Souza

Milkey, J.

(dissenting). The majority’s well-reasoned opinion has a surficial logic that is difficult to contest. In addition, I agree that it is important that judges usurp neither the fact-finding role assigned to juries, nor the gatekeeping role assigned to “qualified examiners” (QEs) pursuant to G. L. c. 123A. Nevertheless, for the reasons set forth below, I ultimately agree with the trial judge that the Commonwealth’s evidence that George Souza is currently a “sexually dangerous person” (SDP), as defined by G. L. c. 123A, § 1, was so insubstantial that, as a matter of law, it cannot justify his continued detention. I therefore respectfully dissent.

In examining the sufficiency of the Commonwealth’s proof, it is important to consider the extraordinary context in which this dispute arises. It is uncontroverted that Souza has both committed odious crimes and fully served his punishment for those crimes; *174 indeed, he already has been deprived of his liberty for almost a decade after his prison term ended. The Commonwealth seeks to have him reconfined not in punishment for his past crimes but in anticipation that he may commit future ones. In this context, the ordinary rule barring propensity evidence does not apply. In fact, propensity is the main focus of SDP proceedings, and experts are called upon to speak directly to that issue (with seeming oracular certitude). Contrast Commonwealth v. Sepheus, 468 Mass. 160, 172 (2014) (defense counsel determined to have been constitutionally ineffective for failing to move to strike expert testimony that went directly to defendant’s guilt).

By definition, preventative detention schemes allow people to be locked up for crimes they indisputably have not committed, even in the face of the constitutional presumption of innocence. As the United States Supreme Court has held, the constitutionality of such schemes depends on the theory that the people so confined suffer from distinct mental conditions that prevent them from controlling their dangerous behaviors in the future. Kansas v. Hendricks, 521 U.S. 346, 358-360 (1997). It necessarily follows that, absent an adequate medical foundation, the constitutionality of continued confinement is called into question. See id. at 373 (Kennedy, J., concurring) (“[I]f it were shown that mental abnormality is too imprecise a category to offer a solid basis for concluding that civil detention is justified, our precedents would not suffice to validate it”). 1 This constitutional overlay needs to be kept in mind in assessing the adequacy of the nature and quantum of the Commonwealth’s evidentiary proof. When such considerations are taken into account, the Commonwealth’s proof here falls short of acceptable norms.

Certainly, the majority is correct that existing cases state that judges in SDP cases must proceed with caution before directing a verdict against the Commonwealth (or issuing a like order finding the Commonwealth’s case deficient as a matter of law). Thus, where there are competing expert opinions on whether *175 someone is an SDP, a judge is not free to pick and choose which opinions to credit; that job falls to the jury. See Commonwealth v. Reese, 438 Mass. 519, 525-526 (2003). However, the cases do not stand for the proposition that once a QE has opined that someone is an SDP, a judge therefore must allow the case to go to the jury. To the contrary, they continue to recognize that a judge properly may terminate an SDP proceeding if the Commonwealth’s evidence is “so incredible, insubstantial, or otherwise of such a quality that no reasonable person could rely on it to conclude that the Commonwealth had met its burden of proof.” Id. at 524 , quoting from Commonwealth v. Blanchette, 54 Mass. App. Ct. 165, 175 (2002). 2 In my view, this is just such a case.

Souza was sixty-nine years old at the time of trial. At that point, the statutory rape he committed was over four decades old, and the indecent assault and battery on a child (the only other sex offense at issue in this case) was over two decades old. As the Commonwealth’s lead expert, Frederick W. Kelso, Ph.D., himself acknowledged, peer-reviewed empirical studies show that once sex offenders reach their sixties and seventies, they “tend not to be very likely to commit future sex offenses.” Of course, that concession by itself does not present an insurmountable obstacle to the Commonwealth. Even if sex offenders generally are not very likely to reoffend at Souza’s age, this does not preclude proof that Souza in particular suffers from mental conditions that render him likely to do so. However, such proof is lacking on the current record.

The Commonwealth’s experts relied in great part on their classifying Souza as a “pedophile” within the meaning of the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (rev. 4th ed. 2000) (DSM-IV). *176 According to them, it was the combination of pedophilia and “antisocial personality disorder” (APD) that created the undue risk that he would reoffend. In the words of the Commonwealth’s second expert, psychologist Niklos Tomich, “Mr. Souza’s Pedophilia results in his deviant arousal and behavior and his Antisocial Personality Disorder provides him the psychological means to engage behaviorally in, and then excuse, his behavior.”

According to the DSM-IV, “a diagnosis of pedophilia requires ‘[a] period of at least six months, recurrent, intense sexually arousing fantasies, sexual urges, or behaviors involving sexual activity with a prepubescent child or children (generally age 13 or younger).’ ” Commonwealth v. Starkus, 69 Mass. App. Ct. 326, 336 (2007), quoting from the DSM-IV. As applied to the facts here, this required proof that the 1971 victim was prepubescent. The trial judge found the Commonwealth’s proof of that point legally insufficient. The majority rejects the judge’s reasoning on three grounds: (1) the Commonwealth is not bound by the definitions of the DSM-IV, (2) the state of the 1971 victim’s anatomical development is irrelevant because she was in any event well below the age of consent, and (3) the Commonwealth put forward sufficient proof that the 1971 victim was prepubescent (thus in any event satisfying the definition of “pedophilia” set forth in the DSM-IV). I address these points in that order.

We have long recognized the DSM as the standard diagnostic authority in the psychiatric and psychological professions. See Lambley v. Kameny, 43 Mass. App. Ct. 277 , 278 n.4 (1997). Nevertheless, as the majority correctly points out, in building a case that a sex offender suffers from a “mental abnormality” or “personality disorder,” within the meaning of the SDP statute, the Commonwealth is not limited to those mental conditions enumerated and defined in the DSM. See Commonwealth v. Husband, 82 Mass. App. Ct. 1, 4-5 (2012), and cases cited. Of course, this does not prohibit Commonwealth experts from relying on the DSM; indeed, given the authoritative stature that the DSM enjoys in the medical community, it is hardly surprising that many experts would base their opinions on that source. Where, as here, the Commonwealth experts did just that, it is fair and appropriate to hold them to this, and the cases that the majority cites are not to the contrary. 3 When the Commonwealth’s case is predicated upon a specific expert diagnosis of pedophilia as defined in the DSM, *177 a lack of evidence of one of the definitional criteria may not be excused. Otherwise, the Commonwealth would be relieved of its burden of proving the underlying facts on which its expert’s diagnosis was based. See Narducci v. Contributory Retirement Appeal Bd., 68 Mass. App. Ct. 127, 135 (2007) (noting the distinction between an expert’s ultimate conclusion and the “assumed” facts, which must be proved, on which the opinion is based).

As the majority also accurately notes, the 1971 victim was well under the age of consent regardless of whether she was prepubescent. Therefore, the state of her anatomical development is irrelevant for purposes of determining whether a crime had been committed. However, whether Souza committed a crime and whether his actions show that he suffered from a particular “mental abnormality” are distinct questions. The DSM-IV does not classify an adult’s attraction to anatomically developed but still underage adolescents as a “mental abnormality.” 4 While the Commonwealth’s experts could have sought to explain why they considered Souza as suffering from “pedophilia” apart from the definition in the DSM-IV, they did not do so. 5

The question remains whether the Commonwealth in fact offered sufficient proof that the victim of the 1971 crime was prepubescent. Although the DSM-IV notes the unremarkable fact that prepubescent children are “generally age 13 or younger,” it of course does not define prepubescence in those terms. It does not follow, except through false logic, that someone who is thirteen or younger therefore must be prepubescent. Even if the judge credited the defense experts’ definition of prepubescence (instead of *178 leaving that question to the jury), her ruling does not depend on this. The overriding point is that the Commonwealth failed to offer the proof that its own experts’ theory of Souza’s alleged “mental abnormality” demanded. Finally, to the extent that the majority concludes that Souza’s isolated references to the 1971 victim as “little” could constitute proof beyond a reasonable doubt that she was prepubescent, I disagree.

With the facts necessary to support the experts’ diagnosis of pedophilia not having been put in evidence, the experts’ opinion on that point cannot be used to avoid a directed verdict. See LaFond v. Casey, 43 Mass. App. Ct. 233, 237-238 (1997). 6 As we recently said, an expert opinion “premised on facts that [the expert] had gratuitously assumed and conjecture drawn from an insufficient evidentiary foundation . . . [is] inherently flawed and legally incompetent.” Commonwealth v. Acosta, 81 Mass. App. Ct. 836, 843 (2012).

To be sure, the Commonwealth’s failure to establish that Souza was properly classified as a pedophile does not mean that it cannot prove that he is an SDP. The majority is correct that the case law makes clear that proof that someone suffers from “antisocial personality disorder” (APD) by itself can be “adequate to satisfy the definitional requirements of” being an SDP. Commonwealth v. Reese , 438 Mass, at 526 n.9. In other words, where the Commonwealth has proven APD, there is no threshold requirement that it prove a second medical condition. However, it does not follow that a diagnosis of APD, without more, constitutes sufficient proof. This is especially true where, as here, the experts testified that it was the very combination of pedophilia and APD that caused the undue risk of sexual dangerousness (thus making proof of both prongs critical).

A close examination of the Commonwealth’s use of APD evidence here reveals why it did not amount to sufficient proof. To demonstrate that Souza currently suffers from APD, the Commonwealth’s experts relied principally on his obstreperous behavior while confined at the treatment center. Granted, Souza’s comportment during his decade of confinement was hardly exemplary. However, his documented violations of Massachusetts Treatment Center (treatment center) rules averaged only about *179 two per year, and they mainly involved minor infractions such as trying to get medication at an incorrect time, “[f]ailure to stand for a [head] count, sleeping during a count, [and] things of that nature.” Notably, none of Souza’s violations of treatment center rules involved any inappropriate sexual behavior. Compare Commonwealth v. Husband, 82 Mass. App. Ct. at 5 (“Commonwealth experts testified that [sex offender’s] personality disorder resulted in his inability to control his sexual impulses as evidenced by both the governing offenses and his extensive record of sexually aggressive and abusive conduct while incarcerated”).

Moreover, as the trial judge cogently observed, even though proof that someone has APD may be sufficient to satisfy the statute’s definitional requirements, this does not relieve the Commonwealth from having to prove that Souza currently has sexual compulsions on which his APD will induce him to act. Absent such proof, Souza cannot constitutionally be preventively detained. Passing over the question of whether there was adequate proof that Souza ever suffered from sexual compulsions that likely would cause him to reoffend, 7 evidence that he continued to have such compulsions at age sixty-nine was conspicuously absent. In fact, the Commonwealth did not present any evidence that Souza exhibited sexually inappropriate behavior of any kind since 1990. 8 In addition, the only objective test administered to Souza by the treatment center showed that he exhibited no clinically significant arousal to any of the sexual stimuli presented to him. 9

*180 Nor do I believe the other factors the Commonwealth’s experts relied upon supplied the missing proof. Both of the Commonwealth’s experts emphasized Souza’s refusal to admit his past sexual abuse of the two victims, something they asserted was a prerequisite to his being able to avoid reoffending. For example, in Tomich’s view, Souza could not progress to the point that he safely could be released until he “squarely face[d] the reasons for his incarceration and for his civil commitment.” Even to the extent Souza denied his offenses, 10 the import of that denial is, at a minimum, subject to significant doubt. The Commonwealth’s lead expert acknowledged that a pre-eminent empirical study found no correlation between denial and recidivism. In the face of that study, the Commonwealth offered no empirical studies or evidence of a medical consensus to support its contrary position that denial is somehow a predictor of future offending. 11

More generally, the Commonwealth’s experts insisted that the risks Souza presented to the community at large should be considered unacceptable until he has completed a treatment program at the treatment center. That view presupposes both that Souza presents unacceptable risks without treatment and that treatment *181 would address such risks. Neither proposition is self-evident, and one searches in vain for evidence to support them here. 12 In fact, the evidence that was presented tended to undercut the Commonwealth’s case. For example, the treatment center itself ruled out one form of treatment — behavioral conditioning — given Souza’s nonresponsiveness to sexual stimuli as measured by the PPG test. 13 The experts’ reliance on Souza’s failure to complete a treatment program is particularly problematic in light of the undisputed fact that Souza has profound cognitive limitations that, at a minimum, make it difficult for him to complete a classroom course of study. 14 Cf. Kansas v. Hendricks, 521 U.S. at 389-393 (Breyer, J., dissenting) (sex offenders cannot be civilly confined without being offered adequate treatment). In addition, it is undisputed that Souza’s efforts to pursue sex offender treatment were interrupted when his participation was suspended as a disciplinary sanction for his not complying with treatment center rules. In other words, for acting out while he was involuntarily cdnfined based on his allegedly not having received adequate treatment, the Commonwealth withheld the treatment that it considered necessary to allow his release.

Finally, I address the Commonwealth’s one attempt to take on Souza’s advanced age with empirically-based proof. Kelso relied in part on the “Static-99R” model, a widely-used tool that at *182 tempts to predict the degree of likelihood that a convicted sex offender will reoffend. As Kelso explained, the Static-99R model was specifically formulated to address the reduction in risk correlated with the aging process. However, a close examination of Kelso’s use of the Static-99R model shows that it provides negligible support for his position that Souza remains an SDR Kelso accepted that Souza had been married, and he acknowledged that his long-term relationship with his wife may well have lasted more than two years. Kelso also acknowledged that if this were so, then by Kelso’s own calculations, Souza would score only a five on the Static-99R test, which would place him outside the category of offenders considered to be at a high risk to reoffend. 15 None of this is to say that a sex offender may be found to be an SDP only if he scores in the high risk category using the Static-99R model. My point is merely that Kelso’s own reliance on empirically-based modeling undercut his claim that Souza was currently at a high risk to reoffend.

In sum, in my view, the trial judge applied appropriate scrutiny to the expert opinions that the Commonwealth offered and — finding them lacking in adequate foundational support — properly terminated the proceeding and ordered Souza’s release. In the face of the Commonwealth’s efforts to portray its case as adorned in the raiments of medical expertise, the trial judge dared to point out that “the emperor has no clothes.” 16

See also Matter of State of N.Y. v. Shannon S., 20 N.Y.3d 99, 109-110 (2012) (Smith, J., dissenting), quoting from Kansas v. Crane, 534 U.S. 407, 413 (2002) (“[U]nless ‘mental abnormality’ is defined with scientific rigor, [sexual dangerousness] statutes could become a license to lock up indefinitely, without invoking the cumbersome procedures of the criminal law, every sex offender a judge or jury thinks likely to offend again[; such statutes] must be limited to people who can be shown by scientifically valid criteria to have a ‘serious mental illness, abnormality, or disorder’ — one that distinguishes them ‘from the dangerous but typical recidivist convicted in an ordinary criminal case’ ”).

The Commonwealth suggests that the QE’s gatekeeping role effectively precludes a trial judge from scrutinizing the sufficiency of the evidence. In my view, the extraordinary context of preventative detention demands that judges continue to play such a role. Moreover, as this case well illustrates, in light of how the SDP scheme is structured, relying on juries to weed out unmeritorious SDP cases goes only so far. Although the Commonwealth was unable at trial to convince the requisite number of jurors to find that Souza remains an SDP, he now — over five years after his G. L. c. 123A, § 9, petition was filed — again faces the prospect of indefinite confinement. After retrial, he could be confined even in the absence of a jury finding that he currently is an SDP so long as a sufficient number of jurors held out for such a finding. This presents serious cause for concern, especially given that the underlying subject area is one that is “ruled by emotions.” Commonwealth v. Sullivan, 82 Mass. App. Ct. 293, 319 (2012) (Milkey, J., dissenting).

Commonwealth v. Reese , 438 Mass, at 520, was an appeal from a judge’s finding of no probable cause after a hearing under G. L. c. 123A, § 12(c). The *177 Supreme Judicial Court explained that at least in that context, the Commonwealth’s expert could rely on clinical observations and experience independent of the DSM criteria to make a diagnosis of pedophilia. Id. at 525-526. Reese thus involved a situation in which the Commonwealth’s expert explained that he was not resting his diagnosis on the DSM-IV. Reese does not say that where an expert relies on the DSM-IV at trial, the Commonwealth is excused from producing evidence that the DSM-IV criteria have been met.

That is hardly surprising given that, as Judge Smith of the New York Court of Appeals trenchantly has observed in writing for a three-judge dissent, “the idea that a man’s mere attraction to pubescent females is abnormal is absurd.” Matter of State of N.Y. v. Shannon S., 20 N.Y.3d 99, 111 (2012) (Smith, J., dissenting).

I recognize that lay jurors presumably would consider Souza a “pedophile” within the far broader everyday use of that term. But that underscores the constitutional concerns raised by allowing experts to untether their opinions from the stricter definitions accepted by the medical community as to what constitutes a “mental abnormality.”

See also Patterson v. Liberty Mut. Ins. Co., 48 Mass. App. Ct. 586, 592-593 (2000), and cases cited (an expert’s opinion must be “based solely on the expert’s ‘direct personal knowledge’ or admissible evidence in the record and not on assumptions that are not established by such evidence”).

This is not a case where the historical pattern of sex offenses itself demonstrated that the offender must have suffered from such compulsions.

Obviously, opportunities for sexual misbehavior may be more limited for someone who is confined, but they are hardly absent. Compare Commonwealth v. Husband, 82 Mass. App. Ct. at 2 (noting a sex offender’s disciplinary record while incarcerated, in which “[h]is reported conduct toward prison female medical personnel included sexual epithets, insults, taunts, threats, exposure, and masturbation”). Moreover, as the evidence in this case revealed, sex offenders who target children sometimes exhibit sexually inappropriate behavior in confinement, such as hoarding pictures of children. There was even testimony about a pornography ring operating inside the treatment center; Souza was not implicated in any such activity.

Kelso discounted the results of the penile plethysmograph (PPG) test, even while acknowledging that respected empirical researchers had concluded that the best predictor of recidivism was sexual deviancy, as measured by PPG tests or other means. This is not to say that the reliability of PPGs has been established, and one of Souza’s own experts stated that he does not put much stock in such tests. However, the fact remains that the one test that the treatment *180 center itself administered to Souza to measure his response to sexual stimuli provided no evidence to support the Commonwealth’s case and, if anything, undercut that case.

The uncontested facts belie any suggestion that Souza has accepted no responsibility for his two sex offenses. Indeed, Souza pleaded guilty to both offenses. In addition, even though his postplea accounts of the 1971 offense have varied somewhat, he has regularly admitted that he had intercourse with the 1971 victim while she was underage and that what he did was wrong. Granted, although Souza pleaded guilty to having indecently touched the 1990 victim, he denied sexually assaulting the boy in his postplea accounts. Souza was also indicted of rape of a child, something he consistently denied. The Commonwealth nol pressed the rape charge (after Souza admitted that he touched the boy’s penis), and it made no independent effort to substantiate that Souza had committed a rape. Nevertheless, the majority goes out of its way to highlight salacious details underlying the rape allegations even though the Commonwealth itself appropriately avoided the issue.

I fully appreciate that the Legislature has made the opinions of QEs admissible in SDP trials regardless of whether they have been demonstrated to be reliable, and that this situation-specific modification of the rules of evidence has been upheld. See Commonwealth v. Markvart, 437 Mass. 331, 339 (2002), citing G. L. c. 123A, § 14(c). However, especially in light of the overlaying constitutional concerns that are implicated, I do not interpret such precedent as barring any judicial inquiry into whether the opinion of the QE enjoys a demonstrated medical foundation. That inquiry need not embroil a trial judge in making credibility determinations or “weighing” the evidence.

The experts’ stance on the need for treatment is better understood as a policy position than as evidentiary proof. That the experts would adopt such a position is consistent with the institutional roles that each played. Kelso was an employee of the private contractor that provided sex offender services at the treatment center, and Tomich was the director of forensic psychological services at the Department of Correction.

Kelso, the Commonwealth’s lead expert, acknowledged that a preeminent empirical study demonstrated only a minor correlation between treatment and recidivism. Again, the existence of that study did not preclude the Commonwealth from proving that Souza’s failure to complete a treatment program mattered, but, again, the Commonwealth offered no empirical studies or evidence of medical consensus to substantiate its position.

It is undisputed that Souza is of borderline intelligence, with an IQ measured between sixty-eight and seventy-one. Treatment center records show that he is able to read at a third-grade level. Kelso acknowledged that Souza’s cognitive limitations presented potential obstacles to his succeeding in the treatment classes made available to him, and Tomich acknowledged that Souza’s cognitive limitations meant that “it may take him longer to benefit from treatment.” There was evidence that programs tailored for people with Souza’s limitations were “sometimes offered” at the treatment center, that at least one treatment component was modified to address those limitations, and that he was able to pass that one (and a “few” classes overall).

Kelso was able to score Souza that high only by crediting him with six 1971 sex crimes, even though five of the six New York charges were dropped, and there was no independent evidence presented in this trial that Souza had committed those crimes.

Because I consider a retrial unwarranted, I would not reach the Commonwealth’s claim that the jury instructions were erroneous. I state no view on the merits of that issue except to note that while I agree with the majority that a narrow reading of Johnstone, Petitioner, 453 Mass. 544, 553 (2009), does not compel the instruction that the trial judge gave, that instruction does find some support in the reasoning on which Johnstone is based. Clarification from the Supreme Judicial Court on this point of law would be beneficial.