In re COMMITMENT OF Gary BURTON.
Gary Burton, Appellant,
v.
State of Florida, Appellee.

*1113 James Marion Moorman, Public Defender, and Deborah K. Brueckheimer, Assistant Public Defender, Bartow, for Appellant.

Charles J. Crist, Jr., Attorney General, Tallahassee, and Janet A. McDonald, and Marilyn Muir Beccue, Assistant Attorney General, Tampa, for Appellee.

LEVENS, WILLIAM P., Associate Judge.

Gary Burton seeks review of his civil commitment as a sexually violent predator under the Jimmy Ryce Act, section 394.910-.931, Florida Statutes (1999) ("the Act"). Burton takes issue with several pretrial rulings of the trial court, several evidentiary rulings that took place during the course of trial, the State's closing argument, and the jury instructions. Our review of the record revealed no reversible error, and we affirm the commitment order in all respects. We write to address the issues that are likely to recur in other commitment proceedings under the Act.

FRYE ISSUES

Prior to trial, Burton filed a motion in limine requesting the exclusion of expert testimony regarding certain tests used to determine whether a defendant is a sexually violent predator. The court conducted a Frye^[1] hearing and denied the motion. At the Frye hearing, the parties presented evidence about seven instruments that experts used to determine whether a defendant qualifies as a sexually violent predator (VRAG, SORAG, MnSOST, MnSOST-R, RRASOR, PCL-R and SVR-20). The expert testimony in this case only relied upon the RRASOR, PCL-R, and SVR-20.

Under Frye, expert opinion testimony that relies on novel scientific principles will only be admitted if it is generally accepted by the members of its particular field. Id. at 1014. We have previously determined that the RRASOR and PCL-R satisfy the requirements articulated in Frye. See Rodgers v. State (In re Commitment of Rodgers), 875 So. 2d 737, 739 (Fla. 2d DCA 2004); Lee v. State, 854 So. 2d 709, 712 (Fla. 2d DCA 2003). We conclude that the SVR-20 is not subject to the requirements of Frye.

The SVR-20 consists of twenty risk factors to be assessed by a clinician during a clinical interview as part of a sexual offender evaluation. The SVR-20 does not produce a numerical score or predict a degree of recidivism risk. Instead, it is a guidepost that assists the professional by focusing the clinical interview on risk factors which research has determined to be relevant to recidivism.

Given the very nature of the SVR-20, it is clearly not the type of objective scientific principle or test with which Frye is concerned. Any testimony regarding the use and application of the SVR-20 is tantamount to pure opinion testimony. An expert's "pure opinion testimony," based "solely on the expert's training and experience," is admissible without having to satisfy the requirements of Frye. Hadden v. *1114 State, 690 So. 2d 573, 579-80 (Fla.1997); Flanagan v. State, 625 So. 2d 827, 828 (Fla.1993). But see Collier v. State, 857 So. 2d 943, 945-46 (Fla. 4th DCA 2003) (holding that the SVR-20 does not satisfy Frye).

VIOLATION OF PLEA BARGAIN

Prior to trial, Burton filed a motion to enforce his plea agreement and to dismiss the civil commitment petition. Burton argued that the State violated the plea agreement (a twelve-year prison sentence followed by fifteen years of probation) by seeking civil commitment pursuant to the Act. The Florida Supreme Court has addressed this issue and determined that the State does not violate a plea agreement to a term of incarceration followed by probation by seeking civil commitment under the Act after the term of incarceration is served. See State v. Harris, 881 So. 2d 1079, 1083 (Fla.2004); Murray v. Regier, 872 So. 2d 217, 224 (Fla.2002); see also Rodgers, 875 So.2d at 741; State v. Heath, 865 So. 2d 633, 634 (Fla. 2d DCA 2004); Cartwright v. State (In re Commitment of Cartwright), 870 So. 2d 152, 155 (Fla. 2d DCA 2004).

HEARSAY

Burton argues that the admission of hearsay under the Act^[2] is unconstitutional because it violated his right to confrontation under the Confrontation Clause of the Sixth Amendment of the U.S. Constitution and the parallel provision in article I, section 16(a) of the Florida Constitution. This court has recently rejected this argument. See Cartwright, 870 So.2d at 156.

"SEXUALLY VIOLENT PREDATOR"

Burton argues that the trial court erred in allowing Burton to be referred to as a "sexually violent predator" at trial. The supreme court has expressly allowed the State and court to use the term "sexually violent predator" to explain to the jury that their job is to determine whether the defendant meets this status. Standard Jury Instructions-Criminal Cases (99-2), 777 So. 2d 366, 367 (Fla.2000). The only caveat to the use of the term is that it must not become a "feature of the trial." Id. The use of the term was proper in this case.

JURY INSTRUCTIONS

Burton argues that the jury was not properly instructed on an essential element of volitional control. Burton claims that the jury should have been required to find that he had a serious difficulty in controlling his dangerous behavior. This issue was not preserved and does not rise to the level of fundamental error. See Westerheide v. State, 831 So. 2d 93, 109 (Fla.2002); Rodgers, 875 So.2d at 741; Lee, 854 So.2d at 716. As we have previously done, we certify the following question as one of great public importance:

MAY AN INDIVIDUAL BE COMMITTED UNDER THE JIMMY RYCE ACT IN THE ABSENCE OF A JURY INSTRUCTION THAT THE STATE MUST PROVE THAT THE INDIVIDUAL HAS SERIOUS DIFFICULTY IN CONTROLLING HIS OR HER DANGEROUS BEHAVIOR?

Affirmed; question certified.

SILBERMAN, J., Concurs.

ALTENBERND, C.J., Concurs with opinion.

*1115 ALTENBERND, Chief Judge, Concurring.

Gary Burton has a very troubled past. At least to a layperson, his current psychological profile suggests that he is the type of sexual offender who would be likely to strike again if released from custody. Our society may very well be safer with Mr. Burton indefinitely confined by involuntary civil commitment. That said, I concur in this opinion with considerable hesitation because the expert testimony that was relied upon to accomplish this result troubles me immensely.

This court is required to perform a de novo review of the admissibility of the actuarial tests, upon which the experts in this case relied to express their opinions, to determine whether those tests are now generally accepted within a relevant scientific community. See Brim v. State, 695 So. 2d 268, 274 (Fla.1997); Brim v. State, 779 So. 2d 427, 428 (Fla. 2d DCA 2000) (citing Hadden v. State, 690 So. 2d 573, 578 (Fla.1997)). We have previously made it clear that this court will not rely upon information outside the record to perform this review. Brim, 779 So.2d at 430. Our record in this case contains no new information since the Frye hearing that occurred four years ago. Thus, for purposes of this review, we must rely on the evidence presented to the trial court and assume that no one has published any article critical of these tests during the last four years. As a result, I must vote to affirm the trial court's order as it affects this case.^[3]

I am inclined to believe, however, that the lawyers and trial judges involved in these cases have not yet identified the issues that need to be examined to determine whether these actuarial tests pass the Frye test and whether evidence regarding the tests is more probative than prejudicial. It is not entirely clear to me that the diagnostic method utilized by these experts is generally accepted within the psychiatric and psychological professions, or that the courts should permit opinion testimony based on these methodologies. The record in this case does not provide this court with a basis to reverse the judgment on appeal. Nevertheless, the judiciary needs to make certain that in future cases we carefully and methodically consider the types of evidence that should be utilized in Jimmy Ryce cases.

This country has a worthy tradition of trial by jury. Juries have long been employed to make decisions that result in the deprivation of individual liberty. We have, however, always expected the jury to make *1116 those decisions based on a unanimous determination of past factual events. We have expected the State to present evidence to those juries that fulfills the highest burden of proof. In Jimmy Ryce proceedings, we are now asking juries to make decisions that will indefinitely deprive citizens of liberty based on predictions of future conduct supported by "clear and convincing" evidence. I remain concerned that the law is expecting these juries to make decisions that humans may currently lack the knowledge and capacity to make. At a minimum, if we are going to use juries for this purpose, we should be extremely careful that the evidence and the burden of proof in these cases are designed to minimize the risk that a jury will imprison a citizen based solely upon a "false positive" result from one relatively new actuarial test.

I. ELEMENTS OF A CLAIM FOR CIVIL COMMITMENT

Under the Jimmy Ryce Act, a person may be involuntarily committed for an indefinite period upon a determination, by clear and convincing evidence, that the person (1) has been convicted of "a sexually violent offense"; (2) "suffers from a mental abnormality or personality disorder"; and (3) the disorder "makes the person likely^[4] to engage in acts of sexual violence if not confined in a secure facility for long-term control, care, and treatment." See § 394.912(10), Fla. Stat. (1999). Pursuant to the United States Supreme Court's opinion in Kansas v. Crane, 534 U.S. 407, 413, 122 S. Ct. 867, 151 L. Ed. 2d 856 (2002), the State must also show that the defendant has "`serious difficulty' in controlling his [dangerous] behavior." See Westerheide, 831 So.2d at 107.

The first two elements of this civil commitment proceeding are not overly controversial. Normally, a prior sexually violent offense can be proven with an accurate copy of an earlier judgment. Likewise, I would not dispute that psychiatrists and psychologists can provide useful expert testimony that a person is currently suffering from a condition that is generally accepted by the psychiatric or psychological professions to be a "mental abnormality or personality disorder."

The stumbling block is the third element necessary to establish this involuntary commitment. The crux of the problem, for both the psychological or psychiatric community and for the legal community, is whether humans have the ability to accurately select those people who, by clear and convincing evidence, are likely to engage in future acts of sexual violence if not confined. Psychologists and psychiatrists must decide whether the members of their profession have a professional capacity to perform this function, and the legal community must decide whether ordinary citizens serving as jurors have this capacity.

Increasingly, psychiatrists and psychologists are responding to outside pressures to develop a scientific method to uniformly identify dangerous sexual offenders. This has resulted in the creation of certain testing instruments intended to establish whether a defendant is likely to engage in future acts of sexual violence. These efforts *1117 are worthy. The more difficult question is whether the experts have yet succeeded to a point where jurors should be called upon to rely upon these tests when deciding issues of liberty.

When presented to a jury, these tests, coupled with the testimony of the experts who rely on them, may "impl[y] an infallibility not found in pure opinion testimony." See Flanagan v. State, 625 So. 2d 827, 828 (Fla.1993). To ensure that a jury will not be misled by experimental scientific methods which may ultimately prove to be unsound, Florida courts apply the Frye test, which provides that expert testimony deduced from a scientific principle is not admissible unless the principle is "sufficiently established to have gained general acceptance in the particular field in which it belongs." Id. (citing Frye v. United States, 293 F. 1013, 1014 (D.C.Cir.1923)). I remain concerned that these actuarial tests may not actually pass the Frye test. Even if they do, the current margin of error in these tests suggests that they may be more unduly prejudicial than probative.

II. THE FRYE HEARING IN THIS CASE AND THE QUESTIONABLE SELECTION OF A RELEVANT SCIENTIFIC COMMUNITY

Prior to the trial in this case, a consolidated Frye hearing was held involving this and twelve other civil commitment cases that were pending in the Tenth Judicial Circuit. The hearing occurred in 2000, and Judge Cecelia M. Moore presided. The State presented testimony from Dr. Dennis Doren, a licensed psychologist in Wisconsin who primarily handles sex offender commitments, and Dr. Ted Shaw, a licensed psychologist in Florida who specializes in forensic testimony in sex offender cases. Both of these men are experienced in the clinical assessment of sexual predators. However, both men would seem to have a professional interest in the admissibility of these new tests. The State did not provide testimony from a psychiatrist. It also did not provide testimony of academic psychologists or psychiatrists who would be expected to peer review new or novel tests and procedures like the ones espoused in this case. Little was done to establish that these tests pass the traditional scientific benchmarks of reliability and validity.

Dr. Stephen Hart and Dr. Randy Otto testified for the various respondents. Dr. Hart is a professor of psychology and a clinical forensic psychologist in Canada. His testimony generally criticized the tests at issue in this case and provided evidence that might be helpful in rejecting the tests under the Daubert standard. See Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993). He testified that actuarial assessment is accepted within the profession of psychology but that these specific tests have not been accepted. Dr. Otto is a licensed Florida psychologist and an associate professor at the Department of Mental Health Law and Policy at the Florida Mental Health Institute at the University of South Florida. His testimony was similar to that of Dr. Hart.

The respondents also filed the deposition of Dr. Fred Berlin. Dr. Berlin was the only psychiatrist who provided evidence at the Frye hearing. He was very critical of these procedures and relied on publications questioning their accuracy.

After hearing this evidence, Judge Moore ruled in favor of the State. Her order is very short and does not contain a thorough discussion of the various tests. She seems to have been impressed by the fact that the State's experts had extensive experience with sex offenders, and to have discounted the testimony of the respondents' experts because they had not done *1118 assessments subsequent to the creation of these tests. Dr. Berlin, whose credentials seem very impressive, was discounted because he "does not use actuarials." If professionally he does not trust these tests, one would hope he does not rely upon them.

Perhaps most troubling, the order finds that these tests are "accepted by a clear majority of the professional community in assessing the risks of recidivism for sexually violent offenders." For purposes of a Frye determination, a trial court must decide whether a test or procedure is generally accepted within "the particular field within which it belongs." Frye, 293 F. at 1014. This has been described as the "relevant scientific community." See Castillo v. E.I. Du Pont de Nemours & Co., 854 So. 2d 1264, 1271 (Fla.2003); Ramirez v. State, 810 So. 2d 836, 843 (Fla.2001); Brim, 695 So.2d at 272. No party objected to Judge Moore's description of the relevant scientific community. Certainly, none of the State's experts could testify about acceptance within the psychiatric community.

The concept that humans can accurately predict the criminal or aberrant behavior of other humans in the long-term future is new and probably novel. I would think that the relevant scientific community that must generally accept these tests and the interpretation of their results should include a broader group of clinical and experimental psychologists and psychiatrists, and not merely the group of licensed professionals who are making a living by relying upon these tests. See. e.g., Ramirez v. State, 810 So. 2d 836, 844 n. 13 (Fla.2001) ("In applying the Frye criteria, general scientific recognition requires the testimony of impartial experts or scientists. It is this independent and impartial proof of general scientific acceptability that provides the necessary Frye foundation."); Sybers v. State, 841 So. 2d 532, 542 (Fla. 1st DCA 2003). But see Roeling v. State, 880 So. 2d 1234 (Fla. 1st DCA 2004) (defining the relevant scientific community that generally accepts these actuarial tests as licensed clinical psychologists specializing in forensic psychology and the evaluation of sexually violent predators). I find it very hard to believe that the membership of the American Psychiatric Association and the membership of the American Psychological Association have reached a consensus that their professions can achieve these predictions with a level of accuracy sufficient to permit the indefinite confinement of an individual.^[5]

Before we admit testimony based on these actuarial tests and before we allow psychologists to quantify their predictions about the future behavior of specific people, it seems to me that we need to have better answers to better questions regarding the specific relevant scientific community and what that community "generally accepts" in this area. I do not profess to have the expertise to even phrase all of the questions, but the judiciary needs to obtain the help of those who can ask and answer the necessary questions.

*1119 III. THE USE OF ACTUARIAL INSTRUMENTS IN PREDICTING FUTURE HUMAN BEHAVIOR, THEIR ERROR RATE, AND THEIR INHERENT RISK OF FALSE POSITIVES

It is important to note that the tests relied upon by the experts in this case are not traditional diagnostic tests of the individual. Rather, they are actuarial tests, designed to establish or define a small group or sub-population of people in which the risk that a member of the group will commit a violent sexual offense is higher than in the population as a whole. I do not question whether psychologists could find these tests useful and generally acceptable within their profession to determine a mode of treatment for a voluntary patient. Even if they do, and even if these professionals would also use the tests for some predictive purpose, the judicial system still needs to address whether it should permit the use of these tests in involuntary civil commitment proceedings. See Ramirez, 810 So.2d at 842 (explaining that all evidence must meet test of legal reliability; scientific reliability is a predicate to legal reliability). Unless I fail to understand actuarial tests, which is certainly a distinct possibility, these tests appear to have an inherent error rate that is unacceptable for purposes of involuntary commitment.

Actuarial devices are designed to identify a small pool of people in which the likelihood of a particular condition is higher than in the population as a whole. But the pool always contains some people who do not have the particular condition. The people falsely identified by the actuarial device determine its error rate. For many purposes, an error rate of 30% or more is quite acceptable. Life insurance companies, for example, usually charge higher premiums or refuse to insure the pool of obese, cigarette smokers because the probability of a premature death is higher among the members of this pool, even though many members of the pool live to an average age. It is one thing to price insurance based on actuarial device with an error rate of 20% or higher; it is quite another to deprive citizens of their constitutional liberty based on actuarial devices with such high error rates. There is no question that the judiciary would not allow expert opinions based on fingerprints or DNA in criminal cases if the methodology and statistical analysis in those fields produced an error rate of even 10%.

In this case, the jury heard testimony from a forensic psychologist that Mr. Burton scored a 4 on the RRASOR test and a score of 4 "related to an approximate 32.7% risk for sexual recidivism within five years, and a 48.6% rate of recidivism within ten years." The RRASOR is the Rapid Risk Assessment for Sexual Offender Recidivism, an actuarial test that was created in 1997 by R. Karl Hanson, Ph.D.^[6] This is not a lengthy test like some of the other well-established psychological tests seen in other types of litigation. The RRASOR considers only four categories. It scores 1 to 3 points based on the person's prior record of sex offenses, including charges that did not result in conviction. It gives 1 point if the person is under the age of 25. It gives 1 point if any victim was male, and 1 more point if the victim was unrelated to the person being tested. Thus, an 18-year-old man with one conviction for lewd behavior involving an unrelated boy scores 4 on the RRASOR.

Intuitively, I find it hard to believe that the knowledge that an 18-year-old man *1120 has one conviction for lewd behavior involving an unrelated boy is sufficient information to conclude that there is a 48.6% probability that the man will commit a violent sexual crime during the next decade. It also troubles me that Mr. Burton can successfully complete a full course of rehabilitation and the RRASOR will not have changed its assessment of him. With or without successful treatment, he has a 48.6% chance of doing a bad act in the future according to this test.

Treatment does not vary his score because the RRASOR is not a test of the individual, but an actuarial device that merely creates a pool of people with a higher risk of being sexual predators. By analogy, I would be willing to bet that there is a 48.6% chance that any member of a pool of 24-year-old married women in households of average income who already have one child will give birth to another child within the next ten years. This is an actuarial device that mail order catalogues might successfully use for marketing products for infants, but the fact that more than half of these women will not give birth to a child should limit the use of such an actuarial device for individualized legal decisions affecting the women's liberty.

It may well be that 48 out of every 100 individuals in the sub-group of people who score 4 on the RRASOR will commit a violent sexual offenses in the next decade. This means, of course, that 52 out of every 100 members of this group will not commit such an offense in the next ten years. Mr. Burton, or any other person who scores a 4 on this test, could be in either group.^[7] I question whether this is the type of quality scientific evidence that jurors should consider for these serious purposes.

When a scientist incorrectly places a person into a group or category based on a scientific test, the scientist regards the error as a "false positive." In the context of Jimmy Ryce proceedings, a "false positive" places a person who will not commit a future act of sexual violence into a civil commitment facility for an indefinite period of long-term confinement. That facility looks very similar to a prison. Once confined, it is very difficult for a person to prove how he or she would have behaved, or may in the future behave in the outside world. As a parent of young children, I understand the temptation to involuntarily commit the 52 "false positives" under the RRASOR in order to protect society from the other 48. As a judge sworn to uphold the constitutional rights of all citizens, I am not so convinced we should accept such a high error rate. Just as we would not allow fingerprint experts or DNA experts to testify if their identifications were wrong 52% of the time, I question the role of these tests and opinions in Jimmy Ryce proceedings.

We have embarked on the first steps into a new world, arguably a science fiction world, in which judges and juries are asked to prevent crimes years before they occur. There are many violent crimes that are associated with mental abnormalities, and sexual crimes may only be the start of this process. Actuarial devices could certainly be established for drunk drivers, burglars, robbers, and car thieves. I, for one, do not yet have faith that it is wise for the judiciary or for society as a whole to rush down this new path before we are confident that both the science of jurisprudence and the sciences of psychology and psychiatry are up to this awesome task. *1121 We need to protect our society, especially our children, from sexual predators, but we also need to honor and trust the heritage of freedom and liberty that has made this country strong.

NOTES

[1] Frye v. United States, 293 F. 1013 (D.C.Cir.1923).

[2] Section 394.9155(5), Florida Statutes (1999), allows for the use of hearsay in civil commitment proceedings under the Act as long as it is reliable and not the sole basis for the commitment.

[3] I doubt that this court's opinion in Lee, 854 So. 2d 709, is dispositive of this case. In Lee, this court concluded that based upon the conflicting evidence presented to the trial court in that case, "the trial court did not err by allowing the test results to be presented to the jury as part of the expert's testimony." Id. at 712. However, we specifically held that "even if there were any error regarding the Frye analysis and the admissibility of actuarial evidence," the error was harmless. Id. at 712-13. Thus, any statement that the tests involved in Lee satisfied the requirements articulated in Frye was at most an alternative holding, and more likely dicta. But see Rodgers v. State, 875 So. 2d 737, 739 (Fla. 2d DCA 2004) (stating that Lee"rejected a Frye challenge" to the use of actuarial tests and therefore "compelled" court to conclude that trial court's permitting testimony regarding the same tests was not error). I recognize that the First District has recently recognized a variety of these tests as sufficiently reliable to pass the Frye test. See Roeling v. State, 880 So. 2d 1234 (Fla. 1st DCA 2004). See also Jackson v. State, 833 So. 2d 243, 246 (Fla. 4th DCA 2002); but see Collier v. State, 857 So. 2d 943, 945-46 (Fla. 4th DCA 2003) (holding SVR-20 test and testimony based upon it inadmissible based upon psychologist's admission that test was "somewhat experimental" and questioned by some in psychological science community).

[4] The statute does not specify when in the future it must be "likely" that these possible crimes might occur. Compare § 394.467(1)(a)(2)(b), Fla. Stat. (2003) (permitting involuntary commitment under the Baker Act when there is a "substantial likelihood that in the near future he or she will inflict serious bodily harm on himself or herself or another person, as evidenced by recent behavior causing, attempting, or threatening such harm"). Typically, experts in Jimmy Ryce cases provide statistical probabilities for five- and ten-year periods.

[5] It is useful to compare a Jimmy Ryce proceeding to a Baker Act proceeding. Under the Baker Act, involuntary admission to a treatment facility must be based on the opinion of a psychiatrist, supported by a second opinion of a psychologist or psychiatrist. See § 394.467(2), Fla. Stat. (2003). The common diagnostic issue is whether the patient "in the near future" "will inflict serious bodily harm on himself or herself or another person, as evidenced by recent behavior causing, attempting, or threatening such harm." See § 394.467(1)(a)(2)(b). Thus, the legislature in the Baker Act has demanded proof of recent conduct that demonstrates unacceptable risks in the near term. It anticipates that both psychiatrists and psychologists will be involved in this assessment. This structure does not appear to strain the limits of science in the way that Jimmy Ryce proceedings do.

[6] Dr. Hanson is a psychologist employed with the Canadian government's Portfolio of Public Safety and Emergency Preparedness Canada (PSEPC) in the area of Corrections Research.

[7] Note that this test does not purport to say what the chances are that Mr. Burton individually will engage in future sexually violent behavior. Rather, it predicts that 48% of people like him in a few ways will commit these acts. A jury unskilled in statistical analysis might misinterpret the results to apply to the individual.