexually abused young and adolescent boys had mothers receiving public assistance. Finally, Faller46 noted that boys abused by relatives were more than twice as likely to be of low socioeconomic status than those abused by nonrelatives (P = .003).
One study suggested that disabled boys may be sexually abused more frequently than nondisabled boys (P = .07).150 Another study reported that 7% of sexually abused males had delayed recall of the abusive experience.104
PERPETRATOR AND EVENT CHARACTERISTICS
Large-sample studies reported that 53% to 94% of perpetrators were men, with up to half of female perpetrators being adolescent-aged baby-sitters.37, 48, 128, 141, 148 Small-sample studies revealed a similar predominance of male perpetrators. One study noted that 98% of these male perpetrators self-identified as heterosexual.82 Studies of children and young adolescents reported that more than 90% of perpetrators were male.38, 46, 82, 121, 143, 151, 154 Studies of older adolescents and young adults reported lower rates of male perpetrator abuse (22% to 73%), and rates of female perpetrator abuse from 27% to 78%.53, 84, 106, 140, 159 Studies of adult samples reported intermediate male perpetrator rates of 63% to 90%.24, 72-73,77, 113-114,132 These findings may suggest that males revise their perceptions as they age such that abusive sexual experiences with females become defined, retrospectively, as normative rather than abusive.
Large-sample studies reported that 54% to 89% of perpetrators were extrafamilial, and that 21% to 40% of these perpetrators were not known to victims.37, 48, 68, 95, 128, 148 Small-sample studies also reported that more than half of perpetrators were extrafamilial, but noted that less than 6% were strangers.76, 93, 121, 132, 154 Boys younger than 6 years were at greatest risk for abuse by family and acquaintances; boys older than 12 years faced an increasing risk of extrafamilial abuse by strangers.37, 46, 60 Findings by Lenderking et al93 suggested that up to 97% of the abuse of adolescent boys (>13 years) is extrafamilial.
The literature varied widely on the reported duration of sexual abuse. Several studies reported that abuse was a 1-time occurrence in 46% to 93% of cases (3 of these were large-sample studies that reported 1-time rates of 46%-73%).7, 48, 53, 128, 144, 147 Many boys (17%-53%), however, reported chronic abuse.7, 93, 99, 128, 151 Durations of abuse ranged from less than 6 months to 18 to 48 months.48, 60, 73-74,88, 113
Male victims typically described 3 or more types of sexually abusive acts, including forced anal penetration of the victim or perpetrator, vaginal penetration of the perpetrator, orogenital contact of or by the perpetrator, manual-genital contact of or by the perpetrator, and exhibitionism.117 Anal penetration was reported by 37% to 70% of victims in 13 studies, but by less than a third in 9 other studies.27, 37, 41, 48, 53, 72-73,76-77,88, 93, 100, 102, 106, 121, 128-130,144-145,164 Anal penetrative abuse was more likely to be repeated than other types of sexual abuse.128 It was reported by less than 10% of subjects victimized prior to age 2 years compared with 71% victimized at ages 9 to 11 years.121 Rhynard et al126 reported that 5% of male high school students had been forced to have intercourse while on a date (date rape). This study did not report perpetrators' sex.
Most studies reported that orogenital contact occurred at rates (12%-55%) similar to penetration: 15% to 38% of victims were fellated, and 12% to 35% of victims were forced to perform fellatio or cunnilingus.27, 37, 41, 53, 72, 76, 88, 100, 106, 117, 121, 129, 145 Fondling (by and of the perpetrator) was the most frequently reported act (55%-91% of cases), and exhibitionism was the least frequently reported act (as low as 6% of cases).41, 48, 88, 93, 100, 106, 145 Rhynard et al126 reported that 13% of male high school students had been fondled without consent while on a date. Rates of fondling and exhibitionism probably were underestimated throughout the reviewed literature since subjects were often instructed to describe only the most violating or disturbing act that occurred during an abusive event.
Many studies reported that physical force occurred in 10% to 25% of abuse events, although 4 studies reported higher rates of 32% to 56%.14, 31, 37, 48, 53, 106, 128, 132, 144 Moisan et al106 reported that weapons were used in 10% of the abuse events of black and Hispanic adolescent boys. Threats of physical force or harm increased with victim age and male perpetration.129, 147 Female perpetrators used persuasion rather than actual or threatened force in 91% of cases, and 26% promised special favors to boys for participation.74, 84, 147 Up to one third of abused boys said curiosity led to their participation.128
Histories of concomitant physical and sexual abuse were common. Seven studies, including 1 large population-based study, noted that 36% to 68% of sexually abused males were also physically abused.64, 70, 97, 100, 104, 136, 164 One study reported a lower rate of 20%.19 Two studies, one of 911 US Army soldiers, reported that histories of physical abuse were more common in men with than without histories of sexual abuse (44%-68% vs 13%-47%; P<.001).104, 136 Males who were sexually abused by family members were at highest risk of concurrent physical abuse.102 Family members of boys who were sexually abused also were more likely to be physically abused.63, 143
SEQUELAE
Negative responses to sexual abuse were reported by only 15% to 39% of male victims and were associated with the use of force, a greater perpetrator-victim age difference, an older perpetrator, and a younger victim.37, 53 Several other studies reported that two thirds of subjects perceived the abuse experiences as negative (especially if male perpetrated or involving penetration or fondling), with 63% reporting disabling obsessive thoughts about the abuse and 54% to 68% recounting strongly adverse life effects.79, 83-84,128 Rew et al125 reported that young men with histories of contact sexual abuse scored in the severely distressed range using a valid and reliable measure of well-being.
Victims who did not experience negative reactions to their abuse experiences had either positive reactions or were equally split between positive and neutral reactions.53-54,114 Of those with positive reactions, 91% recalled the events as physically pleasurable.114 Other factors associated with positive responses included age older than 12 years, longer duration, and female perpetration (88% abused by an adult female viewed the experience as a positive one).37, 67, 114 A meta-analysis (N = 2451) reported that sexual abuse was not associated with poor subjective health in males.59
Studies of actual clinical outcomes (rather than perceptions), however, indicated that sexually abused males were at increased risk for negative clinical sequelae. These sequelae included increased rates of posttraumatic stress disorder, major depression, anxiety disorders, borderline personality disorder, antisocial personality disorder, paranoia, dissociation, somatization, bulimia, anger, aggressive behavior, poor self-image, poor school performance, running away from home, and legal trouble.7-8,13, 16, 20-21,52, 58, 63, 68, 70-71,73, 75, 81, 91, 106, 110, 112, 116, 123-124,130, 132, 136, 145, 164-166 The rates among sexually abused compared with nonabused males were 4-fold for major depression (P<.001); 3-fold for bulimia (P<.05); and at least 2-fold for antisocial personality disorder (P = .002), behavior problems (P = .03), low self-image (P = .04), runaway behavior (P<.001), and legal problems (P = .001).63, 91, 110, 112, 130, 165 Rates of posttraumatic stress disorder and major depression among victims were 25% to 30% and 65%, respectively.130, 145, 166 Paris et al116 noted that sexual abuse was a significant, independent risk factor for the development of borderline personality disorder.
The rate of attempted suicide was 1.5 to 14 times higher among sexually abused compared with nonabused males.7-8,14, 16, 19, 35, 63, 68, 91, 110, 112, 123-124,165 Remafedi et al123 found that sexual abuse did not predict attempted suicide in a sample of gay or bisexual males after controlling for the age of sexual identity identification, illicit drug use history, and presence of feminine gender role.
A strong association between sexual abuse and subsequent substance use also was reported. Harrison et al63 found that sexually abused boys compared with nonabused boys in an inpatient chemical addiction center were more likely to report use of alcohol before age 10 years, marijuana use before age 12 years, and current drug use. Nagy et al110 noted that sexually abused compared with nonabused high school boys were 2 times more likely to use alcohol currently and 5 times more likely to use drugs currently (P<.05). In a similar sample, Nelson et al112 reported that current use of alcohol, marijuana, and cocaine were 2, 4, and 10 times higher (P<.001). Sexually abused and sexually and physically abused sixth-grade boys reported rates of multisubstance abuse that were 12 and 44 times greater than nonabused boys (P<.001); 12th-grade boys within the same abuse groupings reported rates that were 3 and 10 times greater (P<.001).64 In another study, the rate of injection drug use was up to 2 times higher (P<.001).8, 171 The increase in injection drug use was reported to begin during adolescence.72 Many other studies, in both institutional and community settings, supported these findings.8, 28, 31-32,43, 68, 70, 75, 130-131,145, 149 Some investigators postulated that the increased use of psychoactive substances reflected an attempt by abused males to self-medicate. |
