Petitioner appeals from the order and judgment (one paper) of Supreme Court, Bronx County (David L. Lewis, J.), entered April 4, 2023, which determined, after a nonjury trial, that the State of New York failed to sustain its burden of establishing that respondent suffers from a mental abnormality under Mental Hygiene Law article 10 and dismissed its petition. ANIL SINGH, JUSTICE At issue on this appeal is whether Supreme Court’s finding that the State failed to prove that respondent suffers from a mental abnormality is tainted by error requiring a new trial. We find that it is. First, Supreme Court incorrectly conflated the proper legal standards in its assessment of whether the State proved that respondent has a mental abnormality as defined in article 10 of the Mental Hygiene Law. Second, Supreme Court erred in declining to follow binding precedent establishing that a combination of disorders may support a finding of a mental abnormality. Third, Supreme Court improperly analyzed and relied upon extra-record scientific and psychological research on several important issues, including the source of expert bias, the effect of opioid use on sexual offending, and diagnostic criteria, without notifying the parties. We agree with the dissent that, absent serious error, trial court determinations should be accorded great deference. Here, however, the fundamental nature of the errors committed by Supreme Court constitutes reversible error. Factual and Procedural Background This article 10 proceeding arose out of respondent Richard V.’s 2002 conviction of rape in the first degree. In October 2001, respondent and an accomplice posed as plumbers to gain entry to the apartment of a female acquaintance. After the woman brought them inside, respondent subdued her with pepper spray, restrained her, repeatedly attacked her, threatened to kill her, and twice violently raped her. Respondent’s extensive criminal history began at approximately age 12, when he shot and killed a man whom he believed had harmed his mother. He was confined to a juvenile detention facility in Puerto Rico, where he developed an opioid addiction that followed him into adulthood. Following his release as a young teenager, respondent lived with his uncle, who abused him and frequently brought him to the brothel and bar where he worked, allowing respondent to witness and participate in various criminal activities. Respondent was arrested several times for various nonsexual offenses beginning at age 17, including criminal possession of a weapon, assault, and criminal sale of a controlled substance. Respondent’s sexual offending began at age 35. Prior to the qualifying offense, respondent had sexually assaulted three other women, all of whom were acquainted with him. In February 1998, while he was on parole for an unrelated drug offense, respondent threatened his then-girlfriend with a knife and forcibly raped and sodomized her. He was arrested and charged with rape in the first degree, sodomy in the first degree, unlawful imprisonment, and menacing. In December 1999, respondent forcibly raped another female acquaintance on three separate occasions and pleaded guilty to one count of sexual misconduct, for which he was incarcerated for one year. Further, in January 2000, respondent used a knife to threaten a woman that he had met the day before at a social services office, forcibly raped her, and stole $200 in cash from her. He pleaded guilty to one count of sexual misconduct and was incarcerated for one year, which ran concurrently with the sentence for the 1999 offense. Following his conviction for the qualifying offense, respondent was incarcerated until 2013. He amassed 38 disciplinary tickets during that time, although none were for sexual misconduct. In 2013, the State filed its first article 10 petition against respondent. Summary judgment was granted for respondent and the case was dismissed. Respondent was incarcerated for an additional 12 months after pleading guilty to a charge of refusing to obey a direct order. Respondent briefly participated in sex offender treatment while incarcerated in 2011 but was unable to complete the program due to sanctions related to his disciplinary tickets. The State filed a second article 10 petition in 2016; however, upon respondent’s conviction of attempted assault upon a corrections officer, which carried an additional 2 to 4 years’ incarceration, the petition was dismissed without prejudice. The State brought this petition in 2021. Supreme Court’s Decision Following a determination that there was probable cause to believe that respondent is a detained sex offender requiring civil management under article 10, a bench trial was held in Supreme Court, Bronx County (Lewis, J.) in December 2022. Supreme Court dismissed the State’s petition, finding that it failed to prove by clear and convincing evidence that respondent suffers from a mental abnormality under article 10 that predisposes him to commit sexual offenses. In a 185-page decision, Supreme Court discussed the credibility of each expert at length. The opinion relied on scientific research outside the trial record as a basis to reject the testimony of the State’s expert witnesses and made numerous psychological findings that were not included in any of the experts’ testimony. Psychologists Dr. Susan Cox and Dr. Stuart Kirschner both testified on behalf of the State. Dr. Cox is employed with the New York State Office of Mental Health (OMH) and has conducted more than 150 article 10 evaluations for the State. She has found a mental abnormality in approximately 50 percent of those individuals. Dr. Kirschner is a private clinical psychologist who has performed approximately 300 article 10 evaluations for the State. He found a mental abnormality in nearly all of them. Respondent declined both experts’ requests for an interview; therefore, their evaluations were based solely on respondent’s records. The psychologists utilized the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), in their evaluations. Following their review of the records, both State experts diagnosed respondent with antisocial personality disorder (ASPD) and opioid use disorder. Dr. Cox additionally diagnosed respondent with unspecified depressive disorder and psychopathy, and Dr. Kirschner diagnosed him with borderline personality disorder (BPD) and unspecified bipolar disorder. Both experts concluded that respondent suffers from a mental abnormality based on the combination of their respective diagnoses. Forensic psychologist Dr. Leonard Bard testified on behalf of respondent. He has conducted hundreds of evaluations of sexually dangerous persons for defense attorneys and state and federal governments. In addition to viewing respondent’s records, Dr. Bard interviewed respondent in January 2022 via video teleconference and had previously interviewed him in 2014 and 2017, all in connection with article 10 proceedings. Following the 2022 interview, Dr. Bard diagnosed respondent with bipolar disorder and opioid use disorder. He also utilized the DSM in his assessment. Dr. Bard concluded with a reasonable degree of professional certainty that respondent does not suffer from an article 10 mental abnormality. The court considered the State experts’ testimony and concluded that their opinions and diagnoses were less credible than those of Dr. Bard. For a diagnosis of ASPD, Dr. Cox testified that an individual must display evidence of a conduct disorder prior to the age of 15, and then three or more of seven additional criteria must be met as an adult. Key to the court’s analysis regarding the State experts’ diagnosis of ASPD was Dr. Bard’s testimony that respondent’s bipolar disorder and the manic episodes that resulted from a lack of effective treatment better explained his erratic and violent behavior while incarcerated than ASPD would. The court also found no basis for the State experts’ conclusion that respondent exhibits a lack of remorse. Regarding their finding of a conduct disorder prior to the age of 15, the court opined that the State experts gave “short shrift” to respondent’s Puerto Rican culture. When Dr. Kirschner referenced respondent’s writings as to the events surrounding the homicide of the man who harmed his mother, he testified that respondent wrote that the incident “was the beginning of giving him the feeling that he was not a person to be violated, that it gave him a feeling of being macho and to a certain degree fearless.” The court criticized Dr. Kirschner’s use of the word “macho” and failure to perceive it as culturally relevant to respondent’s behavior, rejected his finding that respondent’s behavior as an adolescent was not tied to his culture, and determined that Dr. Kirschner was “not a reliable and credible witness or a credible diagnostician.” The court additionally criticized the State experts for not adequately considering that respondent’s behavior has changed over time and that individuals with ASPD often age out of the diagnosis. As to respondent’s opioid use disorder, which is the only diagnosis that all three experts agreed on, the court discredited Dr. Cox’s testimony that respondent’s “offense supportive beliefs” — namely, that respondent admitted he purposely targeted women who were drug users because he lacked respect for them — were aggravated by his own drug use. Dr. Bard testified that a substance use disorder does not predispose someone to commit sexual offenses but agreed that opiates can have a disinhibitory effect and that respondent’s drug use has “significantly” impacted his life and behavior. The court concluded that respondent’s substance abuse, taken alone, was an insufficient basis upon which a mental abnormality could be found, finding that respondent’s drug use did not fuel his sex offending, as not all the offenses occurred while he was high; rather, “[t]here were long periods of drug use without sexual offense,” and he did not offend during his “alleged and admitted” Suboxone use in 2022. The court took particular issue with Dr. Cox’s diagnosis of psychopathy. Dr. Cox testified that although this diagnosis is not included in the DSM-5, it is “generally accepted as a condition” and is evaluated using the Psychopathy Checklist-Revised (PCL-R). The PCL-R contains 20 areas that are assessed for a possible total of 40 points. She gave respondent high scores in several areas of the PCL-R — including pathological lying, callous disregard, and juvenile delinquency — and assigned him a total score of 30. Respondent had been assessed using the PCL-R twice before, with the examiners scoring him at 22 and 28, respectively. The court concluded that although the PCL-R is “generally accepted in New York” under Frye v. United States (293 F 1013 [DC Cir 1923]), respondent’s “shifting score based on the different evaluators…raises validity questions.” Dr. Cox’s report, as well as her testimony, failed to note what specifically in respondent’s records led her to score him at 30. The court was “unwilling to accept her diagnosis without a foundation,” and credited Dr. Bard’s testimony that the PCL-R is an unreliable instrument, especially with an eight-point deviation between the scores assigned to respondent by the three evaluators. Regarding Dr. Cox’s diagnosis of unspecified depressive disorder, the court noted that because she was unable to interview respondent “specifically on his depressive symptoms” and perceived that “there was no evidence of manic episodes besides his own self-reporting,” she could not be more specific than diagnosing him with some kind of depressive disorder. Dr. Bard, however, identified manic episodes that Dr. Cox did not perceive as manic, resulting in a clear diagnosis of bipolar disorder. The opinion further criticized Dr. Cox’s lack of acknowledgment of respondent’s expressions of remorse, her minimalization of his participation in sex offender treatment, and her “failure of memory of otherwise significant facts,” such as that respondent was sexually abused as a child, that he believed his family was sending him invisible coded messages, and that he attempted suicide while confined in 2016. As a result, the court found that Dr. Cox was “not credible or reliable.” Critically, the court then went outside the trial record and undertook an examination of “bias in forensic psychological settings.” The court found that Dr. Kirschner “demonstrated such bias as discussed in the relevant literature.” The court concluded that Dr. Kirschner’s credibility was undermined by his testimony that he has found a mental abnormality in nearly 100 percent of individuals that he has examined on behalf of the State. The court also condemned Dr. Kirschner’s diagnosis of BPD, as he failed to comment on the fact that respondent was treated with anti-psychotic medication and mood stabilizers and found no evidence of mania, unlike Dr. Bard. Dr. Kirschner’s failure to examine Dr. Bard’s records, as well as lack of remembrance of relevant facts, also made him “unreliable.” Based on the foregoing, the court rejected Dr. Kirschner’s testimony in its entirety. The court credited Dr. Bard’s diagnosis of bipolar disorder above the State experts’ diagnoses of unspecified depressive disorder, unspecified bipolar disorder, ASPD, BPD, and psychopathy, despite also critiquing him for exhibiting examiner bias.1 The court noted that this diagnosis was “consistent with the one assessed by mental health professionals in Puerto Rico” while respondent was detained at the juvenile facility. As Dr. Bard was the only expert to interview respondent, the court concluded that his findings that respondent did not exhibit “enduring and persistent” behavior associated with a personality disorder, that his erratic and violent behavior in 2016 could be attributed to manic episodes as a result of untreated bipolar disorder, and that his criminal acts as a child were reactions to his mother being harmed and suffering from sexual abuse were the most credible of the three experts. The court determined that respondent’s diagnoses do not cause him serious difficulty in controlling his sex offending. Opining that article 10 is “aimed at aberrational sex offenders,” the court did not find any “testimony of sexual arousal or gratification of [respondent's] illicit desires” or “evidence that he has any urges regarding sexual activity at all.” The court disagreed with the State experts that respondent’s behavior met the DSM criteria for ASPD and BPD because “the evidence fails to represent anything that is [] enduring and inflexible over a broad range of social situations and similar ranges.” The court also noted that respondent has not committed any sexual offenses while incarcerated, although it recognized that respondent’s sex crimes were “horrific” and that he has made “limited progress in sex offender treatment.” In evaluating the State experts’ testimony that the combination of their diagnoses predisposed respondent to commit sexual offenses, the court stated that “article 10 should be narrowly interpreted” and “the mental abnormality must be one that prevents sexual impulse control.” As both the State experts “agreed that no one of the diagnoses demonstrates a predisposition to commit sex offenses,” the court posited that “it follows logically that should any one of the diagnoses found not to be credible or validly diagnosed…the entire structure crumbles.” The court therefore found that the State failed to establish a nexus between respondent’s diagnoses and his sexual offending and additionally concluded that “[the] theory of combined impact of the diagnoses when no one of them was sufficient to establish a sex offender’s disposition to commit sex offenses has no support in the scientific community.” Acknowledging that its decision “may well serve to release a potentially dangerous offender into the community…where he may be at significant risk to commit another crime” should he begin using drugs again or decide not to take his medications, the court concluded that civil confinement would be “constitutionally intolerable” where a respondent does not “suffer from a lack of ability to control themselves regarding the potential commission of sex offenses.” Accordingly, Supreme Court ordered that respondent be released into the community because he does not suffer from a mental abnormality. Discussion “Under article 10 of the Mental Hygiene Law, the State must establish at trial, by clear and convincing evidence, that a detained sex offender suffers from a mental abnormality as defined in that statute” in order for that individual to be civilly confined upon their release from imprisonment (Matter of State of New York v. Floyd Y., 22 NY3d 95, 99 [2013]; see Mental Hygiene Law §10.07[d]). A mental abnormality is “a congenital or acquired condition, disease or disorder that affects the emotional, cognitive, or volitional capacity of a person in a manner that predisposes him or her to the commission of conduct anyway constituting a sex offense and that results in that person having serious difficulty in controlling such conduct” (Mental Hygiene Law §10.03[i]; see id. §10.07[d]; Floyd Y., 22 NY3d at 99). Such conditions are not limited solely to sexual disorders (see Matter of State of New York v. Dennis K., 27 NY3d 718, 743 [2016], cert denied 580 US &mdash, 137 S Ct 574 [2016]). Supreme Court Conflated the Applicable Standards for Article 10 Proceedings The sole issue at the bench trial was whether respondent suffers from a mental abnormality that “predisposes [him] to the commission of conduct constituting a sex offense” resulting in “having serious difficulty [] controlling such conduct” (Mental Hygiene Law §10.03[i] [emphasis added]). At the second stage of an article 10 proceeding — the dispositional phase — the standard is whether a respondent has “such an inability to control his behavior that [he] is likely to be a danger to others and to commit sex offenses if not confined to a secure treatment facility” (Mental Hygiene Law §10.03[e] [emphasis added]). The State argues that the court applied the heightened “inability to control” standard, or, at a minimum, conflated the two standards beyond the point at which we could ascertain which one it applied. There can be little dispute that Supreme Court conflated the applicable legal standards. At the very beginning of its extensive opinion, Supreme Court misstates the issues before it and appears to include the heightened “inability” standard.2 The court states that the issue being decided is whether respondent has “such an inability to control his behavior that [he] is likely to be a danger to others and commit sex offenses if not confined to a secure facility.” Throughout its opinion, Supreme Court alternated intermittently between the two standards. In the body of the opinion, the court reiterated that the issue before it “is whether the diagnoses mean that Respondent is unable to control his compulsions to re-offend” and that ” [t]he density of the sexual offenses [does] not suggest that Respondent is presently unable to control himself” regarding his sex offending (emphasis added). The court concluded that the State “ failed to meet its burden of proving, by clear and convincing evidence, that [respondent] is ‘presently unable to control his sexual conduct’ and is thus a dangerous sex offender requiring confinement” (emphasis added). Further, in evaluating the experts’ testimony, the court repeatedly held Dr. Cox and Dr. Kirschner to the standard of proving that respondent’s diagnoses rendered him unable to control his sex offending (“Dr. Cox failed to explain how the purported loss of control, leading to a fight or two, is proof of an inability to control [respondent's] sexual urges”; “The evidenced [sic] when viewed in the light most favorable to the State requires more than Dr. Cox or Dr. Kirschner’s diagnoses. A mere tendency to engage in socially undesirable behavior…is quintessentially insufficient to establish inability where there is an absence of sexually inappropriate conduct while confined”; “[T]he combinations of the diagnoses do not fulfill the rubric of an inability to control [respondent's] sexual behavior.” The court’s credibility determinations were thus inappropriately premised on the State’s experts’ failure to demonstrate that respondent has an inability to control his sexual offending such that he requires civil confinement. Although Supreme Court ultimately set forth the correct “serious difficulty” standard in its conclusion, we find that Supreme Court committed reversible error by conflating the “serious difficulty” standard or the “inability to control” standard in finding that the State failed to prove that respondent had a mental abnormality. Supreme Court Improperly Held that the State Cannot Rely on a Combination of Disorders to Conclude that Respondent Has a Mental Abnormality Binding precedent establishes that an expert may rely on a combination of disorders to establish a mental abnormality when no individual diagnosis would do so (see Dennis K., 27 NY3d at 751-752; see also Matter of State of New York v. David D., 206 AD3d 481, 485 [1st Dept 2022]). In circumstances “where a State expert diagnoses an article 10 respondent with multiple mental disorders, a reviewing court does not view each in isolation. Rather, “the court assesses whether and how those disorders, in combination, predisposed respondent to the commission of conduct constituting sex offenses” (David D., 206 AD3d at 485, quoting Dennis K., 27 NY3d at 750-752 [internal quotation marks omitted]). We disagree with the dissent’s conclusion that Supreme Court did not reject this combined diagnosis principle. The dissent argues that Supreme Court’s repeated citations to binding decisions, including Dennis K. and David D., “acknowledged the holdings of these cases as controlling authority.” While the court cited this line of case law, it nevertheless rejected the Dennis K. principle outright. The record is devoid of any expert testimony supporting Supreme Court’s holding that there is “no support in the scientific community” for permitting a forensic psychologist to rely on multiple diagnoses to conclude that a respondent has a mental abnormality. More importantly, “mental abnormality” is a legal concept rather than a medical one (see Matter of State of New York v. Shannon S., 20 NY3d 99, 106 [2012], cert denied 568 US 1216 [2013]), and the Court of Appeals has found that it may consist of a combination of diagnoses (see Dennis K., 27 NY3d at 750-752). Thus, in our view Supreme Court committed reversible error in finding that the State could not use a “constellation” of conditions, diseases, and disorders to establish that respondent has a mental abnormality. Supreme Court Improperly Relied on Research Outside the Record We agree with our dissenting colleague that deference is typically accorded to the trial court’s assessment of expert witnesses (Matter of Kelly, 265 AD2d 154, 154 [1st Dept 1999]) and that a judgment in a bench trial should be set aside only where it cannot be supported by any fair interpretation of the evidence (Wong v. Hsia Chao Yu, 160 AD3d 549, 549-550 [1st Dept 2018]) adduced at trial. Nevertheless, such a determination cannot be upheld when there is a reasonable possibility that the court could have reached a different conclusion had it not independently reviewed scientific literature and relied on facts not in evidence (see Floyd Y., 22 NY3d at 110; see also People v. Reynolds, 90 AD3d 956, 957 [2d Dept 2011]). Across a wide swath of its lengthy decision, Supreme Court cited to a multitude of facts outside the record that can be reasonably found to have influenced its findings. Accordingly, Supreme Court’s findings are not entitled to the deference typically accorded to such determinations. First, the court cited several pages of external research regarding Suboxone, the opiate that respondent admitted to obtaining and using during his confinement in 2022. Among the information included — both cited and not — was a study regarding the effects of “Suboxone and its chemical composition” on sexual function. The court interpreted this study as showing that a loss of interest in sex was “almost universal” among men who use Suboxone. The court clearly used this study to discredit both Dr. Cox’s and Dr. Kirschner’s testimonies that respondent’s opioid use disorder could fuel his sex offending, stating that “the chemical makeup [of Suboxone] and consequences of its use indicate that the user will likely be unable to behave in a way so as to commit sexual crimes.” While Dr. Bard’s professional opinion is that opioid use disorder would not predispose an individual to commit sexual offenses, none of the three experts testified about the effects of Suboxone on sexual function or sexual urges. Rather, all agreed that opioids and other drugs can have a disinhibitory effect on the user, and those addicted to substances often orient their behavior around obtaining additional drugs at the expense of those around them. Second, the opinion referenced numerous portions of the DSM-5 and the PCL-R manual that were not admitted into evidence.3 Although the Court of Appeals and this Court have referred and cited to the DSM, the degree to which the court analyzed its contents — and those of the PCL-R manual — beyond the experts’ testimony was error (compare Dennis K., 27 NY3d at 725 n 1, 727 n 2, 738 n 3 [quoting definitions of certain disorders to provide additional context]; Floyd Y., 135 AD3d at 73, 75-76 [providing parallel citations to the DSM where trial testimony independently established the same facts]; Matter of State v. Nicholas T., 55 Misc 3d 884, 892 [Sup Ct, NY County 2017] [referencing the DSM's "cautionary statement" to explain that the DSM was developed to meet the needs of mental health professionals rather than those of legal professionals]). Here, Supreme Court opined that “examination of the text of the DSM presents fuller [sic] picture and demonstrates that the prime definitional battle between sides as to whether [a] disorder needs to be currently manifested is a semantic difference, not an actual difference” and relied on its own interpretation of the DSM criteria, particularly regarding the diagnosis of ASPD. The court concluded that a diagnosis of ASPD requires a “current” pattern of relevant behavior within the past “six or twelve months” and concluded that there was no evidence that respondent “met the requisite number of criteria” in 2022. Dr. Bard, however, testified that he considered an ASPD diagnosis to involve displays of antisocial behavior “in the past year or two.” He additionally noted that the DSM does not specify a timeframe when assessing an individual’s behavior for ASPD, but that such behavior should be “enduring,” “pervasive,” and “ongoing,” which he considered in his professional opinion to be interchangeable with “current.” None of the experts testified that an individual must exhibit antisocial behavior within the past “six or twelve months.” Several pages of the opinion also assessed the experts’ testimony in comparison to the DSM’s criteria for ASPD. Similarly, contrary to the dissent’s characterization, the court’s references to the PCL-R manual far exceeded the scope of the expert testimony regarding the diagnosis of psychopathy. The court opined that “[t]here must be a combination of psychopathy with deviant sexual arousal to be predictive of sexual recidivism,” which none of the experts testified to. Finally, while the foregoing is sufficient to find reversible error, the court additionally spent several pages discussing extra-record literature and research as to expert bias.4 The court clearly relied upon these studies in determining that Dr. Kirschner’s testimony should be disregarded in its entirety, stating that “[a] survey of the issue of bias in forensic psychological settings reveals that nonintentional unconscious bias is commonplace and skews analysis. Dr. Kirschner demonstrated such bias as is discussed in the relevant literature” (emphasis added). The opinion further stated that “[t]he relevant literature agrees that legal or forensic psychologists are prone to cognitive biases, such as being adversely affected by irrelevant contextual information, confirmation bias, and allegiance bias.” The court then finally proceeded to discuss its analysis of Dr. Kirschner’s actual testimony. The dissent argues that the court did not draw any conclusions from the cited studies because it acknowledged that “it is unknown whether legal psychologists also evince a bias blind spot” and that “empirical research on this issue is absent.” We disagree. The opinion clearly and consistently conflated both facts inside and outside the record and plainly compared the perceived bias exhibited by Dr. Kirschner to the cited surveys. Next, the dissent is correct that both Dr. Cox and Dr. Bard commented on adversarial bias during their testimony; however, neither expert cited to any specific studies, and the references to such bias are brief, general, and scattered at best. More importantly, this testimony did not open the door to Supreme Court to seek out its own scientific research and then extensively opine on adversarial bias in forensic psychologists, including the terminology involved and conclusions drawn from these studies, which are entirely untethered from the trial record. Supreme Court’s extensive usage of outside research blurred the lines between the roles of judge and counsel, depriving the parties of the opportunity to respond (see generally People v. Arnold, 98 NY2d 63, 67-68 [2002]). We disagree with our dissenting colleague that such usage constituted harmless error (see id. at 69 [remanding for new trial where trial court improperly introduced evidence during bench trial, depriving a party of the opportunity to advance an argument, and there was a significant probability that the error affected the verdict]; Orlich v. Helm Bros., 160 AD2d 135, 141 [1st Dept 1990] ["We modify…to express our disapproval of the court's usurpation of the advocate's role in delineating a theory of liability neither advanced by the parties nor supported by the record"]; see also Matter of Justice v. King, 60 AD3d 1452, 1453 [4th Dept 2009], cert denied 558 US 994 [2009]; cf. People v. Smith, 18 NY3d 544, 552 [2012] [error was not harmless where the trial court "clearly relied on" wrongly admitted evidence in a bench trial]). “The overarching principle restraining the court’s discretion is that it is the function of the judge to protect the record at trial, not to make it” (Arnold, 98 NY2d at 67). Whether the cited sources in the opinion are accurate or not, information gleaned from outside the record “that does not qualify for judicial notice and is not disclosed to the parties is untested by the adversary process” (American Bar Association Committee on Ethics and Professional Responsibility Formal Opinion 478 [2017], https://www.abajournal.com/images/main_images/FO_478_FINAL_12_07_17.pdf [accessed April 4, 2024]). “The key inquiry here is whether the information…is of factual consequence in determining the case. If it is, it must be subject to testing” by the parties (id. at 6). Should additional information be necessary to reach a decision in a particular case, such information should be provided by the parties or be subject to judicial notice (id.). Further, if the purpose of a court’s inquiry is to “corroborate facts, discredit facts, or fill a factual gap in the record” and such facts are adjudicative, the inquiry is improper (id.). Here, the extraneous research was “not only inappropriate, but, in any event, unnecessary to a proper disposition” of this case (Orlich, 160 AD2d at 141-142 [modifying order to further limit discovery where the IAS court "relied on materials and factual assumptions not found anywhere in the record" to support its finding]). Despite the court’s assurances that its conclusions “were not those of a clinician” and that it was not “substitut[ing] its own psychological judgments for that of the three testifying doctors,” there is ample evidence demonstrating that its citation of outside research was not “at most tangential” but rather was at the core of the court’s credibility determination rejecting the testimony of the State’s witnesses. Conclusion In sum, the fundamental legal errors committed here, including the conflation of legal standards, rejection of binding precedent, and — perhaps most critically — the analysis and reliance on scientific research and psychological opinions unmoored from the trial record constitute reversible error. Supreme Court improperly based its decision on evidence not presented in open court and available to all parties and thus strayed beyond the boundaries of its role as a neutral fact finder. Accordingly, the order and judgment (one paper) of Supreme Court, Bronx County (David L. Lewis, J.), entered April 4, 2023, which determined, after a nonjury trial, that petitioner State of New York failed to sustain its burden of establishing that respondent suffers from a mental abnormality under Mental Hygiene Law article 10 and dismissed the petition should be reversed, on the law, without costs, the petition reinstated, and the matter remanded for a new trial on mental abnormality under article 10. The appeal from the order, same court and Justice, entered on or about March 30, 2023, is dismissed, without costs, as subsumed in the appeal from the aforesaid order and judgment. All concur except Gesmer J., who dissents in an Opinion. GESMER, J. (dissenting) I respectfully dissent and would affirm the decision of the trial court, as we are required to do in the absence of serious error. “‘In reviewing a judgment from a bench trial, especially where credibility played an important role, the judgment should only be set aside where it is not supported by any fair interpretation of the evidence’…Because the trial court was in the best position to evaluate the credibility of the witnesses, its determination is entitled to great deference” (Wong v. Hsia Chao Yu, 160 AD3d 549, 549-550 [1st Dept 2018], quoting Abreu v. Barkin and Assoc. Realty, Inc., 115 AD3d 624, 624 [1st Dept 2014]; see also Moore v. Beautiful Spaces, LLC, 192 AD3d 591, 591 [1st Dept 2021]; Nightingale Rest. Corp. v. Shak Food Corp., 155 AD2d 297, 297-298 [1st Dept 1989], lv denied 76 NY2d 702 [1990]). It is particularly incumbent on us to follow this principle in a case like this one where the trial judge is called upon to “evaluate the weight and credibility of the often conflicting testimony of the medical and psychiatric experts” (Matter of Kelly, 265 AD2d 154, 154 [1st Dept 1999], quoting Matter of George L., 85 NY2d 295, 305 [1995]). Here, petitioner presented two expert witnesses who disagreed with each other in major respects as to their diagnosis of respondent, even though they each concluded that the elements of their own diagnosis, taken together, caused respondent to suffer from a “mental abnormality” as defined by Mental Hygiene Law §10.03[i]. Respondent presented his own expert witness who presented a third set of diagnoses and concluded that respondent does not suffer from a mental abnormality. The trial court determined that respondent’s witness’s conclusions were more credible and that the State failed to meet its burden to prove by clear and convincing evidence that respondent suffers from a mental abnormality (id.; Matter of State of New York v. Dennis K., 27 NY3d 718, 752 [2016], cert denied 580 US 1102; Matter of State of New York v. Donald DD., 24 NY3d 174, 187 [2014]). For the reasons discussed more fully below, it is my view that the trial decision is supported by a fair interpretation of the evidence, and therefore we should not disturb it. The majority declines to defer to the trial court because it concludes that the trial court failed to apply the correct legal standard, relied on its own research, and rejected legal precedent. I disagree. First, as the majority concedes, the trial court’s conclusion was based on the correct legal standard.5 Second, I would find that the trial court’s citation to scientific and academic studies regarding Suboxone and inherent bias had no effect on its determination as to the credibility of the witnesses and was therefore harmless error. Similarly, I would find that the trial court’s references to the DSM-5 and the PCL-R manual were consistent with the experts’ testimony. The conclusions that the trial judge reached about the State’s experts’ lack of credibility were supported by the record, and he had strong independent reasons for crediting the testimony of Dr. Bard over the petitioner’s experts. Finally, I disagree that the trial court rejected the principle that a constellation of conditions may establish a mental abnormality (see Dennis K., 27 NY3d at 751-752; Matter of State of New York v. David D., 206 AD3d 481, 485 [1st Dept 2022]). Rather, it is my view that the trial court found that the State’s witnesses, who did not agree on respondent’s diagnosis, failed to prove that respondent suffered from a constellation of conditions which together established a mental abnormality. Accordingly, I respectfully dissent and would vote to affirm the order appealed from. Article 10 of the Mental Hygiene Law sets out conditions under which the State may civilly confine a sex offender who has completed his or her prison term, as respondent has. In order to do so, the State bears the burden to prove, by clear and convincing evidence, that respondent suffers from a “mental abnormality.” This requires proof that respondent suffers from a “condition, disease or disorder that affects [respondent's] emotional, cognitive, or volitional capacity…in a manner” that both predisposes him to commit sex offenses and causes him serious difficulty in controlling that conduct (Mental Hygiene Law §10.03[i]). Where there is conflicting expert testimony, the court is required to evaluate the accuracy and credibility of the testimony (see George L., 85 NY2d at 305; Griffin v. Cerabona, 103 AD3d 420, 421 [1st Dept 2013] ["resolution of issues of credibility of expert witnesses and the accuracy of their testimony are matters within the province of the [fact finder]“] [internal quotation marks omitted]). I disagree with the majority that the court’s discussion of portions of the DSM-5 and PCL-R manual which were not in evidence constituted reversible error.6 As the majority notes, courts can and do reference these manuals, both to provide context and to support trial testimony which is consistent with the language cited. Here, the trial court used these sources in both of these ways. An example of the trial court’s use of the DSM-5 for context is its discussion of the DSM-5′s “Cautionary Statement.” Other courts have discussed the Cautionary Statement because it explains that “that the definition of mental disorder included in the DSM—5 was developed to meet the needs of clinicians, public health professionals, and research investigators rather than all of the technical needs of the courts and legal professionals” (Matter of State v. Nicholas T., 55 Misc 3d 884, 892 [Sup Ct, NY County 2017]) and that courts should therefore be “informed by an awareness of the risks and limitations of its use in forensic settings…. because of the imperfect fit between the questions of ultimate concern to the law and the information contained in a clinical diagnosis” (Matter of State of New York v. Jerome A., 58 Misc 3d 1202[A], 2017 NY Slip Op 51762[U], *5 [Sup Ct, NY County 2017] [cited favorably by this Court on appeal from a later order concerning the same respondent, 172 AD3d 446, 447 (1st Dept 2019)]). Here, the trial court’s critical observation that the State’s experts’ testimony lacked context and thus “appear[ed] to use the DSM-5 criteria as a checklist” was consistent with the Cautionary Statement. Furthermore, stating the obvious fact that the DSM is designed by mental health professionals for use by mental health professionals, and not by lawyers and judges for use in making legal determinations hardly constitutes reversible error.7 Similarly, the trial court’s citation to the 20 categories for scoring the PCL-R was consistent with the testimony by the expert witnesses about the use of the PCL-R by mental health professionals to determine if a person exhibits psychopathy. The majority cites the trial court’s discussion of a study about the drug Suboxone as an improper use of research that was not in evidence. I agree that the study cited was not in evidence. However, I disagree that the trial court used the study to discredit the state’s experts’ testimony about the effect of respondent’s opioid use disorder on the likelihood that he might reoffend. To the contrary, the trial court noted that “Dr. Cox properly identified [respondent's] suboxone use as a very serious relapse issue.” The trial court considered this concern, noting that respondent “has no relapse plan designed to interrupt any possible sexual offense” and acknowledged that his opioid use disorder “may push him into criminal behavior.” Nevertheless, the trial court determined, for reasons having nothing to do with the Suboxone study cited, that the state failed to meet its burden. Accordingly, I would find that the court’s citation to the Suboxone study was harmless error. It is also true, as the majority points out, that the trial decision discusses studies not in evidence which discussed the issue of implicit bias in expert opinions. However, I do not agree that the court’s discussion of these studies caused any reversible error since two of the three trial experts testified at trial about this very issue. Dr. Cox testified, for example, that experts’ results on the test for psychopathy are skewed by what she described as the allegiance effect; that is, by their allegiance to the party who retained them. In addition, Dr. Bard testified that doctors have to be self-conscious about their biases when doing an evaluation. In its decision, the court discussed several articles about bias, but concluded that, although those studies showed that both lay people and experts can exhibit largely unconscious bias, “it is unknown whether legal psychologists also evince a bias blind spot…. [because] empirical research on this issue is absent.” Accordingly, the court did not draw any conclusions from the articles it cited. Moreover, although the trial court found Dr. Bard more credible, it was also critical of instances in which he too appeared to the court to demonstrate bias. This demonstrates that the court was critical of instances of apparent bias among evaluators generally but did not rely on any of the research it cited to make its credibility determinations. Thus, the studies that the trial judge discussed in his decision were both entirely consistent with the testimony at trial, and were at most tangential to his credibility determinations. Moreover, the majority argues that the trial court’s citation to segments of the DSM-5, PCL-R manual, and studies about Suboxone and examiner bias could have affected the trial judge’s ruling as to the credibility of the expert witnesses. However, the court made its determinations as to credibility based solely on its evaluation of the trial evidence. As the trial court noted, the State’s witnesses, Dr. Cox and Dr. Kirschner, did not agree with each other as to respondent’s diagnoses and how their diagnoses together caused respondent to suffer from a mental abnormality. Neither agreed with respondent’s expert, Dr. Bard, as to his overall assessment of respondent. All three experts diagnosed respondent with opioid use disorder, and all agreed this was an insufficient basis by itself for a finding of mental abnormality. The trial court thus appropriately found that the experts’ competing opinions required the court to make a credibility determination. Based on its analysis of their testimony, the trial court determined that Dr. Bard’s conclusions were more credible than those of the State’s experts, and stated ample reasons for making that finding, none of which relied on the court’s independent research. In weighing the evidence at trial, the court cited several bases for its determination that Dr. Bard’s conclusions were more credible than those of either Dr. Kirschner or Dr. Cox. None of these relied on scientific literature outside the record. First, the court found that the State’s experts failed to establish a nexus between their respective diagnoses and a predisposition to commit sex offenses or serious difficulty controlling those urges. The court found that their testimony in this regard was “conclusory” and lacked sufficient analysis of the factors leading to their conclusions to constitute the “detailed psychological portrait” necessary to meet the State’s burden (Donald DD., 24 NY3d at 188; see also Dennis K., 27 NY3d at 726; Matter of State of New York v. Howard H., 151 AD3d 648, 649 [1st Dept 2017]). Second, the court noted that neither of the State’s experts had the opportunity to interview respondent and both based their diagnoses solely on review of his records. In contrast, Dr. Bard had evaluated respondent on three separate occasions in 2014, 2017, and 2022, and had interviewed him each time. The court noted that Dr. Kirschner had not bothered to read any of Dr. Bard’s three evaluations of respondent, and did not state any explanation for his failure to do so. He thus deprived himself of the information that Dr. Bard obtained by interviewing respondent. Third, the court noted that both Dr. Cox and Dr. Kirschner either did not read or did not recall facts from respondent’s records that cast doubt on or were inconsistent with their conclusions. For example, Dr. Cox concluded that respondent did not participate in treatment while confined. However, on cross-examination, she acknowledged that she did not recall treatment notes indicating that respondent’s participation in groups was “excellent” and “very honest.” Similarly, both Dr. Cox and Dr. Kirschner cited respondent’s 2016 assault on a corrections officer as central to their diagnosis of antisocial personality disorder (ASPD) because they claimed his conduct in that incident was evidence of the aggressive, antisocial behavior typical of that disorder. However, neither recalled the records in evidence showing that, at the time of that incident, respondent was so delusional that corrections officers were concerned that he was going to commit suicide. Neither Dr. Cox or Dr. Kirschner testified that respondent’s delusions and suicidality were related in any way to ASPD. Dr. Bard attributed them to bipolar disorder. Thus, the testimony by Dr. Cox and Dr. Kirschner about the 2016 incident provides an example of their disregard of information which was contrary to their conclusions. Fourth, both Dr. Cox and Dr. Kirschner based their diagnosis of respondent with ASPD in part on his alleged lack of remorse. The trial court found that the State’s experts failed to identify the source of their claim that respondent lacked remorse; indeed, there was no support for their claim in any of the documents entered into evidence. On the other hand, the court credited Dr. Bard’s testimony that respondent has expressed remorse. In particular, the court noted that in respondent’s handwritten personal history, which he created in connection with his treatment program and entered into evidence, he acknowledged that he was “becoming a monster” and had “issues that needed to be d[e]alt with,” thus clearly demonstrating his remorse. Finally, the State’s experts each claimed that respondent suffered from a particular combination of diagnoses. In addition to opioid use disorder, Dr. Cox diagnosed respondent with ASPD, unspecified depressive disorder, and the condition of psychopathy; and Dr. Kirshner diagnosed him with ASPD, borderline personality disorder (BPD), and unspecified bipolar disorder.8 The court rejected the State’s experts’ diagnoses of ASPD, BPD, and the condition of psychopathy. It found that the evidence supported Dr. Bard’s diagnosis of bipolar disorder and the diagnosis by all of the experts of opioid use disorder.9 The trial court’s rejection of the diagnoses proffered by the state’s experts, discussed more fully below, provides the strongest basis for finding that he reached his conclusion that the state’s witnesses were not credible independent of any reference to research outside the record. Antisocial Personality Disorder The trial court rejected the State’s experts’ diagnosis of ASPD for three reasons. First, the court found that there was insufficient evidence to support their conclusion, necessary for such a diagnosis, that respondent suffered from a conduct disorder prior to the age of 15. All of the experts agreed that there were no records available from respondent’s adolescence. The court credited Dr. Bard’s testimony that respondent’s admission that he shot and killed a man when respondent was approximately 11 or 12 was not evidence of a conduct disorder. Dr. Bard testified that, when viewed in context, it was a terrible youthful response to respondent’s belief that the man had killed his mother and to the horrific abuse respondent had endured throughout his childhood. Similarly, the court credited Dr. Bard’s testimony that respondent’s childhood and adolescent participation in robberies and presence at bars or brothels where prostitution was occurring were not indications of a conduct disorder since he engaged in these activities under the influence of his adult uncle who functioned as his guardian. Second, the trial court found that the evidence did not support a finding that respondent currently exhibits a pattern of antisocial behavior. The court credited Dr. Bard’s testimony that respondent’s current behavior must show an antisocial pattern of behavior in order to justify a diagnosis of ASPD. Dr. Bard testified that the DSM-5 states that personality disorders involve an “enduring pattern [that] is inflexible and pervasive across a broad range of personal and social situations” and that is “stable over time.”10 The trial court further credited Dr. Bard’s testimony that the evidence of respondent’s possibly antisocial behavior since 2016 was insufficient for a diagnosis of ASPD because it was either too insignificant or infrequent to constitute an enduring pattern that is pervasive and stable over time, or it was caused by his bipolar disorder. Specifically, Dr. Bard concluded that respondent’s disagreement with another inmate in 2022 in which respondent threw a cup of coffee in the direction of the other inmate who then beat respondent up was too insignificant to constitute anti-social behavior. Second, he concluded that other cited events were too infrequent to be pervasive or stable over time. Finally, Dr. Bard determined that respondent’s 2016 assault on a corrections officer was a consequence of psychosis resulting from respondent’s bipolar illness. The court further noted that respondent’s sexual offending behavior occurred over a four-year period culminating in the qualifying offense over 20 years ago, that there was no evidence of such behavior prior to that, and none of respondent’s disciplinary sanctions while incarcerated afterward were for behavior that was sexual in nature. Third, the trial court credited Dr. Bard’s testimony that respondent’s bipolar disorder made it difficult to attribute any of respondent’s behaviors to ASPD. Consequently, the trial court properly did not give any weight to the diagnosis of ASPD given to respondent by Dr. Cox and Dr. Kirschner. Psychopathy Psychopathy is not a diagnosis in the DSM-5.11 The trial court rejected Dr. Cox’s opinion that respondent has the condition of psychopathy for two reasons. First, Dr. Cox testified that she had concluded that respondent has the condition of psychopathy based on her use of the PCL-R. However, while she testified that she scored respondent at 30 points, she failed to produce any of the data underlying her opinion and provided no evidence as to why she assigned a score of 30 to respondent. Second, the court noted that three different psychologists had used the PCL-R to evaluate respondent between 2013 and 2020. Each had given him different scores on the PCL-R: variously 22, 28 and 30. The trial court found that the range of scores called into question the reliability of the PCL-R in this case. The court noted that, while Dr. Cox testified that her higher score was a result of events occurring after the earlier evaluations, she failed to identify what events she considered or how she arrived at a score of 30. Third, the court concluded that the PCL-R is of questionable reliability and validity. The court noted that Dr. Cox had not interviewed respondent although the inventor of the PCL-R recommends that scoring be done through an interview of the subject, since it is difficult to assign an appropriate PCL-R score without direct observation. Moreover, Dr. Cox admitted that studies have shown that the PCL-R is less reliable in a forensic setting and that there is an “allegiance effect” among scorers dependent on who has retained them. The court credited Dr. Bard’s testimony to the effect that studies had found problems with both the reliability and the validity of the PCL-R. Dr. Bard testified that the PCL-R was developed for the prediction of non-sexual violence among a general prison population, and that there are no studies establishing that it is useful in predicting recidivism among sexual offenders.12 Consequently, the court had an adequate basis for declining to credit Dr. Cox’s conclusion that respondent suffered from psychopathy. Borderline Personality Disorder The court rejected Dr. Kirschner’s diagnosis of BPD. It found that he failed to identify which diagnostic criteria respondent met for BPD, just as he had failed to do with regard to ASPD. The court also noted that Dr. Kirschner was the only mental health professional who had ever diagnosed respondent with BPD, even though respondent had undergone evaluations for more than 40 years by multiple evaluators. The court credited Dr. Bard’s testimony that some of the symptoms of bipolar disorder can mimic symptoms of BPD and that BPD is best diagnosed by an interview of the subject. The trial court also credited Dr. Bard’s testimony that he rejected a diagnosis of BPD in favor of bipolar disorder for a variety of reasons, including having witnessed respondent in a manic state indicative of bipolar illness. Dr. Bard also testified that people with BPD usually have great difficulty in maintaining relationships, whereas respondent had been married twice, in one case for five years. Consequently, the trial court had a substantial basis for declining to credit Dr. Kirschner’s diagnosis of respondent with BPD. Bipolar Disorder, Unspecified Depressive Disorder, and Unspecified Bipolar Disorder Finally, the court noted that all three experts agreed that respondent had some kind of depressive disorder. It credited Dr. Bard’s diagnosis of bipolar disorder over Dr. Cox’s diagnosis of unspecified depressive disorder and Dr. Kirschner’s diagnosis of unspecified bipolar disorder. In particular, the court noted that respondent had been diagnosed with bipolar disorder when he was an adolescent. Furthermore, Dr. Bard had the advantage of having previously evaluated respondent during a time when respondent was exhibiting manic behavior that Dr. Bard personally witnessed. In contrast, both Dr. Cox and Dr. Kirschner testified that they did not diagnose respondent with bipolar disorder because they had not witnessed, or were not aware of, respondent exhibiting manic behavior, a necessary factor in diagnosing bipolar illness. Accordingly, I would find that the trial court had overwhelming reasons to credit the testimony of Dr. Bard over the testimony of the state’s expert witnesses, and to find them not credible. Based on this, I would find that the court’s citation to its own research as to bias constituted harmless error, as it could have had at most a minimal role in its findings as to the credibility of the expert witnesses. Moreover, the trial court’s decision concluded that the research it cited did not address whether forensic psychologists “evince a bias blind spot.” Similarly, I would find that the court’s citation to research on Suboxone was completely irrelevant to its credibility determination. Finally, I would find that the trial court’s citation to the DSM-5 and PCL-R manual were consistent with the expert testimony. Furthermore, notwithstanding the majority’s claim, the trial court did not reject the combined diagnosis principle. Rather, the trial court repeatedly cited cases from the Court of Appeals and from this Court (Dennis K., 27 NY3d at 751-752; David D., 206 AD3d at 485) for the proposition that mental abnormality may be based on a diagnosis of “multiple illnesses, not to be viewed in isolation, that work in tandem with one another to predispose the offender to commit a sex offense” and cause the offender “serious difficulty” with controlling that conduct. Accordingly, the trial court acknowledged the holdings of these cases as controlling authority, and his discussion of cases to the contrary is dicta. In this case, Dr. Cox and Dr. Kirschner each testified that all of their diagnoses taken together predispose respondent to commit a sex offense and cause him ‘serious difficulty’ in controlling that conduct. However, the trial court concluded that some of those diagnoses were not supported by the evidence. He thus held that the experts’ conclusion of mental abnormality must fail, in the absence of specific testimony concluding that the remaining credible diagnoses could cause these effects. As discussed above, the trial court did not find credible any of the diagnoses or conditions assigned to respondent by the states’ witnesses except opioid use disorder. All of the experts agreed that the remaining diagnosis of opioid use disorder was insufficient standing alone for a finding of mental abnormality, since it alone would not make respondent predisposed to commit a sexual offense or to cause him ‘serious difficulty’ in controlling that conduct. Moreover, while the trial court credited Dr. Bard’s diagnosis of bipolar disorder, there was no testimony establishing that this condition, either alone or in combination with opioid use disorder, could cause respondent to have a predisposition to commit sex offenses or to have serious difficulty controlling those urges. Accordingly, the court properly found that the State failed to meet its burden to prove that respondent suffers from a mental abnormality. As discussed above, I conclude that the trial court decision reflected a fair interpretation of the evidence. I would find that the court appropriately weighed the evidence in crediting the conclusions of respondent’s expert over the State’s experts. I would find that the trial court’s citation to the DSM-5 and PCL-R manual were consistent with the testimony and its citation to research outside the record was harmless error. I disagree that the trial court rejected the combined diagnosis principle, Accordingly, it is my view that the trial court properly found that the State failed to prove by clear and convincing evidence that respondent suffers from a mental abnormality. Therefore, I respectfully dissent. Order and judgment (one paper) of Supreme Court, Bronx County (David L. Lewis, J.), entered April 4, 2023, reversed, on the law, without costs, the petition reinstated, and the matter remanded for a new trial on mental abnormality under article 10. Appeal from the order, same court and Justice, entered on or about March 30, 2023, dismissed, without costs, as subsumed in the appeal from the aforesaid order and judgment. Opinion by Singh, J. All concur except Gesmer, J. who dissents in an Opinion. Kern, J.P., Singh, Gesmer, Scarpulla, O’Neill Levy, JJ. THIS CONSTITUTES THE DECISION AND ORDER OF THE SUPREME COURT, APPELLATE DIVISION, FIRST DEPARTMENT. Dated: April 23, 2024
