The article entitled, “The Impact of Surgical Castration on Sexual Recidivism Risk Among Sexually Violent Predatory Offenders,” which appears in this edition of the Journal, takes up several important subjects of both a psychiatric and a forensic nature.1 There are two fundamental questions: first, what does the scientific evidence have to say about the likely impact of surgical castration on sexual drive and the enactment of sexually motivated behavior—most specifically, sexual behavior that is criminal in nature? Second, how heavily should the impact of a testosterone-lowering intervention, such as surgical castration, be weighted when attempting to determine the likely risk of future sexual recidivism by a previously civilly committed sexual offender who is seeking possible release into the community? It should be emphasized that the authors have conducted a retrospective review that attempts to predict the possible effects of surgical castration on civilly committed sexual offenders, rather than a prospective study capable of actually demonstrating its impact.
Because, in my judgment, Weinberger et al.1 appear to be most concerned that the castration data reviewed not be too heavily weighted in support of a possible release into the community, this commentary is meant to balance that argument by cautioning against underestimating its possible importance when it comes to supporting such a release. I also argue that both public safety and fairness to those who have been civilly committed are likely to be better served by the community-based provision of psychiatric follow-up interventions capable of decreasing any potential risks that may be of concern. Effective risk management both at the time of release and following re-entry into the community, may better serve the interests of both patients and the public, than continuing to emphasize the current questionably accurate “crystal ball approach” of long-term risk prediction.
The literature review conducted by the authors of the accompanying article appears to establish firmly that lowering testosterone by means of surgical castration is generally associated with a marked decline in sexually motivated behavior, including that of a criminal nature. In virtually all species of animals that have been studied, including humans, castration has generally been associated with either a reduction or a complete elimination of sexual activity and interest. As documented by the authors, studies of medical patients who were castrated because of testicular or prostate cancer, have consistently reported a significant decline in both the intensity of sexual drive (libido) and the frequency of sexually initiated behavior. That has been so despite the fact that some men have maintained their ability to perform genitally (i.e., they have maintained some degree of erectile capacity) following castration. As a consequence, some have also maintained their ability to become at least somewhat aroused genitally in response to visually presented stimuli, such as sexually explicit photographs.
Weinberger et al. seem quite concerned about the potential preservation of some degree of erotic erectile capacity after surgical castration. However, it should be emphasized that the intended goal, when utilizing testosterone-lowering interventions to treat sexually disordered patients (and clearly many who have been civilly committed are sexually disordered), is to decrease the intensity of their paraphilic cravings. Such cravings can function as a motivational factor, which arguably would be of far greater importance in influencing their behavior than would be the capacity to sustain some degree of penile function. The critical question, at least in the case of the sexually disordered patient who is working to succeed, is not whether he still manifests some capacity for erotic arousal, but rather whether he is now better able to suppress it, and, if inappropriate, to resist acting on it. Whether a would-be dieter who has been prescribed an appetite suppressant is going to be successful in changing his behavior, would probably depend, not so much on the capacity of any such drug to eliminate his ability to eat, but more so on its ability to lower the intensity of his cravings to do so.
The authors’ literature review correctly notes that many of the early studies involving surgical castration have suffered from several methodological shortcomings. Nevertheless, all the studies reviewed by them have reported impressively low rates of subsequent criminal sexual recidivism. Those studies have included individuals with a history of rape, as well as individuals with a history of pedophilic misconduct. The most recently reported study, a 1991 investigation by Hansen,2 included a group of sexually violent offenders, arguably similar to at least some of those now categorized as sexually violent predators. Those offenders had committed a variety of serious crimes involving severe bodily injury, rape, attempted murder, and murder. Not one of those castrated sex offenders (n = 21) had recidivated over several years, although two had done so more than 15 years later, but only after having been administered testosterone-replacement therapy. Thirty-six percent of men in the comparison control group, who had not been castrated, did recidivate.
In yet another study reported by the authors, Sand et al.3 examined the records of more than 900 castrated individuals. Eighty-two percent had been documented sexual offenders—many, repeat offenders—prior to castration. Some of the individuals in that research had subsequently been tracked for periods as long as 30 years, and the reported overall sexual recidivism rate following castration was less than 2 percent. Weinberger et al.1 argue that several of the studies reviewed by them may have contained populations that differ significantly in their makeup from civilly committed sexual offenders. However, the fact remains that many studies from a variety of locations, involving a variety of samples, have documented consistently low rates of sexual recidivism following castration. Thus, it is difficult to see why one would expect currently committed patients, many of whom are likely to be similar to many others not under commitment in jurisdictions that do not legislatively allow for it, to be the exception.
Weinberger and colleagues summarize a study by Wille and Beier4 documenting that castration was most effective in reducing libido, as well as any sort of sexual activity, among those aged 45 years and older. Although older men who have not been castrated do not necessarily have significantly lowered levels of testosterone, nor do they necessarily have lowered sexual libido, castrated elderly men do show such reductions. In their study, Wille and Beier report that more than 92 percent of castrated men over the age of 60 years had reported virtually extinct libido and sexual activity following surgery. In addition, according to Wille and Beier, even within the youngest groups of castrated men (ages 30–49 years), only 33 percent had indicated that they could still function sexually following castration. Of those, many had required extensive stimulation to be able to do so. Approximately 67 percent of that younger group had reported that their sexual activity was virtually extinct six months following the surgery. Thus, castration appeared to be generally effective in lowering both libido and sexual activity across a broad age range of sexual offenders. Wille and Beier reported that only 3 percent of those castrated had engaged in recidivistic criminal sexual misconduct following surgery.
Weinberger et al.1 suggest that one should not depend solely on the subjective self-reports of sexual offenders in assessing persons for “deviant” sexual interests. Rather, they argue that penile plethysmography (PPG) is a technology that can be useful in making such an assessment and that it can be helpful when evaluating previously committed sexual offenders who are seeking community release. In terms of group data, more persons in a group of individuals who show evidence of deviant arousal (e.g., erotic arousal in response to children as assessed by PPG) are likely to engage in pedophilic sexual misconduct than would be the case in a comparison group of persons showing no such arousal.5 In contrast, when it comes to making predictions about the long-term risk of sexual recidivism for a given individual, there is evidence documenting that “deviant” arousal, as demonstrated in the PPG laboratory, is not necessarily an accurate predictor of long-term behavioral outcome in the community.6 To put it bluntly, what is important in terms of treatment outcome is not how the penis behaves in the laboratory, but rather how the man himself behaves over time when back in society. The fact that some castrated sexual offenders may continue to show some degree of penile tumescence in the PPG laboratory in response to deviant stimuli should not necessarily be construed as evidence of a high risk of subsequent recidivism in the future.
The authors present two hypothetical clinical vignettes discussing how to go about predicting the likelihood of future recidivism in the case of civilly committed sexual offenders who have been surgically castrated. In both vignettes, they seem to infer that a high risk is still present. Certainly, aspects of the hypothetical vignettes that they have elected to present (e.g., the discovery of a stash of pictures of young boys in a patient’s locker six months following castration) would be cause for concern. Clearly, one would want to give weight to post-surgical evidence suggestive of continued sexually problematic acts. How might the authors’ opinions about risk have changed had the pictures not been discovered or had there been reason to doubt that they did actually belong to the patient in question? The difficulty relates to how to determine with confidence whether the authors’ suggested methods of predicting future risk have validity, in the absence of any data regarding their ability to do so successfully.
Clearly, caution is necessary when making predictions about future recidivism with persons who have committed sexual offenses. In the past, 24 convicted sex offenders treated at a specialized hospital facility in California and released as “cured” had a 20.8 percent (5/24) re-arrest rate within a follow-up period of six and one-half years.7 Seventeen men characterized as “unamenable to treatment” did better rather than worse, with a lower 11.8 percent (2/17) re-arrest rate over a comparable follow-up period. In addition, contrary to what might have been expected, five of six other patients discharged from treatment as “nonamenable and dangerous” did not recidivate during subsequent follow-up.
Weinberger and colleagues1 have rightfully pointed out the obvious—in assessing risk, all factors thought to be relevant should be considered. That would include not only the presence or absence of sexual deviation, as well as knowledge about any interventions such as surgical castration that may have been performed, but also the question of whether any comorbid Axis I or Axis II diagnoses that may be present have been adequately addressed. That admonition not withstanding, based on the data currently available, the fact remains that in all reported instances in which sexual offenders have been surgically castrated, subsequent sexual recidivism rates have consistently been impressively low.
Some researchers have argued that the best way to assess the risk of recidivism with respect to sexual offenders is to make use of statistical actuarial data.8 I do not necessarily subscribe to that argument, because such actuarial methods are intended to make predictions about the behavior of groups, rather than about the behavior of specific individuals.9 For example, insurance companies can make use of actuarial data to predict the percentage of individuals within a given group, perhaps made up of obese, hypertensive, cigarette-smoking males manifesting high levels of cholesterol, who are likely to have a heart attack. However, they cannot utilize such actuarial data to predict accurately which specific individuals within that group are more or less likely to do so. Nevertheless, should one elect to take such an actuarial approach in making predictions, as is often done at civil commitment hearings, the data available regarding castrated sex offenders seem to suggest that that intervention is generally associated with a very low risk of future sexual misconduct. Beyond that fact, unlike research involving actuarial data, conclusions regarding the expected outcome of castration are based on a well-documented scientific understanding of hormonal biophysiology.
Where would the issue of predicting future recidivism be left, were the data regarding the likely effects of either surgical or chemical castration not to be heavily weighted? Arguably, recidivism rates following cognitive behavioral interventions, long-term institutional treatment, and relapse prevention therapies have all been associated with higher rates of subsequent sexual recidivism than have been reported in the case of castration.10 If castration data should not be weighted heavily when considering possible release into the community, in general are there any other treatment data that might more powerfully and validly predict a safe re-entry into society?
Weinberger et al.1 argue that factors unrelated to testosterone levels can be important in effecting the likelihood of future sexual recidivism. While that is undoubtedly true, there is every reason to believe that such factors were also present in the case of those groups of sexual offenders who have undergone surgical castration and have not recidivated.
Castration addresses a fundamental element that has been used to justify the civil commitment of sexual offenders in the first place—that is, the element of volitional impairment.11 The repeat bank robber is not civilly committed because his behavior is not thought to be related to a psychologically impairing mental disorder. However, it is now generally accepted that powerful biologically based cravings, be they related to overeating, alcohol, drugs, sex, pain, or the need for sleep, can sometimes overcome even stern resolve. By decreasing the intensity of sexual cravings capable of wearing down such resolve, castration can concomitantly increase the volitional capacity to maintain appropriate self-control.
The article provides much useful information, and a sound literature review of data, both from the United States and abroad, related to surgical, and in some instances chemical, castration. Weinberger et al.1 have also discussed some of the considerations related to ethics. From a treatment standpoint, the data presented seem to suggest clearly that lowering testosterone can provide many sexually disordered patients with the equivalent of a sexual appetite suppressant, thereby facilitating better behavioral control. Parenthetically, it should be noted that there seems to be little reason to favor surgical castration, given the fact that the same testosterone-lowering effects can be induced pharmacologically. The problem, in the case of the civilly committed sexual offender, lies not in the rationale underlying the use of castration as a form of treatment, nor necessarily even its likely effects on subsequent behavior, but rather in deciding how much weight to give to the fact that it has been done, when trying to predict future risk. Psychiatrists and other mental health providers are better equipped to manage risk—that is, to reduce it—rather than trying to predict it in the abstract.
Most civilly committed sexual offenders are sexually disordered. Sexually disordered offenders—that is, those with a paraphilic disorder—cannot be cured. However, research evidence suggests that many can be successfully treated.12,13 In general, the treatment of any chronic behavioral disorder, be it drug addiction, alcoholism, or a paraphilia, depends on the availability of adequate community-based resources, in some instances following a period of residential care. As noted earlier, rather than predicting how the castrated sexual offender is likely to fare following release into the community, it may be more crucial to manage effectively any risk that may be present at the time of his re-entry. Perhaps that could be accomplished best by insuring that all such releases are conditional (as is the case in several states), as opposed to unconditional, so that ongoing support, treatment, and monitoring can be regulated. In that way, one could guard against the surgically castrated sex offender’s reversing his status by means of the ingestion of exogenous testosterone. A simple mandated blood test could quickly detect any attempt to do so. A variety of other safeguards would then also be possible, including electronic surveillance, and prophylactic recommitment, if indicated. If, in the meantime, predictions must nevertheless still be made, then the available scientific research seems to suggest that, in general, low rates of sexual recidivism can ordinarily be expected following surgical castration.
- American Academy of Psychiatry and the Law