Appellant was convicted of two counts of rape in the first degree, one count of sodomy in the first degree and two counts of sexual abuse in the first degree. He was sentenced to consecutive terms totaling 85 years. The victims were his step-daughters and on this appeal, appellant claims to have been denied a fair trial by virtue of numerous trial errors.
Appellant and the child victims’ mother were married for about two years between 1985 and 1987. After their divorce, the mother made a report to Social Services that she suspected appellant of having sexually abused the children during the marriage. An investigation was commenced, without appellant's knowledge, which resulted in a determination in October, 1987, by the social worker in charge and the police that there was no evidence to indicate sexual abuse. At that time, the children were 7½ and 6½ years old.
In 1990, the children’s mother again contacted Social Services and requested an investigation of the 1985-87 allegations. This request appears to have been motivated, in part, by the mother’s belief that appellant was molesting their natural child during visitation. In response to this, the child victims were interviewed by the social worker, Ms. Handorf, and examined by a physician and a psychiatrist. As a result of the renewed investigation, appellant was indicted.
Of the multiple issues raised, appellant has placed his greatest emphasis on the contention that inadmissible hearsay evidence and opinion testimony were admitted over his objection. The witnesses whose testimony is challenged are the social worker and the psychiatrist and we are thus required to revisit one of the most vexatious questions of law to come before this Court in the last decade: whether and to what extent social workers and physicians may repeat the out-of-court statements made to them by children who may be sex abuse victims.
Testifying on behalf of the Commonwealth was a psychiatrist, Dr. David Hilton. Dr. Hilton indicated generally that he saw the children to evaluate, diagnose, and treat them, but from the substance of his testimony, it is clear that he was not a treating physician as that term is normally understood. He was hired by Social Services for the purpose of evaluation, and any treatment recommendation he made was incidental to his primary duty. Dr. Hilton succinctly described his role as follows: “The principal reason for my seeing the children was not treatment. The principal reason was to determine whether there was a reason for treatment.” Dr. Hilton’s testimony consists of more than thirty pages in the transcript and it would unnecessarily lengthen this opinion to review its entirety. It is sufficient to say that, over appellant’s objections, Dr. Hilton was allowed to repeat what the children told him appellant had done to them and describe their use of anatomically correct dolls to portray what appellant had done. Though not exhaustive, a fair sampling of Dr. Hilton’s testimony is as follows:
“She specifically initially said that Scotty ‘kissed me around my front and private parts’. When I asked, ‘Did he touch your private parts?’ she said yes. She then indicated that he touched her with his fingers, both in the behind, the anal region and also what she called the moosh, the vaginal region. She became very ill at ease, very hesitant to speak, very sad looking, when we were talking about this. She then went on to indicate that he also touched her in the same way using his penis, which she described as a ding dong, in the vaginal region. At this point in time, she also was given the opportunity to use the dolls. She used the adult male doll and the female child doll, and then simulated these activities, showing the male doll touching the female doll with the penis in the vaginal region and then the last comment that she made regarding activities that allegedly occurred was a quote. She said, ‘He tongued me all over my body.’ Meaning that he kissed or licked, as I interpreted it.”
*544Prior to this Court’s decision in Drumm v. Commonwealth, Ky., 783 S.W.2d 380 (1990), there would have been little necessity for a detailed analysis of the challenged testimony. By any reasonable accounting, Dr. Hilton was not a treating physician, his testimony disclosed information which had little or no relevance to treatment, his examination was remote in time from the event which may have given rise to the need for treatment, and it was highly prejudicial. With our decision in Drumm, however, a new day dawned and the answer is not so clear. Drumm began by declaring that “[t]he time has come to expand the hearsay exception for statements for purposes of medical diagnosis or treatment to conform to the modern approach as represented by Rule 803(4) of the Federal Rules of Evidence.” Drumm, at 384. See KRE 803(4), effective July 1,1992. We acknowledged that the new rule “blurs the distinction between treating and testifying physicians,.... ” but made it clear that the distinction had not been abolished. Drumm recognized the desirability of a more flexible approach to the determination of whether such evidence should be admitted or excluded but acknowledged that statements made to a physician who is consulted for a purpose other than treatment have less inherent reliability than such statements made to a treating physician.1 As such, we directed application of a prejudicial effect versus probative value analysis. The rule which now prevails requires a court to begin with the view that statements made to a physician who lacks treatment responsibility have less inherent reliability than traditional patient history. Then the court must decide whether from the totality of the circumstances the probative2 value of the evidence outweighs its prejudicial effect.
Within the last year, this Court has confronted the new rule in two similar cases and it may be helpful to briefly review our holdings. In Jones v. Commonwealth, Ky., 833 S.W.2d 839 (1992), the victim was a three-year-old girl who was taken to the emergency room after a relative discovered redness about the child’s genital and anal areas. On examination and testing, the doctor diagnosed chlamydia, a sexually transmitted disease. Based on the observation of redness close in time to the alleged occurrence and the diagnosis of the disease, the Court found no error in admission of the child’s statements to the doctor as to the acts and identity of the perpetrator. Similarly, in Edwards v. Commonwealth, Ky., 833 S.W.2d 842 (1992), a five-year-old boy was examined by a physician on suspicion of sexual abuse some weeks after the occurrence of the event. On finding a wound on the child’s perianal skin which was neither fresh nor fully healed, and diagnosing the sexually transmitted disease chlamydia, the doctor was permitted to testify that the child identified the act and the person who did it. This Court affirmed the conviction.
Our decisions in Jones and Edwards disclose little discomfort when other indicia of reliability are present. Particularly is this so when the trial court appears to have fully considered the issue as in Edwards and heard the objection in advance of trial *545and rendered thorough findings and conclusions.
When the hearsay testimony of Dr. Hilton is measured against the standard described hereinabove, it clearly falls short of the requirements. It had diminished reliability by virtue of the origin and the otherwise limited relationship between the physician and the children. Moreover, since the children themselves testified effectively, the physician testimony added little of a probative nature, but by virtue of the aura which attends such testimony, the prejudicial effect was significant. We conclude that the trial court committed reversible error with the admission of Dr. Hilton’s hearsay testimony.
We now turn to appellant’s contention that the hearsay testimony of social worker Linda Handorf was erroneously admitted. After receiving a complaint from the children’s mother in 1990 relating to the 1985-87 allegations, Ms. Handorf and police detective Stucker commenced an investigation by interviewing the children. The principal technique utilized in the interviews was to furnish the children with outlines of male and female figures of various sizes and ages and suggest that the children draw upon the outlines or draw their own figures on blank paper to illustrate what had happened to them. The drawings were then made a part of the departmental records. The court subsequently admitted several of the drawings made by each child as Commonwealth’s Exhibits 2 and 3. In response to questioning and by referring to the drawings, Ms. Handorf was permitted to testify in part as follows:
“The next drawing she did for us is after the point where she was telling us of some sexual touching, and she put X’s on the places, on the body of the person that she identified as Scotty, the places there he had touched her with, and she placed X’s on his hand and on his penis. The next picture that she chose was a picture of a little girl that she chose to represent her, and she marked an X on the vaginal area, which she marked meaning that’s what was touched with his hand, and an 0 which she marked in a larger area surrounding the vaginal area, around the thigh area, as what he touched with what she called ‘his thing’, which she identified to us as the male penis. She then chose a rear view of a child to represent herself and marked it with an X and said this had been touched with Scotty’s hand.... We had [A.S.] select a picture to represent herself, and she chose this little girl, and we asked her to mark where she was touched. She drew a circle around the vaginal area and said that’s where he, Scotty, had touched her with his ‘ding dong’, which is how she referred to the male penis. She also did another picture which she marked with X’s up around the breast area and vaginal area, for where Scotty touched her with his hands, and O’s where she marked on the breasts, on the navel area, on the arms, on the vaginal area, and the lower stomach area, on the hand, and even on the legs, where he kissed her. She chose this picture to represent Scotty, and she placed X’s where she touched him and she marked arms, chests — _
And then she drew an 0 around the part of the body which she said he used to touch her, and she drew a circle around the penis and she drew a circle around the mouth, and a circle around the hand.”
As the quotations hereinabove well illustrate, Ms. Handorf extensively repeated the children’s out-of-court statements and described their out-of-court actions which identified appellant as their abuser. In addition, Ms. Handorf’s testimony contains extensive conclusions as to the meaning of the children’s acts and statements.
In Bussey v. Commonwealth, Ky., 697 S.W.2d 139 (1985), this Court refused to apply the so-called residual exception to the hearsay rule as recognized by the Federal Rules of Evidence to the testimony of a social worker. Despite the refusal of the victim to testify, a fact not present here, we held that the testimony was inadmissible hearsay. In Souder v. Commonwealth, Ky., 719 S.W.2d 730 (1986), nonverbal conduct such as demonstrations by children with anatomically correct dolls was held to be embraced within the hearsay rule and no exception was recognized for *546the results of investigations of social workers. Recently, however, in Hellstrom v. Commonwealth, Ky., 825 S.W.2d 612 (1992), four Justices of this Court found no error in the admission of an exhibit which was a drawing with contemporaneous captions made by the victim in the presence of her mother and a police officer. The majority recognized that the writings were out-of-court statements offered for their truth and thus hearsay, but by virtue of their in-court identification by the child who was subject to cross-examination, the Court characterized the error, if any, as harmless. While Hellstrom may be a modest departure from the rigid holdings of Bussey and Souder, by virtue of the narrow majority, the recognition of harmless error, and the significant difference in the quantity and tenor of the exhibits and testimony, Hellstrom provides dubious support for the rulings of the trial court with respect to the testimony of Ms. Handorf.
As the foregoing and other decisions of this Court, which include Lantrip v. Commonwealth, Ky., 713 S.W.2d 816 (1986); Hester v. Commonwealth, Ky., 734 S.W.2d 457 (1987); Mitchell v. Commonwealth, Ky., 777 S.W.2d 930 (1989); and Brown v. Commonwealth, Ky., 812 S.W.2d 502 (1991), well illustrate, this Court has demonstrated its discomfort with convictions for child abuse based upon the hearsay testimony and ultimate fact opinion given by social workers. There may be a temptation among judges to let pity for small children who may have been victimized by vicious adults overcome their duty to enforce the rules of evidence. This temptation is heightened by the frequent inability of children to effectively tell in court what happened to them. Likewise, the social workers of this Commonwealth are often called upon to aid helpless victims and by virtue of their compassion and the intimate knowledge they acquire of the circumstances, they become, in effect, advocates for the child victim. Judges must guard against the temptation, whether personal or of others, to abandon settled rules of evidence to accommodate their sense of justice in a particular case.
“The rules of evidence have evolved carefully and painstakingly over hundreds of years as the best system for arriving at the truth. They bring to the law its objectivity. Their purpose would be subverted if courts were permitted to disregard them at will because of an intuitive perception that to do so will produce a better result in the case at hand. We accept the premise that obeying these rules is the best way to produce evidence of a quality likely to produce a just result. We reject the notion of different rules for different cases because one or the other of the litigants insists that a different ruling will produce a better result in his particular case. Changes in the rules are both inevitable and desirable, but changes should occur only after a judicious Darwinian process.” Fisher v. Duckworth, Ky., 738 S.W.2d 810, 813 (1987).
We reiterate the view expressed in Souder v. Commonwealth, supra, “There is no recognized exception to the hearsay rule for social workers or the results of their investigations.” Id. at 734. On this issue, we likewise reverse the trial court.
The next issue we feel obliged to address concerns the actions of a courtroom bystander who undertook to become a participant in the trial of this case. While the first child to testify was on the witness stand, Kathy Hess, a self-described friend of the family, was observed by another courtroom bystander making signals and gestures to the child witness. When this fact was brought to the attention of the trial court, appellant objected and moved for a mistrial. On the motion, the trial court convened a hearing outside the presence of the jury. The facts which emerged were in substance as follows: The bystander who observed Ms. Hess was convinced that the gestures told the child whether to answer yes or no to the questions asked and whether she approved of the child’s answers. Ms. Hess denied suggesting any answers to the child, but admitted that by gestures she sought to comfort and encourage the child during her testimony. She admitted to mouthing “You’re doing fine” and admitted that she gave approving gestures by winks and a *547thumbs-up sign. During the hearing, Ms. Hess also disclosed that she had informed one or more separated witnesses as to certain portions of the child’s testimony and had entered the room where separated witnesses were being kept. While most of the witnesses had already testified, they were subject to recall by appellant. After expressing annoyance with Ms. Hess and considering whether to grant a mistrial, the trial court overruled the motion.
“It’s causing me problems, and if I let it in, it very may well cause a reversal. When they read all this, I’m not so sure, and I’m going to have to think about it more, but I am going to let the trial continue for the time being.”
While the trial court did not render findings of fact as such, it is undisputed that the bystander gestured to the child witness during her testimony and communicated the substance of some testimony from the courtroom to one or more separated witnesses.
During the trial of many cases, unexpected and improper events transpire. The atmosphere in a courtroom is usually charged with partisanship and the partisans, whether lawyers, parties or spectators, frequently transgress rules of strictly proper behavior. Ordinarily, the question for the court when faced with a motion for mistrial is whether the impropriety would likely influence the jury. As it is impossible to catalog the myriad occurrences which might provoke a motion for mistrial, courts generally hold that the trial court is vested with broad discretion to determine whether a mistrial is necessary upon the occurrence of courtroom misconduct. Jones v. Commonwealth, Ky., 628 S.W.2d 226 (1981). However, that discretion is not unlimited as illustrated in Raney v. Commonwealth, 287 Ky. 492, 153 S.W.2d 935 (1941), in which this Court reversed a conviction in part upon improper conduct by spectators and its presumed effect upon the jury. We also reversed in Ballard v. Commonwealth, Ky., 743 S.W.2d 21 (1988), because the trial court failed to conduct a hearing prior to holding that a witness in violation of a separation order could not testify.
While the parties have not cited a case which directly addresses the issue here, we believe common sense dictates the result. The child victim on the witness stand was a crucial witness for the Commonwealth. Her demeanor during testimony and her ability to withstand cross-examination inevitably influenced the jury as to whether or to what extent she should be believed. Even taking Ms. Hess's version of what she did at face value, the witness received encouragement, approval and comfort at the time her credibility was being assessed by the jury. It would be impossible to say that the witness did not derive confidence and assurance from this positive reinforcement which influenced the jury to believe her. Whether Ms. Hess’s contact with the separated witnesses affected the trial can never be known. After learning that these witnesses had been apprised as to testimony from the courtroom, defense counsel did not recall them to the stand.
We reiterate that the trial court has broad discretion to determine whether a violation of proper courtroom conduct requires a mistrial. Jones v. Commonwealth, supra. We also recognize the natural reluctance to grant a mistrial after substantial time and resources have been expended. Nevertheless, we regard the violations here as so egregious and inimical to the concept of a fair trial that they cannot be disregarded in the name of trial court discretion.
The final issue we will decide concerns appellant’s motion for a directed verdict with respect to the charge that he raped the younger child, N.S. He maintains there was no evidence presented from which the jury could have believed beyond a reasonable doubt that penetration occurred. KRS 510.010(8).
The best evidence for the Commonwealth was the testimony of N.S. wherein she said and demonstrated that appellant touched her “middle part” with his “middle part” or “private thing” and that the act “hurt.” In addition, records of a physician’s examination of the child conducted several years after the alleged offense “indicated a penetrating injury of the hymenal membrane *548occurring in the distant past.” It should be recalled that at the time the crimes were alleged to have occurred, N.S. was between 4½ and 6V2 years old. While such evidence is slight, it was sufficient. A mere touching in ordinary experience would not have been sufficient to produce pain, even for a small child. Her testimony, which described genital touching and resulting pain, taken together with the medical evidence, was sufficient to permit a belief beyond a reasonable doubt that penetration occurred. Particularly is this so when we apply the rule set forth in Commonwealth v. Benham, Ky., 816 S.W.2d 186 (1991), that all fair and reasonable inferences from the evidence must be assessed in favor of the Commonwealth.
Appellant has also made a substantial argument with respect to the erroneous admission of evidence of uncharged crimes allegedly committed by him. There was evidence that appellant showed hard-core pornographic books and magazines to the children, that such publications depicted deviate sexual practices, that he beat the children with a paddle, and that he smothered one of the children with a pillow until her nose bled. The Commonwealth maintains that this issue is unpreserved, but in any event, there was no error.
In recent years this Court has decided a number of cases which deal with the admissibility of uncharged criminal acts or acts of misconduct. The general rule as set forth in O’Bryan v. Commonwealth, Ky., 634 S.W.2d 153 (1982) is that
“Evidence of the commission of other crimes, generally speaking, is not admissible to prove that an accused is a person of criminal disposition.” Id. at 156.
O’Bryan recognizes, however, that as with many rules of law, there are exceptions to this one. Inasmuch as this Court has reversed appellant’s conviction for retrial and this issue may not be preserved, we need not decide whether admission of the complained of evidence of uncharged acts was reversible error. We simply say that the admission of such evidence, unless it is admitted strictly in accordance with an exception'to the general rule, may imperil any subsequent conviction of appellant.
The remainder of appellant’s claims of error have been reviewed and determined to be without merit, harmless, unpreserved, or unlikely to recur upon retrial.
For the foregoing reasons, appellant’s convictions are reversed and this case remanded for further consistent proceedings.
STEPHENS, C.J., and COMBS, LEIBSON and REYNOLDS, JJ., concur. SPAIN, J., dissents by separate opinion in which WINTERSHEIMER, J., joins.. Justice Vance succinctly articulated the common law reasons for distinguishing between the testimony of a treating and a non-treating physician:
"A statement to a treating physician to make a diagnosis and to give proper treatment by a person who wants to get well and knows that his recovery may well depend on his telling the truth to his doctor, has an inherent indication of reliability. It is so likely that one would tell the truth under such circumstances that the statement is sufficiently trustworthy to be admitted into evidence. The same is not true, however, when the physician wears the hat of a paid investigator who has been consulted to testify at trial. The statements made to him are not made under any urgency to tell the truth because the likelihood of recovery of one’s health does not depend upon it. There may be cases where the out-of-court statements to the physician employed to testify were precipitated by the basest of motives, including a possible motive of an adult to coach an infant to make statements designed to secure the conviction of a particular accused person.” Drumm v. Commonwealth, Ky., 783 S.W.2d 380, 386 (1990), Vance, J., dissenting.
. "Probative evidence is testimony carrying the quality of proof and having fitness to induce conviction of truth." Globe Indemnity Co. v. Daviess, 243 Ky. 356, 47 S.W.2d 990, 992 (1932).