The DSM Diagnostic Criteria for Sexual Sadism

by Dr. Richard B. Krueger

 View PDF of this publication includes tables and references.

Abstract
I reviewed the empirical literature for 1900–2008 on the paraphilia of Sexual Sadism for the Sexual and Gender Identity DisordersWorkgroup for the forthcomingfifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM). The results of this review were tabulated into a general summary of the criticisms relevant to the DSM diagnosis of Sexual Sadism, the assessment of Sexual Sadism utilizing the DSM in samples drawn from forensic populations, and the assessment of Sexual Sadism using the DSM in non-forensic populations. I conclude that the diagnosis of Sexual Sadism should be retained, that minimal modifications of the wording of this diagnosis are warranted, and that there is a need for the development of dimensional and structured diagnostic instruments.

Introduction

The paraphilic diagnoses have been criticized as not constituting mental illness or involving society’s use of mental health professionals to constrain deviant behavior (Green, 2002; Moser, 2001, 2002) with some moving beyond mere criticism to recommending frank removal of the paraphilias from the DSM (Moser & Kleinplatz, 2005). The diagnoses of Sexual Sadism and Sexual Masochism, in particular, have been cited as pathologizing, stigmatizing, and discriminating against individuals who engage in alternative sexual practices (Wright, 2006). Indeed, Sweden recently took the step of removing transvestism, fetishism, and sadomasochism from its official list of diseases and mental disorders (The Associated Press, 2008) to avoid such discrimination. Further, although the diagnosis of Sexual Sadism is widely used for forensic purposes, it is not reported in diagnostic codes for outpatient ambulatory care. Survey information from the U.S. National Ambulatory Medical Care Survey was obtained for outpatient visits for diagnoses involving the sexual and gender identity disorders (W. Narrow, personal communication, December 16, 2008). This survey reported on the occurrence of diagnoses for a total of visits to psychiatrists, visits to urologists, visits to general/family/internal medicine physicians, and to obstetricians/ gynecologists. Strikingly, no visits with the diagnoses of Sexual Sadism or Sexual Masochism were recorded. This may reflect concerns about stigmatizing individuals with the application of these diagnoses, as well as absence of presentation of individuals for treatment for these problems.

This article will review the changes in narrative and the critiques of the diagnostic entity of Sexual Sadism, examine existing studies that have used the DSM criteria for Sexual Sadism, and review in particular studies that have examined the reliability, validity, and discriminant validity of such criteria. Because most of the studies have been conducted on forensic populations (consisting of subjects who have been arrested or incarcerated for sexual crimes) who one might expect could differ substantially from non-forensic populations, studies done using the DSM on forensic populations will be examined separately from studies done on non-forensic populations. Finally, discussion and recommendations will be based on the use of this diagnosis for both populations.

Further, for ease of reference, several tables have been developed. Table 1 contains criticisms relevant to Sexual Sadism,Table 2 lists studies that have utilized DSM-criteria in exclusively forensic populations, and Table 3 contains studies that have been conducted on mixed (consisting of both forensic and non-forensic) populations. Finally, also included are Appendixes listing all of the previous DSM criteria sets for Sexual Sadism and commentary (Appendix 1), along with ICD-9 criteria (World Health Organization, 1989), ICD-10 criteria (World Health Organization, 1992), and ICD-10 research criteria (World Health Organization, 1993) for sadomasochism (Appendix 2).

Method

Consisted of a literature search by a librarian at the New York State Psychiatric Institute using the search terms of ‘‘sexual masochism,’’ ‘‘sexual sadism,’’ ‘‘sadomasochism,’’ domination,’’‘‘ bondage,’’‘‘BDSM,’’‘‘perversion,’’‘‘paraphilia,’’‘‘sexual homicide,’’ ‘‘sexual murder,’’ ‘‘lust murder,’’ and ‘‘sex killer’’ of PubMed from 1966 through December 15, 2008, and of PsychInfo from 1900 through December 15, 2008. Additionally, all of the prior DSM manuals were consulted as well as ICD-9 and ICD-10. Articles were culled and attention was focused on articles using the DSM to make diagnoses of Sexual Sadism or offering critiques of the diagnostic criteria for Sexual Sadism or the paraphilias. Discussion of this literature and the diagnostic criteria were engaged in with colleagues.

Results

Summary of Evolution of Diagnostic Criteria for Sexual Sadism in the DSM

Sexual Sadism has been incorporated into the DSM manuals since its inception (American Psychiatric Association, 1952). In DSM-I, this was part of the diagnosis of ‘‘Sexual Deviation,’’ which was reserved for ‘‘deviant sexuality…not symptomatic of more extensive syndromes,’’ and was referred to as ‘‘sexual sadism (including rape, sexual assault, mutilation)’’. Sadism was continued as a ‘‘sexual deviation’’ in DSM-II (AmericanPsychiatricAssociation,1968)and masochism was added as a separate diagnosis (see Appendix 1). DSM-III (American Psychiatric Association, 1980) added specific diagnostic criteria, allowing a diagnosis to be made with one of the following: (1) on a nonconsenting partner, the individual has repeatedly intentionally inflicted psychological or physical suffering in order to produce sexual excitement or (2) with a consenting partner, the repeatedly preferred or exclusive mode of achieving sexual excitement combines humiliation with simulated or mildly injurious bodily suffering, or (3) on a consenting partner, bodily injury that is extensive, permanent, or possibly mortally is inflicted in order to achieve sexual excitement (see Appendix 1).

DSM-III-R (American Psychiatric Association, 1987) modified this to require: A. Over a period of at least six months, recurrent intense sexual urges and sexually arousing fantasies involving acts (real, not simulated) in which the psychological or physical suffering (including humiliation) of the victim is sexually exciting to the person; B. The person has acted on these urges, or is markedly distressed by them (see Appendix 1).

DSM-IV (American Psychiatric Association, 1994) added ‘‘behaviors’’ to the Criterion A requirement of sexual urges and sexual arousing fantasies, and added the conjunctive ‘‘or’’ so that any of these entities (sexually arousing fantasies, sexual urges, or behaviors) was sufficient in Criterion A and changed Criterion B, removing the terminology that a person had ‘‘acted’’ on these, and replacing this with the criteria that these caused ‘‘clinically significant distress or impairment in social, occupation, or other important areas of functioning’’(see Appendix 1).

Finally, DSM-IV-TR (American Psychiatric Association, 2000) returned to the criteria that an individual had ‘‘acted’’ on these urges with a nonconsenting person, and continued with the criteria of ‘‘marked distress or interpersonal difficulty’’ (see Appendix 1).

This last change, returning to the criteria of DSM-III-R,was to avoid the unintended consequence of the removal of the requirement that an individual had acted on such urges in DSM-IV. This deletion would, in the case of an individual with pedophilia, for instance, have not allowed for a diagnosis of pedophilia to be made for an individual who had acted on such urges, but was not distressed by them or socially or occupationally impaired by them (First & Pincus, 2002; Hilliard & Spitzer, 2002). The editors of DSM-IV, regarding the changes in sexual sadism from DSM-IV to DSM-IV-TR, went on to say:

Because some cases of sexual sadism may not involve harm to a victim, such as inflicting humiliation on a consenting partner, the wording for sexual sadism involves a hybrid of the DSM-III-R and DSM-IV text. The DSM-IV-TR version states: ‘‘The person has acted on these urges with a nonconsenting person, or the urges, sexual fantasies, or behaviors cause marked distress or interpersonal difficulty.’’ (p. 291)

In a later communication, the editors of the DSM-IV-TR (First & Frances, 2008) indicated that the addition of the phrase ‘‘or behaviors’’ to Criterion A in DSM-IV had allowed forensic evaluators to conclude that an individual who had committed a sexual offense (e.g., rape) would qualify for the diagnosis of a mental disorder solely on the basis of repeated acts of sexual violence alone, without establishing the underlying condition of deviant urges or fantasies requisite to establishing that a mental illness existed and they recommended removing the phrase ‘‘or behaviors’’ from the DSM-IV criteria. They cautioned that ‘‘tinkering with criteria wording should be done only with great care and when the advantages clearly outweigh individuals who had nonstandard sexual interests. He proposed an alternative classification, Sexual Interest Disorder, to focus on sexual behavior that becomes a problem that would not identify specific sexual interests, such as sadism, as being pathological in and of themselves. This would have two criteria: A: Specific fantasies, sexual urges, or behaviors that cause clinically significant distress or impairment in social, occupational, or other important areas of functioning; B: The sexual interest is not better accounted for by another Axis I disorder, not due to the effects of a general medical disorder, and is not the result of substance use, misuse, or abuse.

Doren (2002) discussed many issues related to the diagnosis of paraphilias in forensic settings. He made the point that in the case of pedophilia one could define a numerical threshold (such as being caught more than 2 or 3 times) for this diagnosis because the penile plethysmographic (PPG) literature suggested that if a child molester had been caught on several occasions, there was a very strong likelihood (i.e., 80% or more) that he was a pedophile (Freund & Watson, 1991). On the other hand, attempts to develop the same sort of behavioral definition based on PPG literature had not shown consistent results for men who had assaulted adults. Some rapists showed clear sexual arousal to depictions of rape in PPG laboratories, and some did not, and this precluded using a numerical threshold for defining a rape-related paraphilia in the same way that one could for pedophilia.

Marshall and Kennedy (2003), in an extensive review of Sexual Sadism in sexual offenders, reported that while most of the authors in the studies they reviewed indicated that they used DSM or World Health Organization’s International Classification of Diseases (ICD) criteria to diagnoses their subjects, the criteria that they specified did not comply with either of these systems and each researcher chose an idiosyncratic list of criteria which included some features from both DSM and ICD but also included other features not mentioned in these documents. They rather pessimistically concluded:

In conclusion then, after more than 100 years of research and clinical observations we seem no closer to a satisfactory, agreed upon, and reliable diagnosis of sadism than was true when [von] Krafft-Ebing (1886)…first described a series of cases he called sadistic. Our review of the evidence does not encourage confidence that things will improve in the future, so we recommend abandoning the diagnosis. Instead, we suggest that researchers rely on behavioral data to identify their subjects along various dimensions of brutality. These dimensions should include the degree of aggression or force, the enactment of degrading or humiliating behaviors (acts as well as speech), and the magnitude of the victim’s injury.

Berner, Berger, and Hill (2003) reviewed Sexual Sadism and presented follow-up data on an earlier evaluated forensic sample. They suggested that more recently there had evolved a different distribution of Sexual Sadism versus Sexual Masochism, with masochism being predominant in outpatient psychiatric facilities and sadism prevailing in forensic settings, supporting the concept of separated diagnoses of sadism versus masochism.

Moser and Kleinplatz (2005) reviewed the paraphilic diagnoses in all of the DSMs, and argued that paraphilias did not meet the definition of a mental disorder and that the DSM presented ‘‘facts’’ to substantiate various assertions in the text, but they found little evidence to support these assertions. They opined that the paraphilias section was so flawed that it should be removed from the DSM. They suggested that an alternative would be to change the definition of a mental disorder or of paraphilia or both, correct factual statements, adjust criteria for inclusion of a diagnosis, and add safeguards to prevent the misuse of the diagnoses. They indicated that other psychological characteristics described individuals now diagnosed with a paraphilia who sought psychotherapy, and said that these concerns more accurately reflected their concerns than their sexual interests did. They stated:

It is not their sexual interests, but the manner in which they are manifest that can be problematic at times and is a more appropriate focus for therapy. The confusion of variant sexual interests with psychopathology has led to discrimination against all ‘‘paraphiliacs.’’ Individuals have lost jobs, custody of their children, security clearances, become victims of assault, etc., at least partially due to the association of their sexual behavior with psychopathology. (p. 107)

Spitzer (2005) responded to the above saying that the concept of ‘‘medical disorder’’ could be applied to human behavior, and doubted that anyone had been hurt by being given a diagnosis of a paraphilia. Fink (2005) maintained that it was important to retain diagnoses to differentiate between normal and abnormal ways in which people become aroused and that retaining paraphilic diagnoses was important ‘‘to save some people from jail and others from themselves’’ (p. 118).

Kleinplatz and Moser (2005) said that Drs. Spitzer and Fink earlier did not dispute their analysis of the problems with the DSM-IV-TR criteria for paraphilias and that conservative organizations had flagrantly misrepresented their statements and intent at a symposium they had presented it at. They stated that public opinion and not science were the main reasons the paraphilias had been kept in the DSM.

Reiersøl and Skeid (2006) focused their efforts and criticism on the ICD-10, concluding:

The ICD diagnoses of Fetishism, Transvestic fetishism and Sadomasochism are outdated and not up the scientific standards of the ICD manual. Their contents have not undergone any significant changes for the last hundred years. They are at best completely unnecessary. At worst, they are stigmatizing to minority groups in society. There are people who are suffering from stigma and emotional distress because of the diagnoses. (p. 260)

Marshall and Hucker (2006) summarized their research on Sexual Sadism, which included an initial study showing that experienced forensic psychiatrists did not accurately employ many of the important diagnostic criteria and a second demonstrating that ‘‘internationally-renowned’’ forensic psychiatrists could not reliably apply the diagnosis, and indicated that they were in the process of developing a Sexual Sadism Scale.

Kirsch and Becker (2007) reviewing information on psychopathy and Sexual Sadism, wrote:

Overall, the difficulties in defining and operationalizing sexual sadism, the unreliability of the diagnoses (Marshall, Kennedy, & Yates, 2002), and findings that normal males report occasional sadistic sexual fantasies (Cre´pault & Couture, 1980), have led some to argue for a dimensional approach to defining the disorder (Marshall & Kennedy, 2003). Given that little work has examined the appropriateness of this approach and the available research to date has used a categorical classification system, this paper will consider sexual sadists to be a discrete group, though the reader should be aware that the reliability of the diagnosis of sexual sadism is an issue that warrants greater empirical attention. (p. 908)

Finally, Fedoroff (2008) in a recent review raised several questions, without answering them, concerning the A criterion for Sexual Sadism in DSM-IV-TR: ‘‘Why 6 months? What does recurrent mean? What does intense mean? Is it meaningful to discuss sexual urges independent of sexual fantasies? Why distinguish between real and simulated acts? Appearing to be a fairly inclusive criteria, why is humiliation specifically identified in addition to psychological and physical suffering?’’ He concluded:

This review indicates that sexual sadism, as currently defined, is a heterogeneous phenomenon. To date, research has often failed to clearly define the population under study and therefore conclusions are limited. This makes generalization from research findings to specific patients problematic. Of particular concern is the possibility that correlations and outcomes from studies consisting of samples of convenience may be interpreted as verified causal relations between unconventional sexual interests and nonconsensual sexual violence… (p. 644)

To summarize the above, the DSM has been criticized for many years for its poor reliability, particularly in issues involving its use in forensic venues. Better interrater reliability has been achieved through structured instruments, education of raters, and appropriate selection of samples. The paraphilias have been criticized as not being mental disorders, and, through inclusion in the DSM enabling society to pathologize and discriminate against people who practice alternative sexual lifestyles. Those critics maintain that there is no evidence that these lifestyles are associated with any significant degree of psychopathology.

Some experts, reviewing Sexual Sadism, have concluded that the diagnostic reliability is so poor that the use of this diagnosis should be abandoned in favor of dimensional approaches to assessment, perhaps involving sexual arousal, or degree of violence, that could be of use in treating individuals. Others have concluded that the possibility of using a threshold number of sexual assaults, for instance, to diagnose Sexual Sadism, or another possible paraphilia of nonconsensual rape, is not supported by penile plethysmographic data supporting differential arousal of rapists to violent stimuli.

Further, some have criticized the facts presented in the narrative sections of the DSM concerning paraphilias, alleging they are inaccurate and provide misinformation. Finally, many questions could be raised about the wording of the criteria for Sexual Sadism that also apply to other paraphilias (e.g.,why is 6 months of duration required, what does ‘‘recurrent’’ or ‘‘intense’’ mean, and how are these operationalized? Should ‘‘preferential’’ be added to the criteria for Sexual Sadism as a threshold for making the diagnosis, or as a qualifier, for instance?).

Review of Diagnostic Studies Involving Use of the DSM in Forensic Populations

Virtually all of the published papers using DSM criteria for Sexual Sadism have been done on studies of forensic populations. Many of these studies have involved sexual homicides of one sort or another, despite the fact that these are exceedingly rare events. Chang and Heide (2009) reported, for instance, that in 2004 sexual homicide accounted for approximately 1.1% of 14,121 murders in the United States.

An early study Packard and Rosner (1985) reviewed records of 95 defendants charged with sexual offenses evaluated in a forensic psychiatric clinic between 1980 and 1883. DSM-III criteria were used and only 6.3% of individuals received a diagnosis of a paraphilia, without further qualification.

Langevin, Ben-Aron, Wright, Marchese, and Handy (1988) reported on a small study of 13 sex killers who were interviewed because they had murdered someone in conjunction with erotic arousal, and compared this with a sample of 13 nonsexual homicide perpetrators. Seventy-five percent of the group who had murdered someone in conjunction with erotic arousal had sexual sadism; 0% of the nonsexual homicide perpetrators received diagnosis of Sexual Sadism. Phallometric testing was offered in 17 cases; 9 of the subjects refused.

Dietz, Hazelwood, and Warren (1990) authored an oft-cited study of 30 sexually sadistic criminals; DSM-III-R criteria were not formally used, but for a case to be admitted into the study, all three of the study authors, on the basis of a retrospective chart review, had to agree that the subject had to have been sexually aroused in response to images of suffering or humiliation on two or more occasions spanning at least six months. Documented or self-reported sexual acts were used to infer arousal. Seventy-seven percent of the subjects engaged in sexual bondage and 100% in intentional torture of the victim.

Yarvis (1990) reported on 100 murderers he had examined between 1980 and 1988. It appeared that 3 of 10 subjects who committed a homicide/rape received a diagnosis of Sexual Sadism. None of the other subjects received this diagnosis.

Bradford, Boulet, and Pawlak (1992) reported on information obtained from 443 males who were consecutively admitted to the Sexual Behaviors Clinic at the Royal Ottawa Hospital, using 11 items from their Male Sexual History Questionnaire. Formal DSM criteria were not used and there was no mention of sadism or masochism. Thirty subjects admitted to rape and 56 to attempted rape. The authors suggested reviewing diagnostic criteria for paraphilias and that a class of ‘‘coercive paraphilia’’ be considered for the DSM.

Gratzer and Bradford (1995) compared offender and offense characteristics reported on in the 30 sexually sadistic criminals studied by Dietz et al. (1990) and compared these with 29 sexually sadistic criminals and 28 nonsadistic sexual offenders at the Royal Ottawa Hospital. Sexual sadists were more likely to engage in physical and psychological torture of the victim. Some of the offender and offense characteristics were not specific to sexual sadism.

Yarvis (1995) reported on a sample of 180 murderers that he had interviewed over a 13-year period using DSM-III criteria (used for consistency, even though DSM-III-R and DSM-IV were published during this period). Only individuals committing sex crimes received a diagnosis of Sexual Sadism, with 6.5% of rapists and 30% of sexual murderers receiving a diagnosis of Sexual Sadism.

Geberth and Turco (1997) reported on a study of 232 serial murderers who had violated their victims sexually (selected from a group of 387 serial murderers) identified from the media and the FBI’s National Center for the Analysis of Violent Crime. They used a case history protocol based upon the DSM-IV criteria of antisocial personality disorder and sexual sadism, and found that 68 cases met the criteria for antisocial personality disorder and Sexual Sadism. These diagnoses were not separated.

Firestone, Bradford, Greenberg, and Larose (1998) reviewed information collected on 48 homicidal sex offenders assessed between 1982 and 1992, and studied these in relation to a comparison group of incest offenders. History, psychological inventories, phallometric assessments, and DSM diagnoses were collected on each group. DSM-III diagnoses reliably discriminated between the groups, with 75% of homicide offenders and only 2% of incest offenders receiving diagnoses of Sexual Sadism. Forty percent of homicidal offenders and two percent of incest offenders received diagnoses of Pedophilia and Sexual Sadism. Psychiatrists made diagnoses before they had psychological test scores of results of phallometry.

Raymond, Coleman, Ohlerking, Christenson, and Miner (1999), using a structured clinical interview for the paraphilias, interviewed 45 males with pedophilia. They found, tabulating lifetime diagnoses, that two of this group had Sexual Sadism and none had Sexual Masochism.

Berger, Berner, Bolterauer, Gutierrez, and Berger (1999) reported on a study that involved the assessment of sadistic personality disorder, other personality disorders, and Sexual Sadism in 70 sex offenders (27 child molesters, 33 rapists, and 10 murderers). This was a prospective study with informed consent. At least two investigators for each case made DSMIII- R diagnoses on the basis of separate interviews, arriving at a consensus. The diagnosis of a paraphilia and the assessment of sexual fantasies were assisted by a separate informal interview with the patient’s therapist. All available sources of information, such as criminal records and court reports, were used. Forty-two percent of subjects had sexual sadism by the DSM-III- R criteria, 19% admitted to sadistic fantasies during masturbation and only 6% admitted that they carried out sadistic activites during intercourse or masturbation. In a follow-up study Berner et al. (2003) following 60 of 70 patients for an average of 6 years, reported there was a trend towards those with sexual sadism having a higher relapse rate.

Holt, Meloy, and Strack (1999) examined records from a nonrandom sample of 41 inmates at a maximum security prison, making a diagnosis of Sexual Sadism using threshold criteria from the DSM-IV and data from the subject’s prison file and a structured clinical interview. Only three individuals received a diagnosis of Sexual Sadism.

Stone (2001) reported on 98 men who had committed sexual homicide, whose biographies he had complied through publicly available information. He reported that 18 of these 98 were reported as having the paraphilia of ‘‘sexual sadism with orgasm.’’

Marshall et al. (2002) extracted archival data on 59 sexual offenders who had been diagnosed by experienced forensic psychiatrists in the Canadian prison system using DSM-III-R or DSM-IV criteria. Forty-one of the cases were diagnosed as sexual sadists and 18 had been given other diagnoses. Printouts of information from all 59 offenders were independently coded by two of the authors into 40 categories (consisting of 18 offense features, 10 self-report categories, 7 phallometric profiles, and 5 diagnoses). They found, comparing sadists with non-sadists, that far more nonsadists were deemed to have various personality disorders other than antisocial personality disorder; that sadists differed from non-sadists in only 2 of 18 categories of offense characteristics (beating and torture) with nonsadists displaying higher frequencies, and that there were no significant differences on self-reported fantasies or acts. Regarding phallometric data, nonsadists showed greater arousal to ‘‘nonsexual violence’’ and sadists showed greater arousal to ‘‘consenting adult’’ stimuli. Marshall et al. concluded that the frequency with which sexual offenders diagnosed as sadists displayed features identified in the literature as being associated with sadism was lower than previously observed and that the diagnosis of Sexual Sadism did not differentiate those deemed to be sexual sadists from those who were not. They suggested that either there were poor diagnostic practices in the Correctional Services of Canada or that the criteria for Sexual Sadism were insufficient.

Marshall, Kennedy, Yates, and Serran (2002) conducted a study of 24 psychiatrists deemed to be expert in forensic diagnosis. Each was sent 12 vignettes of men, half of whom had been diagnosed in their earlier study as being sexual sadists and half of whom had not received this diagnosis. However, only 15 psychiatrists completed and returned the questionnaire. The authors computed, using Cohen’s method for estimating interjudge agreement, a kappa of 0.14, well below acceptable levels. They also found that three features that there was agreement on regarding the diagnosis of Sexual Sadism were cruelty or torture, sexual mutilation, and deviant sexual arousal. They suggested that these features, unlike control and humiliation, were not a common feature of most sexual assaults and that these might constitute a subclass of very dangerous sexual offenders, and that the diagnosis of Sexual Sadism should be restricted to those who met these three criteria.

Langevin (2003) compared 33 sex killers with 80 sexual aggressives who had engaged in sexual activity and killed or attempted to kill their victims before, during, or after the sexual activity. These cases were extracted from a database of more than 2,800 cases; three comparison groups were selected, including a sample of 80 nonhomocidal sexually aggressive men and 23 nonhomocidal sadists. Each person had been interviewed and various tests were administered, including the Clarke Sex History Questionnaire for Males and the Freund Phallometric test of erotic preference in selected cases. Seventy percent of sex killers, 30% of sexual aggressives, and 4% of all sex offenders were identified as having ‘‘sadomasochism.’’

Becker, Stinson, Tromp, and Messer (2003) reported on a review of the legal files of 120 sexual offenders, the entire population up to the time of the study of men who were petitioned for civil commitment in Arizona. Of these offenders, 8.5% received diagnoses of Sexual Sadism and 2% Sexual Masochism.

Levenson (2004a) reported on a study that consisted of a review of diagnostic data drawn from a sample of 450 male convicted sex offenders inFlorida prisons who had received an independent in-person evaluation by at least two psychologists or psychiatrists for SVP civil commitment during the 2000 and 2001. The purpose of the study was to calculate the interrater reliability for, among other things, the DSM-IV diagnoses used to assess whether an offender had a mental abnormality. A total of 277 men were included and kappa was computed for eight DSM-IV diagnoses. Overall, kappa was found to be poor to fair (kappa=0.23–0.70) with the kappa for Sexual Sadism being only 0.30 (poor). Levenson concluded that because the DSM was the only diagnostic taxonomy recognized by U.S. courts, it was critical to improve diagnosis and that diagnosis was difficult because an evaluator must infer arousal to sadistic acts in cases where clients did not readily admit such arousal. In a separate article, the rate of Sexual Sadism was reported as being 4% (Levenson, 2004b).

Packard and Levenson (2006) reanalyzed their 2004 sample after concluding that there were significant limitations to using kappa in reliability studies. They used new statistical analyses measuring raw proportions of agreement, odds and risk ratios, and estimated conditional probabilities to examine reliability. The proportion of total agreement in diagnostic decisions for Sexual Sadism was 97%. They concluded that kappa could be misleading when used exclusively, and that overall the civil commitment evaluation was a highly reliable process.

Hill, Habermann, Berner, and Briken (2006) examined a group of court reports on 166 men who had committed a sexual homicide. Psychiatric court reports were evaluated by three raters. Twenty forensic psychiatrists had written the reports. Psychiatric disorders were diagnosed by the raters according to DSM-IV. A total of 61 (36.7%) men received a diagnosis of Sexual Sadism; no significant differences in sociodemographic characteristics or intelligence were found. About 14 percent of the sexually sadistic offenders were diagnosed with Sexual Masochism. A subsequent study by Hill, Habermann, Berner, and Briken (2007) reported on interrater reliability that was assessed evaluating 20 reports by all three raters. For all Axis I disorders, Cohen’s K ranged from 0.61 to 1.0 with a mean K=0.82, but Sexual Sadism was not specifically reported on. Another study by Hill, Habermann, Klusmann, Berner, and Briken (2008) for an estimated recidivism rate at 20 years at risk disclosed no relationship with Sexual Sadism.

Elwood, Doren, and Thornton (2008) reported on data retrieved from an archival database of 331 sexual offenders held under Wisconsin’s sexual offender statute. Diagnoses had been made by doctoral level licensed psychologists, using the DSM-IV criteria. A total of 8.5% had Sexual Sadism.

McLawsen, Jackson, Vannoy, Gagliardi, and Scalora (2008) sent an anonymous and confidential survey through the Association for the Treatment of Sexual Abusers (ATSA) and the American Psychology-Law Society (AP-LS) e-mail list to professionals who made diagnoses of Sexual Sadism. Sixty participants completed the survey. Participants had made an average of 2.54 diagnoses of Sexual Sadism. Sixty-two statements were included in the survey, drawn from four conceptualizations of Sexual Sadism, with items culled from an extensive literature review. Participants were asked to rate each statement on a 7-point Likert-type scale from ‘‘not at all essential’’ to ‘‘absolutely essential’’ for making a diagnosis of Sexual Sadism. The items were divided into two mutually exclusive categories: Sexual Sadism (39 items) and a general sexual offending category (23 items). Overall, ratings of the two categories differed significantly, indicating that participants were able to differentiate Sexual Sadism from general sexual offending. Behaviors that were common to three of the four conceptualizations were ‘‘slapped or punched victim during the sexual act; cut, stabbed, strangled, bit, or beat victim during sexual act; and, physical restraints used during sexual act’’ (p. 294).

Beauregard, Stone, Proulx, and Michaud (2008) reported on a small study in which 11 sexual murderers of children and 66 sexual murderers of adult women were interviewed. Although no diagnostic instruments or criteria were described, it was concluded that because sadism was a recurrent theme among sexual murderers that future studies should be undertaken to validate a diagnostic instrument of sadism.

So, to summarize the above, some 27 studies have utilized or referred to DSM criteria for the evaluation of subjects in forensic populations. Most studies were not prospective, i.e., they relied on data that had already been obtained by interviewers. Some relied not on direct interviews but on criminal records or information from the media. In those studies that relied on clinical information, almost none of the primary interviewers had utilized structured diagnostic instruments specifically geared towards making diagnoses of the paraphilias or, for that matter, of any of the psychiatric disorders. This is important in that it is conceivable, given the association of Sexual Sadism with Sexual Masochism, for instance, that one might find a substantial occurrence of Sexual Masochism in individuals with Sexual Sadism. Yet, the study design and data collection did not allow for this data to be generated and we do not, in fact, know, if questions pertaining to sexual masochism or the other paraphilias were even regularly included in interviews or assessments.

Few studies have examined interrater reliability. Those studies that have are not entirely comparable. Some have found good interrater reliability and some have found poor reliability. It is not apparent, however, that this poor interrater reliability is a consequence of ambiguous or poor criteria for Sexual Sadism. It could as well be that lack knowledge about diagnostic criteria, lack of training in those conducting the primary interviews, or failure to use structured instruments could account for poor interrater reliability.

Summary of Studies with any Mention of Sexual Sadism Utilizing the DSM in Samples Drawn from Clinical or Not Clearly Forensic Populations.

Abel et al. (1987) and Abel, Becker, Cunningham-Rather, Mittelman, and Rouleau (1988) reported on an outpatient population of 561 men seeking voluntary evaluation and treatment for possible paraphilias in Memphis, Tennessee or in New York City. In the Memphis sample, all categories of paraphilias were evaluated; in the New York sample, mostly subjects with a diagnosis of rape or child molestation were seen.DSM-II and DSM-III criteria were used, with all subjects reporting recurrent, repetitive urges to carry out deviant sexual behaviors. Subjects were not included in the research solely because they had committed the paraphilic behavior. One-third of this sample was referred from legal or forensic sources, one-third from mental health sources, and one-third from other sources. A total of 28 men were diagnosed with sadism, 17 with masochism, and 126 as rapists.

Kafka and Prentky (1994) collected data prospectively on 63 consecutively evaluated outpatient males. Three men were excluded. Thirty-four were seeking treatment for paraphilic disorders and 26 for paraphilia related disorders. A questionnaire was used along with a structured interview to establish a diagnosis, which represented a lifetime diagnosis. It was not clear which paraphilia was the focus for treatment. Twelve percent of the paraphilic group was diagnosed with Sexual Sadism and 9% with Sexual Masochism. Kafka and Prentky recommended that future studies should utilize structured diagnostic interviews and blind interviewing techniques

In the volume Dangerous Sex Offenders (American PsychiatricAssociation, 1999), there were some data in the form of a personal communication from Dr. Gene Abel on a sample of 2,129 patients evaluated at 140 sexual treatment clinics in North America, who presumably answered questions on the Abel Assessment of Sexual Interest, although this was not explicitly stated. Of this sample, 2.3% reported they had engaged in sadism and 2.5% in masochism, but the methods and questions used to obtain this information were not described.

Kafka and Hennen (2002, 2003) reported on a population of 120 consecutively evaluated outpatient males with paraphilias (N=88, including 60 sex offenders), and paraphilia-related disorders (N=32). Structured interviews and DSM-IV criteria were used to make lifetime diagnoses. Eleven percent of the paraphilic sample had Sexual Masochism and 5% Sexual Sadism. Kafka and Hennen noted that there were no rating instruments with documented reliability and validity to diagnose both paraphilias and paraphilia related disorders. The index paraphilia for which treatment was sought was not specified.

The above four studies are the only studies I have found which apply DSM criteria for Sexual Sadism to populations that are not exclusively forensic, and each of these studies has a substantial component of forensic cases. This implies that researchers are not using criteria from the DSM to conduct research on non-forensic community populations or populations seeking treatment, and/or that individuals with Sexual Sadism are not presenting in any substantial numbers in a nonforensic way for treatment.

Other Issues

Research on Sadomasochism in the Community

Moser and Levitt (1987) reported that general population surveys had not established the proportion that identified as S/M and noted that it was not clear if any specific behaviors could be classified as S/M specifically. Paraphilic disorders have, to date, not been included in any of the broad epidemiological surveys of mental disorders (Kessler et al., 2005). Yet S &M behavior would appear to be fairly common. Kinsey, Pomeroy, Martin, and Gebhard (1953) reported that 26% of females and 26% of males reported a definite and/or frequent erotic response to being bitten. In a survey of sexual behavior in the United States involving 2,026 respondents in 26 cities, Hunt (1974) found that 4.8% of males and 2.1% of females reported ever having obtained sexual pleasure from inflicting pain, and 2.5% of males and 4.6% of females from receiving pain. Females appear to have a significant presence among S & M practitioners. Breslow, Evans, and Langley (1985, 1995) reported on a study in which questionnaires were placed in two publications that catered to sadomasochists; of 182 individuals who responded, 130 were males and 52 females, indicating a significant female presence in the subculture. Finally, studies from the S & M population could have much to contribute to an understanding of sexual sadism. For instance, Cross and Matheson (2006) suggested that power, and not the giving and receiving of pain, was at the core of S & M. Again, it is important, however, to distinguish individuals practicing S & M as part of consensual sexual activity from individuals who have been arrested for such activity and are in the forensic system.

There also is little information on how many individuals seek help because of their sadomasochistic orientation.Weinberg (2006) concluded his review of the social and psychological literature by stating that ‘‘sociological and social psychological studies see SM practitioners as emotionally and psychologically well balanced, generally comfortable with their sexual orientation, and socially well adjusted’’(p.37). Ina study of 245 manifestly sadomasochistic West German men, Spengler (1977, 1983) reported that 20% rejected their sadomasochistic orientation, 70% accepted it, and 9%‘‘didn’t know.’’ Ninety percent had never visited a doctor, psychiatrist, or psychologist because of their sadomasochistic deviation, but 10% reported doing this at least once. Another study by Moser and Levitt (1987) reported on the results of a questionnaire given to 178 men self-defined as S & M. Most respondents were satisfied with the S & M part of their sexuality, but 6% expressed distress concerning their behavior and 16% had sought help from a therapist for their S &M desires.

Finally, the focus and nature of therapy for those from the community who might present to practitioners is different from the focus of those who are in forensic situations. One might anticipate that therapy for those practicing S&M may involve issues other than their S&M or involve‘‘normalizing’’ (i.e., making acceptable) their sexual fantasies or behavior (Kleinplatz & Moser, 2004; Nichols, 2006). With forensic populations, the focus would be on controlling or suppressing sadistic arousal and behavior (Krueger & Kaplan, 2002).

These observations suggest that there is a substantial occurrence of sadomasochistic behavior in the community, that some research is being done on it, and that some people seek out consultation from mental health professionals for this. It would appear, however, that the DSM is not being used for research purposes for this population and perhaps not for clinical purposes either.

Relationship and Cultural Context

Mitchell and Graham (2008) raised the issue that relationship influences are not considered in the diagnosis of sexual disorders and Tiefer (2004) and Tiefer, Brick, and Kaplan (2003) noted that both relationship and cultural context are important in assessing and treating sexual disorders. It is notable that the paraphilias, presumably because some of these behaviors are illegal and nonconsensual, do not include any relationship specifiers. Given that sadomasochismis one of the paraphilias that could occur in the context of a relationship (along with transvestic fetishism, and perhaps some of the other unnamed paraphilias), it might make sense to consider including this dimension in the criteria.

Misuse of DSM in Child Custody Proceedings and Discrimination

Klein and Moser (2006) described the case of the misuse by forensic professionals of the DSM criteria in a child custody suit, suggesting that these not infrequent cases should be an impetus to the editors of the DSM to reevaluate its classification of atypical sexual behavior as pathological and to strengthen its warnings against misuse. Wright (2006) presented information on violence and discrimination against SM-identified individuals; of 1017 SM individuals surveyed, 36%had suffered some sort of violence or harassment because of their SM practices, and 30% had been victims of job discrimination.

Recommendations and Discussion

Should Sexual Sadism Be Retained in the DSM?

Yes. The above summaries make clear that Sexual Sadism is a prominent diagnosis and entity in forensic populations. It, along with other psychiatric diagnoses, presents a clear target of treatment. Treatment of psychiatric conditions is a cornerstone in addressing and reducing risk in forensic populations. In some places in the narrative section, there are descriptions of sadistic behavior or other assertions without the caution that much of the information is derived from forensic populations and may not apply to community populations. The narrative section of the DSM should be rewritten to reflect this. Additionally, caveats circumscribing the application of the DSMin forensic matters, particularly as regards Sexual Sadism and Sexual Masochism, should be reviewed and strengthened.

Should There Be Any Changes in the Diagnostic Criteria?

Yes. The current criteria are listed in Appendix 1. I would recommend the following changes:

1. The phrase ‘‘or behaviors’’ be deleted from criterion A. This would address the concerns raised by the editors of DSM-IV-TR (First & Frances, 2008) that inclusion of the term ‘‘or behavior’’ allowed for the inappropriate conclusion that an individual qualified for a mental illness solely on the basis of repeated criminal acts.

2. The phrase ‘‘real, not simulated’’should be deleted from the A Criterion. I cannot see that this adds any real distinction. This appears to have been added in the second criterion (2) in DSM-III for the diagnosis of Sexual Sadism, but there is no information as to the reason this was added.

3. Should the criteria be expanded to include cruelty or torture, sexual mutilation, and deviant sexual arousal as Marshall et al. (2002) have suggested? Should the criteria be modified to include behaviors that were common to three of the four conceptualizations identified by McLawsen et al. (2008), and summarized by the following phrases: ‘‘slapped or punched victim during the sexual act; cut, stabbed, strangled, bit, or beat victim during sexual act; and, physical restraints used during sexual act?’’
No. I think that each of these studies does not present enough evidence to expand on or alter the definitional items in Criteria A. I would strongly recommend the development and use of structured diagnostic instruments for the validation of diagnostic criteria and exploration and validation of other possible items that may be relevant to Sexual Sadism in the clinical and forensic areas. An abundant literature supports the utility of such structured instruments in increasing interrater reliability in other areas of psychiatric diagnosis (Kranzler et al., 1995;Miller, Dasher, Collins, Griffiths, & Brown, 2001; Shear et al., 2000; Steiner, Tebes, Sledge, & Walker, 1995) and I would suggest creation of structured diagnostic instruments for the paraphilias and questionnaires that could yield survey more information about other features or behaviors associated with this diagnosis. Further, sexual surveys are done in an annual way on all sorts of sexual behavior by the U.S. Government and, with appropriate protections related to self-incrimination, identity protection, and sensitively designed survey questions, I see no reason why structured instruments could not be developed for the paraphilias in future government or academically conducted surveys.

4. What about dimensional ascertainment for Sexual Sadism and poor interrater reliability? Marshall and Kennedy (2003) recommended abandoning the present diagnostic criteria and shifting to a dimensional approach to defining sadism. I am in favor of exploring dimensional approaches, but not of abandoning the diagnostic criteria.

It should be noted that this summary reflects my original literature review. Subsequently, interactions with other members of the workgroup and advisors have resulted in modification of these initial suggestions.

Acknowledgments
This article was prepared with the assistance of Dr. Meg Kaplan. The author is a member of the DSM-V Workgroup on Sexual and Gender Identity Disorders (Chair, Kenneth J. Zucker, Ph.D.). I wish to acknowledge the valuable input I received from members of my Paraphilias subworkgroup (Ray Blanchard, Marty Kafka, and Niklas Langstrom) and Kenneth J. Zucker. Reprinted with permission from the Diagnostic and Statistical Manual of Mental Disorders V Workgroup Reports (Copyright 2009), American Psychiatric Association.

Appendix 1: Sexual Sadism

Diagnostic Criteria for Sexual Sadism from DSM-I to DSM-IV-TR

DSM-I (American Psychiatric Association, 1952)

The only mention of sexual sadism occurs under the categorization of Sociopathic Personality Disturbance (000-x60):

Sexual Deviation. This diagnosis is reserved for deviant sexuality which is not symptomatic of more extensive syndromes, such as schizophrenic and obsessional reactions. The term includes most of the cases formerly classed as ‘‘psychopathic personality with pathologic sexuality.’’ The diagnosis will specify the type of the pathologic behavior, such as homosexuality, transvestism, pedophilia, fetishism and sexual sadism (including rape, sexual assault, mutilation).

DSM-II (American Psychiatric Association, 1968)

Sadism is classified as one of the Sexual Deviations (302.6):
Sexual Deviations. This category is for individuals whose sexual interests are directed primarily towards objects other than people of the opposite sex, toward sexual acts not usually associated with coitus, or toward coitus performed under bizarre circumstances as in necrophilia, pedophilia, sexual sadism, and fetishism. Even though many find their practices distasteful, they remain unable to substitute normal sexual behavior for them. This diagnosis is not appropriate for individuals who perform deviant sexual acts because normal sexual objects are not available to them.

DSM-III (American Psychiatric Association, 1980)

Sexual sadism is classified as one of the paraphilias, with one of the following criteria necessary for the diagnosis:
(1) on a nonconsenting partner, the individual has repeatedly intentionally inflicted psychological or physical suffering in order to produce sexual excitement
(2) with a consenting partner, the repeatedly preferred or exclusive mode of achieving sexual excitement combines humiliation with simulated or mildly injurious bodily suffering
(3) on a consenting partner, bodily injury that is extensive, permanent, or possibly mortal is inflicted in order to achieve sexual excitement.

DSM-III-R (American Psychiatric Association, 1987) The diagnostic criteria for sexual sadism were revised as follows:
A. Over a period of at least six months, recurrent intense sexual urges and sexually arousing fantasies involving acts (real, not simulated) in which the psychological or physical suffering (including humiliation) of the victim is sexually exciting to the person.
B. The person has acted on these urges, or is markedly distressed by them.

DSM-IV (American Psychiatric Association, 1994)
The diagnostic criteria for sexual sadism were:

A. Over a period of at least 6months, recurrent, intense sexually arousing fantasies, sexual urges, or behaviors involving acts (real, not simulated) in which the psychological or physical suffering (including humiliation) of the victim is sexually exciting to the person.

B. The fantasies, sexual urges, or behaviors cause clinically significant distress or impairment in social, occupational, or other important areas of functioning.

DSM-IV-TR (American Psychiatric Association, 2000)
The change in the B. criterion from DSM-IV to DSM-IV-TR represents one of the few changes in criteria from DSM-IV to DSM-IV-TR. This change was made to all of the paraphilias which involved a victim, to remove any ambiguity about whether acting out sexual urges with others was sufficient for a diagnosis; some had argued that an individual with a paraphilia who was not distressed about his or her behavior could not be diagnosed with a paraphilia, and this new wording allowed for a diagnosis to be made in such a circumstance.

The diagnostic criteria for sexual sadism were revised from DSM-IV:
A. Over a period of at least 6months, recurrent, intense sexually arousing fantasies, sexual urges, or behaviors involving acts (real, not simulated) in which the psychological or physical suffering (including humiliation) of the victim is sexually exciting to the person.
B. The person has acted on these sexual urges with a nonconsenting person, or the sexual urges or fantasies cause marked distress or interpersonal difficulty.

Suggested Criteria Following Literature Review for DSM-V

These criteria reflect my initial suggestions. Subsequently, interactions with other members of the workgroup and advisors have resulted in a modification of these initial suggestions.
A. Over a period of at least 6months, recurrent, intense sexually arousing fantasies or sexual urges involving acts in which the psychological or physical suffering (including humiliation) of the victim is sexually exciting to the person.
B. The person has acted on these sexual urges with a nonconsenting person, or the sexual urges or fantasies cause marked distress or interpersonal difficulty.

Appendix 2: Sexual Sadism
The ICD-9 and ICD-10 Criteria for Sexual Sadism and Sexual Masochism and the ICD-10 Diagnostic Criteria for Research for Sadomasochism
The ICD-9-CM Diagnostic Criteria for Sadism and Masochism (World Health Organization, 1989) (p. 229) are: 302.8 Other specified psychosexual disorders 302.83 Sexual masochism 302.84 Sexual sadism

The ICD-10 International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (World Health Organization, 1992) (p. 367) criteria are:
Disorders of sexual preference includes: paraphilias F65.5 Sadomasochism

A preference for sexual activity which involves the infliction of pain or humiliation, or bondage. If the subject prefers to be the recipient of such stimulation this is called masochism; if the provider, sadism. Often an individual obtains sexual excitement from both sadistic and masochistic activities.
Masochism
Sadism

The ICD-10 Classification of Mental and Behavior Disorders Diagnostic criteria for research (World Health Organization, 1993) are:

F65.5 Sadomasochism (p. 137)
A. The general criteria for disorders of sexual preference (F65) must be met.
B. There is preference for sexual activity, as recipient (masochism) or provider (sadism), or both, which involves at least one of the following: (1) pain; (2) humiliation; (3) bondage.
C. The sadomasochistic activity is the most important source of stimulation or is necessary for sexual gratification. F65 Disorders of sexual preference (p. 135)
G1. The individual experiences recurrent intense sexual urges and fantasies involving unusual objects of activities
G2. The individual either acts on the urges or is markedly distressed by them.
G3. The preference has been present for at least 6 months.