Opinion
MOSK, J.
Defendant appeals from a judgment convicting him of multiple counts of child molestation. He contends that certain evidentiary rulings at his trial violated his psychotherapist-patient privilege and his right to con[509]*509frontation. As will appear, we conclude that both points are well taken and compel reversal of the judgment.
During a 15-month period ending May 1981 defendant allegedly engaged in various acts of fondling, mutual masturbation, and oral copulation with his stepdaughter Sarah.1 When Sarah’s mother—defendant’s wife—learned of these activities she arranged for her daughter and her husband each to see Dr. Walker, a licensed clinical psychologist. During Sarah’s counseling session on July 28, 1981, she revealed that she had engaged in sexual activity with her stepfather. Dr. Walker reported the conversation to the child welfare agency that same afternoon.2 The agency in turn relayed the information to the sheriff’s office.
The next day Deputy Buttell of the sheriff’s office telephoned Dr. Walker to investigate the child abuse report. Dr. Walker told Buttell that he had seen Sarah the day before, and related the substance of her discussion of sexual relations with her stepfather. He also informed Buttell that he was scheduled to meet with defendant himself later that afternoon, July 29, and with Sarah’s older sister two days later, July 31. The deputy asked the doctor to call back after his session with Sarah’s sister because he was concerned that she might also be the victim of child abuse. However, he hesitated on the issue of defendant’s communications, acknowledging there might be a “confidentiality” problem. This telephone conversation was tape recorded.
Defendant saw Dr. Walker as scheduled, and during his conversation with the psychotherapist discussed his sexual relations with Sarah. Deputy Buttell telephoned the doctor again the next day, July 30, to inquire further about the reported child abuse. When Dr. Walker expressed reservations about disclosing defendant’s confidential communications, Buttell read him Penal Code section 11171, subdivision (b), part of the Child Abuse Reporting Act, which he described as providing an applicable exception to the psychotherapist-patient privilege. The doctor then recounted the substance of defendant’s session of July 29. This telephone conversation was also tape recorded, and a written report summarizing the conversation was prepared.
[510]*510At the opening of trial defendant moved that Dr. Walker’s testimony be excluded on the basis of the psychotherapist-patient privilege. (Evid. Code, § 1014.) The court held that Penal Code section 11171, subdivision (b), provides an applicable exception to the privilege and ruled the testimony admissible.
The court also conducted a pretrial hearing to determine whether Sarah could be declared unavailable as a witness under Evidence Code section 240, subdivision (a)(3), so that her preliminary hearing testimony could be admitted as an exception to the rule against hearsay. (Evid. Code, § 1291.) Based solely on Sarah’s mother’s testimony and on its conclusion that Sarah’s testimony would not be damaging to defendant’s case, the court held that Sarah was suffering from a mental illness or infirmity and was thus unavailable as a witness. (Evid. Code, § 240, subd. (a)(3).) Her preliminary hearing testimony was therefore ruled admissible.
At trial Dr. Walker testified, over objection, regarding his July 29 consultation with defendant. To refresh the doctor’s memory, the district attorney showed him a copy of the report summarizing his second telephone conversation with Deputy Buttell and, in the absence of the jury, played the tape recording of this conversation for him. Sarah’s mother also testified and Sarah’s preliminary hearing testimony was read to the jury. Deputy Ho-berg testified regarding her interview with Sarah, in order to impeach Sarah’s preliminary hearing testimony.
Defendant was convicted of one count of lewd and lascivious conduct with a minor, a felony, in violation of Penal Code section 288a, subdivision (b)(2); one count of misdemeanor child molestation, a necessarily included lesser offense under this section, based on an act of oral copulation; and seven counts of misdemeanor child molestation in violation of Penal Code section 647a. The verdict on one of the latter counts was set aside on defendant’s motion to dismiss. Defendant was sentenced to three year’s probation with ninety days in the county jail.
I
The Psychotherapist-Patient Privilege and the Child Abuse Reporting Act.
Defendant first contends that Dr. Walker’s testimony regarding the consultation of July 29 was erroneously admitted at trial in violation of the psychotherapist-patient privilege, a relationship subsumed in the right to privacy and defined by statutory provision. On the facts of this case, we agree that the doctor’s testimony should have been excluded.
[511]*511Evidence Code section 1014 provides in part that “the patient, whether or not a party, has a privilege to refuse to disclose, and to prevent another from disclosing, a confidential communication between patient and psychotherapist . . . .” We acknowledged in In re Lifschutz (1970) 2 Cal.3d 415, 421 [85 Cal.Rptr. 829, 467 P.2d 557, 44 A.L.R.3d 1], “the growing importance of the psychiatric profession in our modern, ultracomplex society.” Thus for reasons of policy the psychotherapist-patient privilege has been broadly construed in favor of the patient. (Roberts v. Superior Court (1973) 9 Cal.3d 330 [107 Cal.Rptr. 309, 508 P.2d 309]; Grosslight v. Superior Court (1977) 72 Cal.App.3d 502 [140 Cal.Rptr. 278].) Confidential communications between psychotherapist and patient are protected in order to encourage those who may pose a threat to themselves or to others, because of some mental or emotional disturbance, to seek professional assistance. (Grosslight v. Superior Court, supra, 72 Cal.App.3d at pp. 507-508.)
The psychotherapist-patient privilege has been recognized as an aspect of the patient’s constitutional right to privacy. (Cal. Const., art. I, § 1; In re Lifschutz, supra, 2 Cal.3d at pp. 431-432, citing Griswold v. Connecticut (1965) 381 U.S. 479, 484 [14 L.Ed.2d 510, 514-515, 85 S.Ct. 1678]; Ceasar v. Mountanos (9th Cir. 1976) 542 F.2d 1064, 1070.) It is also well established, however, that the right to privacy is not absolute, but may yield in the furtherance of compelling state interests. (Britt v. Superior Court (1978) 20 Cal.3d 844, 855 [143 Cal.Rptr. 695, 574 P.2d 766]; Jones v. Superior Court (1981) 119 Cal.App.3d 534, 550 [174 Cal.Rptr. 148]; Board of Medical Quality Assurance v. Gherardini (1979) 93 Cal.App.3d 669, 680 [156 Cal.Rptr. 55].)
Thus in Lifschutz we held that the patient-litigant exception to the psychotherapist-patient privilege (Evid. Code, § 1016), if narrowly drawn, does not impermissibly invade the patient’s right to privacy; “Even though a patient’s interest in the confidentiality of the psychotherapist-patient relationship rests, in part, on constitutional underpinnings, all state ‘interference’ with such confidentiality is not prohibited.” (In re Lifschutz, supra, 2 Cal.3d at p. 432.) Similarly in Jones v. Superior Court, supra, 119 Cal.App.3d at page 550, the court concluded that “The constitutional right is by no means absolute. The state’s interest in facilitating the ascertainment of truth in connection with legal proceedings is substantial enough to compel disclosure of a great variety of confidential material, including even communications between a psychotherapist and his patient.”
To determine whether the psychotherapist-patient privilege embraced by the right to privacy has impermissibly been violated, we begin by considering the state’s competing interest. Here that interest is the detection and [512]*512prevention of child abuse, and is expressed in the recently enacted Child Abuse Reporting Act. (Pen. Code, § 11165 et seq.) Section 11166, subdivision (a), of the act provides in part that “any child care custodian, medical practitioner, nonmedical practitioner, or employee of a child protective agency who has knowledge of or observes a child in his or her professional capacity or within the scope of his or her employment whom he or she knows or reasonably suspects has been the victim of child abuse shall report the known or suspected instance of child abuse to a child protective agency immediately or as soon as practically possible by telephone and shall prepare and send a written report thereof within 36 hours of receiving the information concerning the incident.” Section 11165 of the act provides the following: subdivision (g) defines “child abuse” to include “the sexual assault of a child . . subdivision (i) defines “medical practitioner” to include licensed psychiatrists and psychologists; and subdivision (k) defines “child protective agency” to include a “police or sheriff’s department” and a “county welfare department.” The scope and substance of the reporting requirement are set out in section 11167. Subdivision (a) thereof states that “A telephone report of a known or suspected instance of child abuse shall include the name of the person making the report, the name of the child, the present location of the child, the nature and extent of the injury, and any other information, including information that led such person to suspect child abuse, requested by the child protective agency.” Subdivision (b) further provides that “Information relevant to the incident of child abuse may also be given to an investigator from a child protective agency who is investigating the known or suspected case of child abuse.”
Together these provisions impose on psychotherapists the affirmative duty to report to a child protective agency all known and suspected instances of child abuse. Lest there be any doubt that the Legislature intended the child abuse reporting obligation to take precedence over the physician-patient or psychotherapist-patient privilege, section 11171, subdivision (b), explicitly provides an exception to these very privileges: “Neither the physician-patient privilege nor the psychotherapist-patient privilege applies to information reported pursuant to this article in any court proceeding or administrative hearing.” The Legislature obviously intended to provide specific exception to the general privileges set out in the Evidence Code (Evid. Code, §§ 994, 1014) so that incidents of child abuse might be promptly investigated and prosecuted.
Defendant neither challenges the constitutionality of the child-abuse reporting exception to the psychotherapist-patient privilege, nor argues that the state’s interest in protecting children is less than compelling. Rather, he contends that on the particular facts of his case the exception provided in Penal Code section 11171, subdivision (b), was un[513]*513necessarily and therefore erroneously applied to his confidential communications with Dr. Walker. We agree.
We begin by recognizing our obligation to construe narrowly any exception to the psychotherapist-patient privilege: we must apply such an exception only when the patient’s case falls squarely within its ambit. (In re Lifschutz, supra, 2 Cal.3d at p. 435.) We therefore examine in detail the sequence of events in this case.
The record reveals that Dr. Walker contacted the welfare agency immediately after his consultation with Sarah on July 28, before he met with defendant. When Deputy Buttell first telephoned to investigate this report Dr. Walker elaborated in detail on Sarah’s revelations. As defendant concedes, Sarah’s communications were not privileged because Evidence Code section 1027 provides an exception when, as here, the patient is under 16 years of age and the psychotherapist has “reason to believe that the patient has been the victim of a crime and that disclosure of the communication is in the best interest of the child.” (Fn. 1, ante.)
After Dr. Walker had seen defendant on July 29, Deputy Buttell called again and asked the psychotherapist to disclose the substance of defendant’s communications. Although Buttell persuaded him that the act overrode the privilege and thus elicited the therapist’s response, under the circumstances the doctor was not legally obligated to discuss Sarah’s case with him again. It is clear from the record that in his own therapeutic consultation defendant gave Dr. Walker no reason to suspect any additional criminal activity, beyond the incidents described by Sarah earlier and already reported. In his first tape-recorded conversation with the deputy Dr. Walker indicated that Sarah had been uncertain how many fondling incidents had occurred but, in response to his questioning, had said she thought there might have been 10 or more. He also related that Sarah had denied there had been oral copulation or sexual intercourse. In his trial testimony Dr. Walker stated that defendant had told him there had been approximately six fondling incidents and, again, no oral copulation or sexual intercourse. Thus defendant at most only confirmed what the doctor had already reported to Deputy Buttell in their first conversation, following Sarah’s consultation.
Dr. Walker was under no statutory obligation to make a second report concerning the same activity. Had he learned from defendant of possible further child abuse—whether additional incidents involving Sarah, or other incidents with another child—he would, of course, have been required to report these new suspicions. Or, if Dr. Walker had first learned of the fondling incidents from defendant himself, he would have been bound to report that information as provided in the act. However, on the facts of this [514]*514case, we conclude that Dr. Walker satisfied his statutory reporting obligation when he divulged Sarah’s revelations; he was not required to reiterate his suspicion following consultation with defendant.
The exception to the psychotherapist-patient privilege set out in the Child Abuse Reporting Act applies only to “information reported pursuant to this article . . . .” (Pen. Code, § 11171, subd. (b).) In this case, Dr. Walker reported his suspicion of child abuse following his consultation with Sarah, pursuant to section 11167, subdivision (a), of the act. He was not then required to make a second report of the same incidents, based on defendant’s subsequent redundant communications. Although section 11167, subdivision (b), provides that a psychotherapist “may” report “information relevant to the incident of child abuse” beyond the fact of the incident itself, it would be highly inappropriate to apply subdivision (b) in this case. The record makes clear that although Dr. Walker voluntarily reported Sarah’s disclosure of her sexual relations with her stepfather, he did not want to disclose defendant’s confidential communications on the identical subject. He did so only at the behest of deputy Buttell, who misled him into believing he was required to do so by law. It was therefore error to admit Dr. Walker’s testimony concerning his consultation with defendant.
We have recognized the contemporary value of the psychiatric profession, and its potential for the relief of emotional disturbances and of the inevitable tensions produced in our modern, complex society. (See, e.g., In re Lifschutz, supra, 2 Cal.3d 415, 421; Tarasoff v. Regents of University of California (1976) 17 Cal.3d 425, 440-441 [131 Cal.Rptr. 14, 551 P.2d 334, 83 A.L.R.3d 1166].) That value is bottomed on a confidential relationship; but the doctor can be of assistance only if the patient may freely relate his thoughts and actions, his fears and fantasies, his strengths and weaknesses, in a completely uninhibited manner. If the psychiatrist is compelled to go beyond an initial report to authorities regarding a suspected child abuse and must thereafter repeat details given to him by the adult patient in subsequent sessions, candor and integrity would require the doctor to advise the patient at the outset that he will violate his confidence and will inform law enforcement of their discussions. Under such circumstances it is impossible to conceive of any meaningful therapy. Ironically, in this case medical help was initially what this distraught family sought as a result of these tragic events.
II
The Right to Confrontation and Witness Unavailability.
Defendant also contends that by declaring Sarah unavailable as a witness and thus allowing her preliminary hearing testimony to be read to [515]*515the jury the court deprived him of his federal and state constitutional right to confront witnesses against him and violated similar statutory guarantees. We conclude he is correct on this point as well.
The United States Supreme Court has established that a defendant’s Sixth Amendment right to confrontation is a fundamental right, applicable to the states through the Fourteenth Amendment. (Pointer v. Texas (1965) 380 U.S. 400 [13 L.Ed.2d 923, 85 S.Ct. 1065].) The California Constitution now provides a specific guarantee of the right to confrontation: “The defendant in a criminal cause has the right ... to be confronted with the witnesses against the defendant.” (Cal. Const., art. I, § 15.) A similar guarantee is codified in section 686, subdivision 3, of the Penal Code, which provides that in a criminal action the defendant is entitled “to produce witnesses on his behalf and to be confronted with the witnesses against him, in the presence of the court . . . .”
In California v. Green (1970) 399 U.S. 149 [26 L.Ed.2d 489, 90 S.Ct. 1930], the United States Supreme Court defined the three-fold purpose of the confrontation requirement: (1) to insure reliability by means of the oath, (2) to expose the witness to the probe of cross-examination, and (3) to permit the trier of fact to weigh the demeanor of the witness. (Id. at p. 158 [26 L.Ed.2d at p. 497]; see also People v. Green (1971) 3 Cal.3d 981, 989 [92 Cal.Rptr. 494, 479 P.2d 998].) It is well settled, of course, that the right of confrontation is not absolute. In Barber v. Page (1968) 390 U.S. 719 [20 L.Ed.2d 255, 88 S.Ct. 1318], the high court recognized that “there has traditionally been an exception to the confrontation requirement where a witness is unavailable and has given testimony at previous judicial proceedings against the same defendant which was subject to cross-examination by that defendant.” (Id. at p. 722 [20 L.Ed.2d at p. 258]; accord, People v. Enriquez (1977) 19 Cal.3d 221, 235 [137 Cal.Rptr. 171, 561 P.2d 261, 3 A.L.R.4th 73].) The California Evidence Code is consistent with this formulation. Section 1291 provides in relevant part: “(a) Evidence of former testimony is not made inadmissible by the hearsay rule if the declarant is unavailable as a witness and: . . . [f] (2) the party against whom the former testimony is oifered was a party to the action or proceeding in which the testimony was given and had the right and opportunity to cross-examine the declarant with an interest and motive similar to that which he has at the hearing.”
Here defendant did have the opportunity to cross-examine Sarah at the preliminary hearing, although his counsel elected not to question the girl at that time. This tactical decision may not be construed as a waiver of the right to confront the witness. In Barber v. Page, supra, 390 U.S. 719, the Supreme Court addressed a situation in which defense counsel had de[516]*516dined to cross-examine a preliminary hearing witness who, by the time of trial, had been imprisoned in a foreign jurisdiction. The court held that defendant had not waived the confrontation challenge by counsel’s failure to cross-examine the witness at the earlier opportunity, and offered the following observations regarding. the preliminary hearing procedure: “we would reach the same result on the facts of this case had petitioner’s counsel actually cross-examined [the witness] at the preliminary hearing. [Citation.] The right to confrontation is basically a trial right. It includes both the opportunity to cross-examine and the occasion for the jury to weigh the demeanor of the witness. A preliminary hearing is ordinarily a much less searching exploration into the merits of a case than a trial, simply because its function is the more limited one of determining whether probable cause exists to hold the accused for trial. While there may be some justification for holding that the opportunity for cross-examination of a witness at a preliminary hearing satisfies the demands of the confrontation clause where the witness is shown to be actually unavailable, this is not, as we have pointed out, such a case.” (Id. at pp. 725-726 [20 L.Ed.2d at p. 260].)
Although defense counsel had the opportunity to cross-examine Sarah at the preliminary hearing, the Evidence Code section 1291 exception to the rule against hearsay requires that the preliminary hearing witness be “unavailable” at the time of trial. Evidence Code section 240 in turn defines “unavailable as a witness” to include the witness who is “unable to attend or to testify at the hearing because of then existing physical or mental illness or infirmity.” (Evid. Code, § 240, subd. (a)(3).) The burden of proof on the issue of witness unavailability rests with the proponent of the evidence, and the showing must be made by competent evidence. (People v. Williams (1979) 93 Cal.App.3d 40, 51 [155 Cal.Rptr. 414]; People v. Enriquez, supra, 19 Cal.3d at p. 235.)
Here the only witness at the unavailability hearing was Sarah’s mother, defendant’s wife. She testified that Sarah had been suffering for some time from emotional difficulties, that she had experienced audio and visual hallucinations, that she had intentionally cut herself several times— most recently two days prior to the hearing—and that she had just been hospitalized following this latest incident. Although the court invited the prosecuting attorney to come forward at a later time with medical testimony in support of the mother’s declarations, no physician, psychiatrist, or psychologist ever testified on the unavailability issue.
We conclude that Sarah’s mother’s testimony on the issue of her daughter’s mental health is legally insufficient to support a finding of witness unavailability. Reviewing courts have typically and properly required either [517]*517expert testimony on the witness’s present condition, or the witness’s own express refusal to testify at trial.
For example, in People v. Gomez (1972) 26 Cal.App.3d 225 [103 Cal.Rptr. 80], the Court of Appeal based its decision sustaining a finding of unavailability on the testimony of two physicians, both staff psychiatrists at the hospital where the witness, the victim of the alleged sex crime, was then under treatment. (Id. at p. 228.) The court distinguished cases relied on by the defendant in which police officers had testified regarding the mental condition of the absent witness, but in which no medical testimony had been presented. (Id. at p. 230, citing Sanchez v. Bagues & Sons Mortuaries (1969) 271 Cal.App.2d 188, 193-194 [76 Cal.Rptr. 372].) While refusing to specify the actual nature or severity of the disability required for a finding of unavailability, the court did offer the following general standard: “We agree with appellant that the illness or infirmity must be of comparative severity; it must exist to such a degree as to render the witness’s attendance, or his testifying, relatively impossible and not merely inconvenient. ” (Italics added.) (26 Cal.App.3d at p. 230.)
In People v. Williams (1979) 93 Cal.App.3d 40 [155 Cal.Rptr. 414], the Court of Appeal rejected a finding that a rape victim who had testified at the defendants’ first trial was unavailable to testify at the second trial of a defendant. At the unavailability hearing several persons were called, including the judge from the first trial. He testified that the witness had been under severe emotional strain during the proceedings and on one occasion had collapsed. Although he also stated that the physician who had examined the witness had testified that she was then suffering from colitis, no such evidence was offered at the unavailability hearing to show that this condition had persisted or recurred. The police officer who had testified at the first trial reiterated his description of the witness’s physicial condition immediately after the sexual attacks. The witness’s girl friend, who had accompanied the witness to court during the first trial, testified that the witness had undergone extreme emotional and physical distress and, on learning of the new trial, had told her “Barbara, I can’t go through with that again. I can’t. I want to put my life together and go on. I am being torn apart.” (Id. at p. 50.) Finally, the witness’s former boyfriend testified that she “had become hysterical” on learning that she would have to testify again. (Ibid.)
However, no medical evidence was presented in Williams regarding the witness’s alleged “mental infirmity” at the time of the hearing. The court concluded that “In the absence of medical testimony such as was introduced in Gomez, there was no credible evidence ... to support a finding that, if required to testify [the witness] would suffer any substantial impairment to her mental or physical health—either permanently or for any significant [518]*518period of time.” (Id. at p. 54.) Adopting the test articulated in Gomez, the court held that the evidence established only that the absent witness’s “mental, emotional and physical condition rendered her ability to testify merely inconvenient and not ‘relatively impossible.’” (Ibid.)
As respondent here points out, neither the Gomez nor the Williams decision explicitly calls for medical testimony as a specific requirement for a finding of unavailability under the “mental illness or infirmity” provision of Evidence Code section 240. Indeed there is competent evidence in this case, from Sarah’s mother, that Sarah was hospitalized and undergoing psychotherapy at the time of the hearing. However, although the mother may be capable of testifying to the existence of her daughter’s mental illness or infirmity, section 240, subdivision (a)(3), requires that the witness be “unable to attend or to testify at the hearing because of then existing . . . mental illness or infirmity.” (Italics added.) In other words, the evidence must establish not only that there is an illness or infirmity, but also that the condition renders it “relatively impossible”—to use the language of Williams—for the witness to appear. This determination generally calls for expert opinion, with supporting reasons, as to the likely effect of the court appearance on the physical or mental health of the witness. The mother’s understandably protective testimony was therefore insufficient as a matter of law for this critical purpose.
As an alternative ground for sustaining the finding that Sarah was unavailable as a witness, respondent calls our attention to People v. Rojas (1975) 15 Cal.3d 540 [125 Cal.Rptr. 357, 542 P.2d 229, 92 A.L.R.3d 1127]. In Rojas the witness himself testified at the hearing on his unavailability and stated that he had received threats by telephone and by letter, that various acts of violence had been directed toward him and his family, and that he was therefore in desperate fear of his and his family’s safety. (Id. at pp. 547-550.) He did not invoke the privilege against self-incrimination, and was physically and mentally capable of testifying, yet he absolutely refused to testify. He was found in contempt of court and was detained in juvenile hall for the remainder of the trial. Although no medical testimony was presented at the unavailability hearing, we held that the term “mental infirmity” includes a mental state induced by fear that impels a witness to refuse to testify. (Id. at p. 551; accord, People v. Quaintance (1978) 86 Cal.App.3d 594, 600 [150 Cal.Rptr. 281].)
The record here indicates that Sarah spent one night in juvenile hall for refusing to cooperate at the preliminary hearing and for repeatedly answering “I don’t know” or “I don’t remember” when questioned by the prosecuting attorney. Respondent therefore contends that the situation is analogous to Rojas, and argues that Sarah has “constructively” refused to testify [519]*519for fear of returning to juvenile hall. This contention is without merit, and Rojas is clearly distinguishable.
In Rojas the witness who had testified at the preliminary hearing and at the first trial himself appeared at the beginning of the second trial and stated that he would refuse to testify out of fear. The court was therefore able to observe his demeanor to determine whether his fear amounted to a mental infirmity that would render it “relatively impossible” for him to testify. Sarah, however, did not testify at the hearing on her unavailability. The prosecution offered only her mother’s testimony that Sarah was afraid that if she took the stand she might again be sent to juvenile hall for failing to answer questions properly.3 While in Rojas there was “no challenge made to the sufficiency of the evidence supporting the trial court’s determination that the fear was justified” (People v. Rojas, supra, 15 Cal.3d at p. 550), here it is the very sufficiency of the mother’s testimony, supported neither by expert evidence nor by Sarah’s own words, that is at issue. In Rojas we were asked to determine whether fear might constitute a mental infirmity for purposes of witness unavailability; here Sarah’s subjective fear has not been established by the evidence.
Furthermore, although in Rojas we held that fear may constitute a mental infirmity rendering a witness unavailable, we also compared the witness who refuses to testify with the witness who invokes the privilege against self-incrimination or who is physically absent from the hearing: “No [] [sufficient] reason appears to us why the former testimony of a witness who is present in court but refuses to testify because he is in fear . . . should not be used when that of a witness, who claims privilege or who is absent from the hearing and his attendance cannot be compelled or procured, can be used.” (Id. at p. 551.) No such analogy to the privilege exemption or to physical absence is possible in this case, because Sarah herself did not personally refuse to testify.
We conclude that Sarah should not have been declared unavailable as a witness under Evidence Code section 240, subdivision (a)(3). Her preliminary hearing testimony was therefore erroneously admitted into evidence, as was Deputy Hoberg’s testimony purporting to impeach Sarah’s declarations. Defendant was thus denied the constitutional right to confront the principal witness against him, the alleged victim of the offenses charged, while respondent was permitted to impeach that witness’s former testimony.
[520]*520Because this error is of federal as well as state constitutional dimension, violating as it does the defendant’s right to confrontation guaranteed by the Sixth Amendment to the United States Constitution, we must apply the reversible error test set out in Chapman v. California (1967) 386 U.S. 18 [17 L.Ed.2d 705, 87 S.Ct. 824, 24 A.L.R.3d 1065]. (People v. Arcega (1982) 32 Cal.3d 504, 525 [186 Cal.Rptr. 94, 651 P.2d 338].) This test provides that “before a federal constitutional error can be held harmless, the court must be able to declare a belief that it was harmless beyond a reasonable doubt.” (386 U.S. at p. 24 [17 L.Ed.2d at pp. 710-711].) The burden is on the beneficiary of the error “either to prove that there was no injury or to suffer a reversal of his erroneously obtained judgment. ” (Ibid.; see also People v. Spencer (1967) 66 Cal.2d 158, 168 [57 Cal.Rptr. 163, 424 P.2d 715].) The Chapman court reiterated the approach it adopted in Fahy v. Connecticut (1963) 375 U.S. 85 [11 L.Ed.2d 171, 84 S.Ct. 229]: “The question is whether there is a reasonable possibility that the evidence complained of might have contributed to the conviction.” (Id. at pp. 86-87 [11 L.Ed.2d at p. 173]; see also People v. Powell (1967) 67 Cal.2d 32, 56-57 [59 Cal.Rptr. 817, 429 P.2d 137].) Where a fundamental constitutional right is at issue, erroneous evidentiary rulings are seldom harmless under this standard: “An error in admitting plainly relevant evidence which possibly influenced the jury adversely to a litigant cannot ... be conceived of as harmless.” (Chapman v. California, supra, 386 U.S. at pp. 23-24 [17 L.Ed.2d at p. 710].) Further, we must weigh the impact of the error not only on the decision of the jury, but also on the course of the trial. (People v. Spencer, supra, 66 Cal.2d at p. 163.)
It is surely “reasonably possible” in this case that Sarah’s preliminary hearing testimony influenced the jury in its decision to convict defendant of the one felony and several misdemeanor charges. Viewing the record without Sarah’s former testimony we are left with very little evidence. We exclude Deputy Hoberg’s testimony in this analysis because it was introduced solely for impeachment. Because we have determined that Dr. Walker’s testimony was erroneously admitted in violation of the psychotherapist-patient privilege, we view the record absent this testimony as well. We are left, then, with only the testimony of Sarah’s mother, defendant’s wife, who contributed negligibly to the prosecution’s case. She testified at some length as to Sarah’s personality and described her mental and emotional problems. When asked if she had discussed with her husband his sexual relations with Sarah she responded, reluctantly: “All right. It was something of the nature that Sarah had told me that some things had occurred between she and my husband, also between she and a shepherd and between she and a neighbor boy and that I had taken Sarah to see Dr. Walker and that I felt that we had a problem which had to be faced. Whether or not what Sarah said was the truth was a problem in our household.” Attempts to elicit more specific [521]*521testimony as to the nature of defendant’s sexual relations were unsuccessful. It is therefore probable that the jury would not have convicted defendant of any crime based on her testimony alone.
Looking at Sarah’s preliminary hearing testimony itself to evaluate its possible injurious effect on defendant’s case, we are persuaded that her statements, however contradictory, reluctant or confused, did contribute to the conviction. While she was often evasive and uncertain in her responses, she conceded that a number of mutual fondling episodes had taken place and, if not in explicit language, suggested that an alleged oral copulation incident had occurred, described an incident of masturbation and ejaculation, and stated that one or two incidents involved the use of a vibrator. Although Sarah’s credibility may have been questioned by the jury, her preliminary hearing testimony undoubtedly contributed to defendant’s detriment.
Considering the effect of the error on the course of defendant’s trial, we note again that but for the admission into evidence of Sarah’s former testimony, Deputy Hoberg would not have been permitted to testify to impeach this testimony. That witness would have been limited to testimony regarding her own observations of Sarah during their interview, and to disclosing whatever of Sarah’s statements might have been admissible under some other exception to the rule against hearsay, if any. Thus we are unable to conclude that the erroneous admission into evidence of Sarah’s preliminary hearing testimony was harmless beyond a reasonable doubt.
The judgment is reversed.
Bird, C. J., Broussard, J., Reynoso, J., and Grodin, J., concurred.