IN THE NEWS: Junk-science paraphilias remain popular despite official rejection, study finds

Sometimes, you can’t win for losing.

Just over a decade ago, opponents of junk science in court won a hard-fought battle when they succeeded in keeping two unreliable sexual-deviance diagnoses from debuting in the fifth edition of the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM).

Now, a new study finds that the rejection did nothing to stop the introduction of these diagnoses in court. Rather, they are being snuck into forensic reports and testimony through the back door, via two vague catchall labels inserted into the DSM manual in 2013. And although proponents had argued at the time that these residual labels would reduce confusion and improve diagnostic reliability, the study suggests that the opposite has occurred.

Long-time readers of this blog may recall the brouhaha over the two novel conditions of “hebephilia” and “nonconsent.” Both were considered but rejected for the sexual disorders (“paraphilias”) section of the 2013 DSM. Their rejection owed to their lack of proven reliability or scientific validity. Neither condition has a standard definition, which is a basic precursor to accurate scientific measurement. Hebephilia generally references a sexual attraction to youths in the pubertal stage of development, while nonconsent refers to attraction to sexual coercion.

A single niche

The single niche where the two labels are in widespread use is a forensic one: Sexually violent predator (SVP) litigation. That’s because the indefinite civil confinement of serial sex offenders has been ruled unconstitutional except in cases where an offender poses a substantial future danger to the public due to a formal mental disorder. The lobby to create the new disorders of nonconsent and hebephilia was led by forensic psychologists working in the SVP trenches, along with psychologists at a Canadian clinic with outsized influence over the paraphilias section of the 2013 DSM manual. The American Psychiatric Association’s refusal to label rapists as mentally ill has encouraged some evaluators to “bend the language of the DSM” to make it work.

The current researchers found that “nonconsent” and “hebephilia” are the two most common bases for invoking an idiosyncratic catchall label of “Other Specified Paraphilic Disorder” (OSPD). Their findings are consistent with a recent review of U.S. legal cases that found that large proportions of civilly committed sex offenders – including about half in California and 43% in Washington - are diagnosed with "OSPD-nonconsent."

The study, published in the journal Sexual Abuse, is the first to systematically analyze the prevalence and patterns of use of OSPD and another vaguely defined label, “Unspecified Paraphilic Disorder” (UPD), in sexually violent predator litigation. It analyzed SVP evaluations in Florida over a four-year period. Because the researchers aimed to calculate the reliability of the disputed labels, only cases in which a convicted sex offender was evaluated by two different psychologists were included. In all, 190 separate cases involving 380 forensic reports were analyzed.

At least one paraphilia was diagnosed in four out of five cases reviewed. Pedophilia was the most invoked, followed by the catchall categories of OSPD and UPD.

OSPD’s reliability – or the agreement among two psychologists evaluating the same man – was abysmal. In cases where one evaluator assigned a diagnosis of OSPD, there was a less-than-chance likelihood that a second evaluator would agree. The kappa reliability statistic was a very poor .21, far below chance agreement. Kappas of below 0.4 are generally considered to be below the minimum reliability threshold in the forensic arena.

Evaluator disagreement was even more profound with Unspecified Paraphilic Disorder, with two psychologists agreeing about its presence only 30% of the time. That comes as no surprise. That label, as critics have long pointed out, is inherently unreliable, in that it is designed to be used in circumstances in which there is not enough information to make a specific diagnosis, or a clinician “chooses not to specify the reason” why it is being assigned, according to the manual’s instructions.

One of forensic psychology’s dirty little secrets is that the assignment of controversial labels often hinges as much on evaluator whims as on the facts of the case. For example, research has found that some evaluators routinely assign higher scores than others on measures of psychopathy, an especially prejudicial label. The current research showed this same problematic pattern with diagnoses of OSPD. Two of the 21 psychologists under study proffered that catchall diagnosis in most of their cases, whereas 38% of the clinicians assigned it in fewer than one out of four cases; one evaluator never used it at all. This suggests that case outcomes are being influenced not only by offender characteristics but by which psychologist happens to be assigned to the case.

Similar evaluator variability was evident when the researchers zoomed in on OSPD diagnoses in which either hebephilia or nonconsent were proffered as its basis. Three evaluators used the term “hebephilia” in half of their OSPD diagnoses, while nine evaluators never used hebephilia-related terminology at all. And evaluators agreed on the hebephilia label in only about one out of four instances. Regarding nonconsent, 13 evaluators invoked it in at least half of their evaluations, whereas five evaluators never used that specifier.

The study’s authors theorized that the widely ranging rates of use of the OSPD and UPD labels likely reflect hesitancy by some psychologists to proffer diagnoses “with vague diagnostic criteria and debatable level of empirical support.”

What all this suggests is that whether an offender is said to have a mental disorder pertaining to an attraction to pubescent minors and/or rape hinges in large part on the luck of the draw as to whether they are assigned to Dr. Jones versus Dr. Smith.

The large variance among evaluators is especially remarkable in that “adversarial allegiance” was not in play. This forensic bias becomes an issue when evaluators’ opinion are influenced by whether they were retained by the prosecution or the defense. Here, all of the evaluators were members of the same ostensibly neutral panel of contracted psychologists. If adversarial allegiance had come into play, the divergences in diagnoses likely would have been even more profound.

Highlighting the higgledy-piggledy nature of any ad-hoc diagnosis, the researchers found that the so-called “specifiers” – or specific rationales – attached to OSPD diagnoses were highly idiosyncratic. Examples included descriptions of behaviors that are illegal but not necessarily evidence of mental disorder, such as “OSPD-Non-Consensual Sexual Activity with Adolescent,” “OSPD-Attraction to Adolescent Females” and an even more bizarre “OSPD-Sexting.”

Custom-tailored labels

“[O]ne may be particularly concerned that several of the labels appear custom to the facts of the specific case rather than resting on any empirically derived diagnosis,” the study’s authors noted.

I witnessed this first-hand last month, when a psychologist testified in federal court that a sex offender the government was aiming to civilly commit had a novel combination of sexual interests that cumulatively rose to the level of a unique mental disorder called “OSPD-deviant sexual interests in hebephilic, sadistic, exhibitionistic and voyeuristic behavior.”

Fortunately, the federal judge at this particular trial was skeptical. Pointing out that “OSPD-hebephilia” was rejected from the DSM and remains controversial in the psychological community, he wrote in his opinion that he was “troubled by the combination of multiple insufficient specifiers, which does not appear to have been contemplated by the DSM-5-TR.”

No matter how nonconsent or hebephilia were defined in the specific psychological reports, the interrater agreement – or concordance between evaluators – remained poor across the board, and far below recommended reliability for diagnoses in routine clinical practice, much less the forensic arena in which precision is especially critical.

"Bad science"

“Relying upon diagnoses with poor empirical support can perpetuate the use of bad science in the courtroom,” the authors concluded. “While it is certainly true that there are high-risk individuals who are likely to sexually recidivate upon their release from prison, providing makeshift diagnoses to satisfy civil commitment criteria significantly questions the ethical practice of psychological decision making.”

A survey of legal cases found a smattering of successful challenges to these controversial diagnoses. These Daubert and Frye evidentiary challenges focused on definitional problems, an absence of substantial research support, and a lack of general acceptance. In State of New York v. Jason C., for example, the court wrote:

“This Court cannot help but ask, if this disorder exists, why isn't there convincing evidence that it exists outside the realm of civil commitment? If this disorder is a matter of the human condition, then shouldn't this paraphilia be seen outside of SVP proceedings?”

The diagnosis was similarly excluded in a Missouri case, In Re: Stanley Williams, on the basis of a high error rate, a dearth of peer-reviewed publications, poor validity, and lack of general acceptance. The judge in that case wrote:

“Using diagnostic language which has been rejected from inclusion in the DSM does not indicate general acceptance by the relevant community, but rather an unwillingness to accept the given methods and language in question.”

The study, "Other Specified Paraphilic Disorder: Patterns of Use in Sexually Violent Predator Evaluations," is authored by Nicole Graham, Cynthia Calkins and Elizabeth Jeglic of the John Jay College of Criminal Justice in New York.