concurring.
I concur in the judgment of the court reversing the grant of the writ of habeas corpus. I write separately to set forth my views on the appropriate method of analysis in determining whether the admission of T’s statements to Dr. McLean conformed to the requirements of the Confrontation Clause. In particular, I do not agree with the majority's assertion that "T's unavailability was not a constitutional sine qua non of allowing her out of court statements to be used in evidence against her father.” Rather, I believe that the Confrontation Clause required the prosecution to show that T was not available to testify at trial and that the statements themselves had particularized guarantees of trustworthiness. In my view, the prosecution has complied with both requirements. Consequently, I believe that the admission of T’s statements through Dr. McLean did not violate the defendant’s sixth amendment rights.
I.
In Ohio v. Roberts, 448 U.S. 56, 100 S.Ct. 2531, 65 L.Ed.2d 597 (1980), the Supreme Court appeared to fix both the unavailability of the declarant and the reliability of the statement as enduring lodestars in Confrontation Clause cases. In Roberts, a case involving testimony given at a preliminary hearing, the Court stated that generally, “the prosecution must either produce, or demonstrate the unavailability of the declarant whose statement it wishes to use against the defendant.” Id. at 65, 100 *1231S.Ct. at 2538. The Court, however, emphasized that a showing of unavailability is not required in all situations. Id. at 65 n. 7, 100 S.Ct. at 2538 n. 7.
The general language contained in Roberts prompted many courts to conclude that a demonstration of the declarant’s unavailability was a constitutional pre-condition to the admission of most hearsay statements. See United States v. Caputo, 758 F.2d 944, 950 n. 2 (3rd Cir.1985) (collecting cases). The validity of this interpretation of Roberts, however, is suspect in view of United States v. Inadi, 475 U.S. 387, 106 S.Ct. 1121, 89 L.Ed.2d 390 (1986). The precise holding in Inadi was that no unavailability showing is required when statements are admitted pursuant to Federal Rule of Evidence 801(d)(2)(E) — co-conspirator statements made during the course and in furtherance of the conspiracy. The Court justified this result on three grounds. First, the Court found that these statements derive much of their significance from the context in which they are made and that this context cannot be replicated at trial. Id. at 395, 106 S.Ct. at 1126-27. Second, the Court noted that the declarant will often have a motive to testify falsely about the statements in view of the specter of imminent indictment or prosecution. Id. Finally, the Court emphasized that the benefit to the defendant of having a co-conspirator testify is minimal and is substantially outweighed by the burden to the prosecution resulting from an unavailability requirement. Id. at 396-400, 106 S.Ct. at 1127-29. In the course of the opinion, however, the Court employed language that could be interpreted as repudiating the unavailability discussion in Roberts.
The Inadi decision has created an unfortunate vacuum in the Confrontation Clause realm, for at present it is not clear if a showing of unavailability is required for most types of hearsay statements. See Mosteller, Child Sexual Abuse and Statements for the Purpose of Medical Diagnosis or Treatment, 67 N.C.L.Rev. 257, 289 n. 9 (1989). Given its broadest construction, Inadi stands for the proposition that the unavailability of the declarant is a relevant constitutional factor only when the hearsay statements involve testimony given at a preliminary hearing. In my view, however, Inadi does not represent a repudiation of Roberts’ unavailability discussion. Rather, I believe that Inadi merely reaffirms and applies the Roberts principle that a showing of unavailability is not required in all situations. See Lee v. Illinois, 476 U.S. 530, 549 n. 2, 106 S.Ct. 2056, 2067 n. 2, 90 L.Ed.2d 514 (1986) (Blackmun, J., dissenting). Thus, to me, Inadi represents a directive to lower courts to carefully analyze the facts of a given situation before concluding that a showing of unavailability is constitutionally required. Moreover, I believe that Inadi provides us with the factors to be used in making this determination — significance of the context in which the statements were made, potential change in the status of the declarant, burden to the prosecution of demonstrating unavailability and benefit to the defendant of having the declarant testify.
The delineation of these factors does not of course end the inquiry, for the question of application remains. In my view, there are two possible approaches to this problem although both lead to the same result in the present case. The first approach, which may be termed the case-by-case approach, involves analyzing the facts of a particular case in terms of the criteria delineated in Inadi. Under this approach, I believe that the state was constitutionally required to demonstrate T’s unavailability. Unlike Inadi, T’s statements were not made during the course of the critical event but rather represented her recollection of past events. Thus, it cannot be said that the statements derived much of their significance from the context in which they were made. Second, T’s status did not change appreciably in the interval between the making of the statements and the trial —T was a little girl who was not facing the threat of prosecution and had no motive to testify falsely at trial. Third, the burden to the prosecution of producing T or demonstrating her unavailability was not substantial. T was readily identified and located and could have easily been transported to trial. Conversely, the benefit to the *1232defendant of having T testify at trial was substantial. T’s statements went to the essential issue in the case — the identity of the abuser. Moreover, T was the only individual who had the opportunity to observe the abuser and, in fact, gave conflicting accounts of his identity. Under these circumstances, T’s testimony at trial could have been critical to the defendant for if T repudiated her identification of the defendant on the stand, in all probability a judgment of acquittal would have been entered.
The case-by-case approach is appealing in the abstract, for in theory, it should accurately identify the cases where a showing of unavailability would be pointless. On further reflection, however, significant problems with this approach emerge. In particular, I believe that a case-by-case approach will engender a great deal of uncertainty in a critical area of the law. Thus, neither party will know prior to trial whether a showing of unavailability will be required. Consequently, the prosecution may in fact expend a great deal of effort in order to produce a declarant only to discover that the presence of the declarant at trial was not required. Moreover, it will often not be possible for the trial judge to accurately evaluate the Inadi factors at the time the prosecution offers the statements. In particular, the benefit to the defendant of the declarant’s testimony may not become apparent until other witnesses testify.
A second possible approach would link the unavailability requirement to the particular hearsay exception. Under this approach, a determination that a statement fits within the medical diagnosis exception will necessarily dictate a particular result on the unavailability issue. By its terms, this approach demands assumptions about the nature of evidence admitted under the particular exception in criminal cases and the identity of the typical declarant. These assumptions in turn are to be evaluated in light of the criteria delineated in Inadi. As with all generalizations, there is always the possibility of an erroneous result in a particular case. In my view, however, the danger of erroneous results when this exception is involved is slight and is substantially outweighed by the certainty that this approach would provide. The need for certainty, a primary goal of Confrontation Clause jurisprudence, see Ohio v. Roberts, 448 U.S. 56, 66, 100 S.Ct. 2531, 2539, 65 L.Ed.2d 597 (1980), is particularly acute in the case of this particular exception for the medical diagnosis exception is being used with increasing frequency in prosecutions for child abuse. See United States v. Shaw, 824 F.2d 601 (8th Cir.1987); United States v. Renville, 779 F.2d 430 (8th Cir.1985); United States v. Iron Shell, 633 F.2d 77 (8th Cir.1980); United States v. Nick, 604 F.2d 1199 (9th Cir.1979).
As noted above, the Supreme Court in Inadi identified several factors in determining that the prosecution need not demonstrate the declarant’s unavailability when statements fall under 801(d)(2)(E)— significance of the context in which the statements were made, potential change in the status of the declarant, burden to the prosecution of producing the declarant, and minimal benefit to the defendant of having the declarant testify. In my view, these factors are inapposite to the typical case involving Rule 803(4) or its state counterpart. Consequently, I would hold that when statements are admitted pursuant to the medical diagnosis exception to the hearsay rule, a showing of unavailability is always required.
First, I do not believe that statements ordinarily admitted under the medical diagnosis exception derive much significance from the context in which they are made. Unlike the co-conspirator context, the de-clarant under the medical diagnosis exception is not engaged in an illicit enterprise at the time the statements were made; in fact as in this case, the statements are generally not made when the relevant act is taking place but rather represent the declarant’s recollection of past events. In this respect, the statements are akin to trial testimony. Thus, I believe that “when the declarant takes the stand, his in-court testimony [generally] will reproduce a significant portion of the evidentiary value of the statements.” Inadi, 475 U.S. at 395, 106 S.Ct. at 1126.
*1233Second, unlike Inadi, the status of the typical declarant under this particular exception will not have changed materially between the time the statements were made and the time of trial. The declarant under this exception will generally be a victim not a criminal facing the threat of imminent prosecution, and will usually have no motive to testify falsely about his or her prior statements.
Finally, in my view, the outcome of Ina-di’s balancing test is reversed when the medical diagnosis exception is involved. In Inadi, the Court emphasized the burden to the entire criminal justice system of adding another tier of appellate review in complex cases and the burden to the prosecution of identifying, locating and transporting the declarant to trial. These conclusions are premised on the special nature of conspiracies which often involve several individuals scattered throughout a wide area. These factors, however, are not present under the medical diagnosis exception. In a typical case under this exception there is a single declarant, — the victim, who is easily identified and located.
On the other hand, the value of the testimony by a declarant under this exception is generally substantial. As noted above, the typical declarant under this exception is the victim, and generally is the only individual with first-hand knowledge of the crime. Moreover, the statements themselves generally involve central issues in the case such as identity, United States v. Renville, 779 F.2d 430, 437 (8th Cir.1985), cause, United States v. Nick, 604 F.2d 1199, 1201-02 (9th Cir.1979) or fact of injury. Given these facts, in most cases the appearance of the typical declarant under the medical diagnosis exception is of great value to the defendant for by discrediting the declarant’s testimony, the defendant necessarily casts doubt upon the principal evidence in favor of the prosecution.
Thus, under either a case by case or exception by exception approach, I believe that the Confrontation Clause required the prosecution to demonstrate T’s unavailability as a condition to the admission of the statements through Dr. McLean. Having said this, I nevertheless conclude that the prosecution has succeeded in demonstrating that T was unavailable. I do not reach this conclusion due to T’s age. While the fact that T was three years old at the time of trial is certainly a relevant factor in determining unavailability, it is not in my view dispositive since extremely bright, verbal three year olds can provide meaningful testimony at trial. See e.g., United States v. Frazier, 678 F.Supp. 499 (E.D.Pa.1986). Rather, I believe that the prosecution clearly established that T could not provide meaningful testimony about the alleged assault. The testimony of Dr. McLean, the only professional who had significant contact with T, indicated that the victim, even on the eve of trial strongly resisted making any statements about her father. Moreover, the defendant introduced no evidence that even obliquely questioned the validity of this assertion. Given these facts, I believe that the prosecution established that T was not available to testify at trial.
II.
In addition to establishing T’s unavailability the prosecution also had to demonstrate that the statements had sufficient indicia of reliability. In Roberts, the Court held that the reliability requirement could be satisfied in two ways. First, “reliability could be inferred without more when the evidence fell within a firmly rooted hearsay exception.” 448 U.S. at 66, 100 S.Ct. at 2540. Second, even if the evidence did not come within a firmly-rooted exception, reliability could be shown if the evidence had particularized guarantees of trustworthiness.” Id.
In Bourjaily v. United States, 483 U.S. 171, 107 S.Ct. 2775, 97 L.Ed.2d 144 (1987), the Supreme Court elucidated the meaning of the term “firmly-rooted.” There, the Court held that a firmly-rooted hearsay exception is one that has “a long tradition of being outside the compass of the general hearsay exclusion.” Id. 107 S.Ct. at 2783. Thus, the test of “firmly-rooted” is simply the longevity of the exception’s existence. The Court noted, however, that when the *1234evidentiary rule deviates from the common law approach an independent reliability inquiry is required. Id.
The Wisconsin Supreme Court held that T’s statements identifying the defendant as the abuser fit within the medical diagnosis or treatment exception to the hearsay rule and the validity of this interpretation is not an issue in this habeas corpus case. At common law, the admission into evidence of statements made to others for purposes of treatment was a well-established practice. See United States v. Nickle, 60 F.2d 372, 373-74 (8th Cir.1932). The common law version of this exception, however, did not embrace statements of fault; the admission of these types of statements under this particular exception is the product of the Federal Rules of Evidence (enacted in 1974) and recent court decisions. See People v. LaLone, 432 Mich. 103, 437 N.W.2d 611 (1989). Since this development represents a marked deviation from and expansion of the common law approach, Bourjaily dictates that an individualized examination into the reliability of T’s statements is required.
As noted above, under the second Roberts test statements are admissible if they have particularized guarantees of trustworthiness. In my view these guarantees are present in this case. First, as the majority notes, the graphic description of the nature of the abuse in childlike language clearly established that sexual abuse had taken place. Second, T told several other individuals who testified at trial that the defendant had abused her. Finally, shortly after the alleged acts had taken place, T began manifesting an abnormal and uncharacteristic terror at the prospect of seeing her father. Given these facts, I believe that T’s statements to Dr. McLean identifying the defendant as the abuser had particularized guarantees of trustworthiness within the meaning of Roberts.
In sum, I believe that the prosecution was required to show that T was not available to testify at trial and that the statements she made to Dr. McLean had particularized guarantees of trustworthiness. In my view, the prosecution satisfied both of these requirements. Consequently, I agree with the majority that the admission of T’s statements to Dr. McLean at trial did not violate the defendant’s sixth amendment rights.