Commentary: Muddy Diagnostic Waters in the SVP Courtroom

The broad mission of the paper by Michael First and Robert Halon¹ addresses the abuses of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM‐IV‐TR)² in civil commitment hearings of sex offenders. The narrower mission of the paper focuses on the second‐prong requirement of these Sexually Violent Predator (SVP) statutes. These statutes, all of a similar kind, require evidence of a nexus between “mental abnormality,” or personality disorder, and prior bad (sexual) acts. There is a lengthy history of the use of mental disorder as a threshold decision in civil commitment of sexual offenders, dating back to the 1939 Minnesota Supreme Court case State ex rel Pearson v. Probate Court.³ As First and Halon point out, the constitutionality of these laws rests now, as it has in the past, on the presence of some form of mental disorder. The precise guidelines for what constitutes a statutory mental disorder have never been elucidated, however, beyond the characteristic presence of impaired volition.⁴ Impaired volition has never been elucidated either. Thus, despite the criticality of a statutory mental disorder, no clear lines define the boundaries of this elusive human condition.⁵

First and Halon tackle this nosological problem by first pointing to the APA's Task Force Report on Dangerous Sex Offenders,⁶ which raises a seeming conundrum—only paraphilias appear to satisfy the statutory intent of a mental abnormality, but the incidence of paraphilias among sex offenders is quite low. More prevalent diagnostic conditions, such as anger‐management problems, substance abuse, antisocial personality traits, and deficits in social and interpersonal skills “have little explanatory connection to the offender's sexual behavior” (Ref. 6, p 9) and thus fail to satisfy the second (diagnostic) prong of the statute. Consequently, despite their low incidence, paraphilias have assumed center stage in the adjudication of SVP cases.

First and Halon contend that: …during the process of adjudication of SVP commitment trials, profound and avoidable errors are made by some mental health professionals who invalidly diagnose paraphilia, assert that there is volitional impairment based solely on the fact that the offender has a paraphilia diagnosis, and thus wrongly claim that the statutorily defined SVP commitment criteria are adequately addressed by the clinical diagnoses [Ref. 1, p 444].

A similar conclusion was reached in APA's earlier Task Force Report, “Use of Psychiatric Diagnoses in the Legal Process.”⁷ Under the topical heading of “Unfounded Intuitions about Mental Disorders and Individual Control,” Halleck et al. noted that, “The problem of volition is especially vexing when the disorder is largely defined on the basis of behavioral characteristics” (Ref. 7, p 5). Halleck et al.⁷ use paraphilias as their example. We argue that First and Halon's contention is not only accurate but more than likely an understatement (i.e., the diagnoses made in SVP cases frequently are reflexive and indefensible).

The reason is twofold. First, the DSM was never designed to provide taxonomic differentiation among sex offenders. First and Halon allude to taxonomic differentiation, albeit with reference to paraphilic fantasies, by citing, somewhat incorrectly, the Crime Classification Manual.⁸ The subtypes reported by First and Halon were described by Groth many years earlier.^9,10 Douglas et al.⁸ created a rationally derived, highly differentiated, hierarchical system that includes approximately 16 categories, four of which were borrowed from Groth.

The essence of the problem is that sex offenders constitute a markedly heterogeneous group of criminals (e.g., Refs. 11 and 12). It was never the mission of the DSM to provide a taxonomic home for subtypes of criminals, including sex offenders. The net result is that the DSM provides little help in differentiating among sex offenders. A one‐size‐fits‐all category (pedophilia), for example, subsumes all child molesters that meet the time requirement. Although pedophilia is an appropriate omnibus category for the purposes of the DSM (i.e., it recognizes a paraphilia characterized by sexual attraction to children), it does not serve the needs of the courts. In effect, all child offenders (who meet the six‐month criterion) are automatically classified as pedophiles, regardless of criminal history and regardless of evidence of volitional impairment.

From a legal standpoint, the courts must regard the diagnosis of pedophilia as a psychometric example of the delimitation of validity as a function of enhanced reliability.¹³ Reliability is a necessary but insufficient criterion for validity. For a diagnosis to be valid, assignments must be reliable. As the boundaries of the diagnosis increase, so does reliability (i.e., the larger the diagnostic target, the easier it is to hit). With regard to pedophilia, we have highly reliable assignments to a category that tells us very little about risk or serious difficulty.

Child molesters, at least, have a home in the DSM, albeit a specious one.

Rapists, on the other hand, have none, resulting in the creation of a diagnosis of convenience (paraphilia: NOS‐nonconsent). Although we concur with First and Halon that it is plausible that some individuals evidence preferential arousal to nonconsenting partners, in practice this newly contrived diagnosis reflects bad faith, bad science, and, often, bad clinical judgment. In practice, a differential diagnosis between nonconsent and sadism is essential. Moreover, rather than the diagnosis’ being liberally applied to most rapists petitioned for commitment, it should be sparingly applied to very few. Finally, it is incumbent on the clinician, per First and Halon's comment about “Doren's formulation” (Ref. 1, p 451), to provide clear and direct evidence that the nonconsent of the victim is consistently associated with sexual arousal (as opposed to myriad other cues). Given the absence of any reliability and validity data on this diagnosis, the court should demand the “clear and direct evidence” mentioned earlier.

Logically, it appears to make little sense to rely on a newly crafted diagnostic category (paraphilia: NOS‐nonconsent) that has no empirical support, has no established criteria for classification, and is of highly questionable reliability, when there is a considerable body of science that has identified factors associated with sexual aggression against women, including misogynistic anger, negative or hostile masculinity, rape‐related cognitive distortions (rape myths) and entitlement, and impersonal sex (e.g., Refs. 14, 15). The answer, of course, is equally simple. The components of Professor Malamuth's Confluence Model of Sexual Aggression, or diagnostic categories issuing from it, do not appear in the DSM.

A second explanation for the abuses of the DSM rightly decried by First and Halon, as well as the APA Task Force Reports and several recent law review articles^4,16 is the highly adversarial nature of these SVP cases, which sacrifice objectivity for a priori judgment (clinical opinions reached prior to conducting an independent evaluation). Rather than gathering all clinical data before arriving at one's ultimate conclusion, the pivotal conclusion is reached first (i.e., whether the respondent has or does not have a mental abnormality) and data are gathered to defend the a priori conclusion. This latter, all too common approach promotes weak arguments and unsupportable conclusions. Halleck et al.⁷ pointed out that, “Ultimate legal determinations are rightly the domain of representatives of the community. The moral dimension of determining responsibility is not a task to be ceded to psychiatrists” (Ref. 7, p 7). It appears that clinicians testifying in SVP cases are increasingly guided by a moral high ground of protecting community safety or the respondent's liberty interests.

The core of the First and Halon paper presents a three‐step process for reforming the diagnosis of paraphilias in SVP proceedings. The first step is to determine that the respondent can be classified legitimately as having any paraphilia. The second step, given the presence of a paraphilia, is to determine whether the sexual (battery) offenses occurred “as a direct consequence of that paraphilia” (Ref. 1, p 448). The third step is to determine if the respondent with a paraphilia causally linked to sex‐offending behavior is volitionally impaired. The authors acknowledge that Step 3 is fraught with uncertainty. We take no exception to anything that First and Halon have proposed, only the practicality of it.

With regard to Step 1, First and Halon might remind us of the 1999 Task Force Report, which observed that “…these conditions [paraphilias] appear to be absent in most offenders” (Ref. 1, p 444). Although empirical data on base rates of paraphilias in sex offenders are minimal, the most often cited study was that of Abel and colleagues.¹⁷ They gathered confidential self‐report data on 561 paraphilics in treatment programs. Of this sample, 10.7 percent sexually assaulted adult women and 49 percent sexually assaulted children. These 561 paraphilics disclosed 291,737 paraphilic acts. Of these, .5 percent involved rape of adult women, a percentage smaller than acts of bestiality (.9%). As Abel et al. observed, “Rape, although frequently getting media coverage, is a very infrequent act, relative to other paraphilic behaviors” (Ref. 17, p 681). If the base rates for paraphilias are indeed this low among rapists, relying on paraphilias as a solution to the second‐prong mental abnormality requirement is suboptimal (i.e., it will exclude the substantial majority of all respondents). First and Halon may have precisely the right idea (i.e., a statutory screening mechanism that selects only a very small proportion of the petitioned offenders). The likely outcome, however, is that virtually all offenders with child victims will still be called pedophiles, and virtually all offenders with adult victims will be classified as having paraphilia: NOS. Hence, we will be back where we started from.

With regard to Step 2, as First and Halon are well aware, determining direct consequence is a difficult statistical problem, typically requiring regression analysis. We argue that it is rare that record data provide evidence permitting such conclusions. Rather, what one typically sees is abundant evidence suggesting associations between events (i.e., correlative evidence). If clinicians know they are expected to testify that the sex offending behavior was directly related to a legitimate paraphilia, the most likely result will be a clinical opinion that such a relationship exists (or does not exist), absent any hard (empirical) evidence.

Step 3 (for the miniscule number of respondents that make it past Steps 1 and 2) requires a determination of incapacity or lack of control. Assessing volitionality is perhaps the most hopeless of all diagnostic quagmires.^18–20 First and Halon are abundantly clear about this, quoting a 1983 APA article: “The line between an irresistible impulse and an impulse not resisted is probably no sharper than that between twilight and dusk” (Ref. 21, p 22). With regard to volitional impairment, we have remarked that, “we are saddled with an 800‐lb. diagnostic gorilla, an essential defining characteristic of an indispensable statutory element that the weight of scholarly opinion regards in the same vein as divination.”⁵

There is no answer to this insoluble problem, other than to excise “volitional impairment” from statutory language (unlikely) or to provide strict operational guidelines for what constitutes incapacity (equally unlikely). We commend First and Halon for drawing attention to a serious problem, one that undermines the integrity of the legal system in general and the SVP adjudicatory process in particular, the integrity of the science that is admitted as evidence, and the integrity of the professional organizations (“The misuse of psychiatric diagnoses has a corrosive effect on the psychiatric profession” [Ref. 7, p 7]) that represent the clinicians who testify in these cases.

We have commented elsewhere that, “The tolerance by the legal system for nonstandard and nonauthoritative diagnoses suggests strongly that the legal system's reliance on diagnostic testimony is largely pretextual” (Ref. 4, p 382). The reference to Professor Perlin's legal “fictions” is clear. He warned that “Toleration of ‘sleight of hand’ in the law's theoretical bases breeds cynicism and fosters an atmosphere of systemic manipulation by litigants, legislators, litigators, and courts” (Ref. 22, p 632). We have argued that the misuse of science, inclusive of the misuse of the DSM, in the SVP courtroom is a variation of pretextuality.

Since there is no empirical remedy on the horizon, we must rely on the courts to exercise their gatekeeping function with a tight reign. Courts should demand that expert testimony be based on empirically defensible diagnostic principles, and that opined diagnoses have a sound empirical relationship to reoffense risk in the respondent. More specifically, we have argued that the courts must demand that experts identify legitimate DSM diagnoses that are recognized by the psychiatric community, defend the appropriateness of that diagnosis for the respondent, and defend the third‐prong risk relevance of the diagnosis for the respondent.⁴ These three steps are roughly equivalent to First and Halon's, with the exception that our third step is not quite as restrictive (i.e., we recognize a potentially broader range of legitimate diagnoses than just the paraphilias). First and Halon require demonstration of incapacity. We agree. We argue, however, that the courts must operationalize incapacity (volitional impairment) with clear behavioral anchors that distinguish the common lack of self‐control that characterizes most criminals from the pathological impairment of self‐control that is explicit in SVP case law. Without firmer control from the courts, experts will continue to flounder, and their opinions will continue to be opaque and of questionable probative value.

• American Academy of Psychiatry and the Law