SCHUMAN, S. J.
Defendant was charged with one count of first-degree sexual abuse, ORS 163.427, and two counts of attempted first-degree sexual abuse, arising out of his alleged contact with the victims, J and K. In a motion in limine, defendant sought to exclude the testimony of experts from Kid’s Intervention and Diagnostic Service Center (KIDS Center) who had interviewed J and K. According to defendant, one witness’s testimony included a diagnosis of sexual abuse and was inadmissible under State v. Southard, 347 Or 127, 140-41, 218 P3d 104 (2009), because it was scientific evidence that was highly prejudicial and only minimally probative. Other witnesses’ testimony was inadmissible, defendant maintains, because, although it did not directly vouch for the credibility of the victim, it amounted indirectly to the same thing, contrary to the rule in State v. Lupoli, 348 Or 346, 234 P3d 117 (2010). The trial court denied the motion. The court also denied defendant’s motion to sever and motion for a mistrial, in which he contended that trying the counts involving J and K together was prejudicial. Midway through a jury trial, the state dismissed the charges relating to J. Defendant was subsequently convicted of one count of first-degree sexual abuse involving K. He appeals, renewing the arguments that he made below.1 We conclude that the trial court did not err, and we therefore affirm.
The relevant facts are not in dispute. J and K, ten-year-old girls, were babysitting for defendant’s son and were asleep on the floor of defendant’s house when defendant and his wife returned home. K awoke to find defendant kneeling down almost on top of her with his hand down her pants. K heard a door open, and defendant left to join his wife in another room. K sat up, and defendant came back into the room. K and defendant engaged in a brief conversation, and defendant left the room again. K woke J and told her that defendant had been touching her in her “private area” and that she wanted to go home. J’s father picked up the girls, [839]*839who told him of defendant’s conduct; the father then took them to a police station, where they were interviewed. He then took them to a hospital for an evaluation.
At that time, K was examined by White, a sexual-assault nurse. White testified at trial that K had reported “rubbing of her genitalia area,” and that, consistent with what K had reported, White’s examination of K revealed an “acute injury” in the genital area — abrasions, redness, and swelling that one would not expect to see from typical wiping or cleaning. White described “increased redness and swelling of the clitoris *** and at the labial folds on the upper portion,” and “increased redness in the upper and lower” parts of K’s hymen, “which is also consistent with some type of rubbing or blunt trauma.” White further testified that K’s labia “appeared abraded, as if there’s been some shaking, some rubbing.” Her written report also stated those findings.
Six weeks later, K was interviewed by Zancanella, a staff person from KIDS Center. Zancanella testified at trial about what KIDS Center looks for in evaluating a child for sexual abuse. Zancanella explained that she looks at whether the child provides information in narrative form and spontaneously. She explained that, when the child can do that, “I’m going to get more details and I’m going to get more accurate information.” Zancanella further testified that she looks at whether the child has consistently reported the “core details” to police, medical personnel, and the KIDS Center interviewer. Zancanella explained that, when a child is able to give more details, including sensory details, “there’s a very low percentage of making stories up ***. [T]hat tells us that there’s some additional validity.”
Glesne, also of KIDS Center, interviewed K, and described at trial the observations relating to K that she considered significant:
“[T]he core details of the information [K] provided [to KIDS Center] were consistent with what she had provided to * * * the sexual assault nurse examiner and the Redmond Police Department. They were also consistent with what had been provided by another child that had been interviewed previously at the KIDS Center.
[840]*840“[K] used both words, her body, a drawing, to describe what had happened in relating the disclosure that she made at the KIDS Center.”
K was also examined by Dr. Kyriakos of KIDS Center. At trial, Kyriakos described what KIDS Center looks for in evaluating a child for sexual abuse. Like Glesne and Zancanella, Kyriakos testified that KIDS Center is “looking to see if there is a physical sign” and “the consistency of the core details that the child has given over time.” Kyriakos explained that she looks to see whether the child “was able to give multiple details” and if the “child was able to give information in more than one * * * media form.” She testified:
“Q. When you’re evaluating a child for any kind of abuse, what is it that — what is it that you’re looking for specifically with regard to child sexual abuse, or the criteria that you’re looking for?
“A. Well, in the examination you’re looking to see if there is a physical sign, something that you see on examination, that would be diagnostic of abuse. The other things that we’re looking at is if the history that we’ve gathered supports a diagnosis of child sexual abuse.
“So we’re looking at a lot of things, but that — some of the main categories are looking at consistency of the core details that the child has given over time. So we compare the core details of what was said previously to law enforcement, or maybe to a parent, maybe to DHS, and then to us during the interview and examination.
“We look to see if the child was able to give multiple details, so — so a lot of depth in detail; talking about who, what, where, when, sometimes how. Are they able to provide other contextual details, like describing furniture in the room or clothes that were being worn? So some of the other details during the incident.
“One thing we look at is if a person — child was able to give information in more than one — we call it more than one media form. And those mean things like able to give descriptions verbally, able to give description by drawing, able to communicate by using the child’s body as a reference.
“One — another thing is how the disclosure was made. So was it — disclosure spontaneous or not, which is — we weigh that heavily, because ‘spontaneous’ means that the [841]*841child made the disclosure without having to be asked about it. So it wasn’t something an adult came and asked the child about; something that they spontaneously provided without being asked.
“Another thing that — that weighs into, if the child’s able to give any kind of very specific details on — on sensory-type information. So something that they felt would be an example. And those type of details are important because sensory details usually are only something you know and if the child actually experienced the actual event.
“Behavioral changes; we look at that, too. Behavioral changes are not diagnostic of abuse, because the changes can be due to other stressors in the child’s life. But if those behavioral changes are a change from what the child’s baseline is, then that’s very concerning for abuse.”
Kyriakos then testified that the core details that K described were consistent with what Kyriakos had previously read in the reports and what K’s mother had reported. Kyriakos also noted that K described sensory details, like “stinging,” gave information using more than one media form, and used her body as a reference. Kyriakos testified that K’s comment about experiencing a stinging sensation
“was important to me, because when we get those kinds of sensory motor details *** it weighs into our decision-making, because * * * it’s always likely the kid will describe a specific sensation if they’ve actually experienced themselves. So *** her describing stinging was important to me.”
Kyriakos testified that, although she found no physical signs of abuse when she examined K six weeks after the alleged abuse, that was to be expected. Unlike the other KIDS Center witnesses, who had testified as to what characteristics might indicate credibility and then also testified that K showed those characteristics — so called indirect vouching— Kyriakos then offered the opinion that K had been sexually abused — a diagnosis of sexual abuse and, allegedly, direct vouching. As noted, a jury convicted defendant of a single count of sexual abuse.
We first address defendant’s argument that the court should have granted his motion in limine to exclude [842]*842Kyriakos’s diagnosis of sexual abuse. That argument is based on Southard, in which the Supreme Court held that
“a statement from an expert that, in the expert’s opinion, a child had been sexually abused was inadmissible in the absence of physical evidence of abuse, because it does not tell the jury anything that the jury could not have determined on its own, and, therefore, the probative value of any such testimony is outweighed by the danger of unfair prejudicial effect under OEC 403.”
Lupoli, 348 Or at 357 (summarizing the court’s holding in Southard, 347 Or at 142). Defendant notes that, although Kyriakos had reviewed White’s report before making her diagnosis, her written assessment did not refer explicitly to White’s physical findings or state that the physical findings were relevant to the diagnosis of abuse. Rather, according to defendant, Kyriakos explained that her conclusion was based on a review of K’s history and the statements K made during her evaluation at KIDS Center. Kyriakos’s written assessment emphasized that the core details provided by K were consistent; that she was able to rely on multiple methods of explaining what had happened to her; that her disclosure was spontaneous; and that she provided sensory details during the examination. In defendant’s view, the lack of explicit reliance on the physical evidence of abuse shows that Kyriakos’s determination was based primarily on an assessment of K’s credibility and therefore was inadmissible because the diagnosis’s probative value was outweighed by its potentially prejudicial effect; the diagnosis was of minimal probative value because it added little to what the jurors could have determined on their own, and was highly prejudicial because it came from an expert. According to defendant, the diagnosis was therefore inadmissible under OEC 403 and the Supreme Court’s application of that rule in Southard.
Defendant also asserts that the court’s holding in Southard is not limited to cases in which the diagnosis is based solely on the child’s statements, but also applies to any case in which the diagnosis of sexual abuse is based primarily on a credibility determination that is within the jury’s ability to evaluate rather than “on an abstruse matter of science.” Southard, 347 Or at 134-35. In this case, defendant [843]*843contends, to the extent that Kyriakos relied on White’s physical findings, those findings were not “diagnostic”; the primary bases for Kyriakos’s diagnosis were her assessments of K’s credibility, which the jury could have determined on its own. The abuse diagnosis, defendant asserts, is therefore inadmissible.
The state responds that, under Southard, the sexual abuse diagnosis is admissible because it is corroborated by physical findings of abuse. The state notes that, although Kyriakos did not specifically refer to the physical findings in describing the bases for her diagnosis, she did testify that the physical findings were corroborative of the alleged abuse. The state notes that, under Southard, the physical evidence of abuse need not be “diagnostic” (a term that neither Kyriakos nor defendant defines) nor conclusive, but only corroborative of the allegation of abuse. In this case, the state asserts, the physical evidence meets that test.
In State v. Ovendale, 253 Or App 620, 626, 292 P3d 579 (2012), rev den, 353 Or 714 (2013), decided after the briefing in this case, we rejected the defendant’s contention that Southard precludes the admission of any diagnosis of “sexual abuse,” whether or not the diagnosis is supported by physical evidence. We explained that Southard held that, in the absence of physical evidence of abuse, a diagnosis of sexual abuse is inadmissible under OEC 403. We further explained, however, that Southard did not also hold the converse — that when there is physical evidence of abuse, a diagnosis of sexual abuse is categorically admissible. Id. at 630. Rather, the admissibility of a diagnosis of sexual abuse continues to be subject to a balancing test under OEC 403, even when physical evidence of abuse is present. Id.
In Ovendale, we described three criteria to be considered in determining that the physical evidence is sufficiently corroborative, i.e., that there is “sufficient physical evidence so that the diagnosis of sexual abuse tells the jury something that it is not equally capable of determining on its own”:
“(1) the significance of that physical evidence is ‘the sort of complex factual determination that a lay person cannot make as well as an expert,’ (2) the physical evidence [844]*844is corroborative of the type of abuse actually alleged, and (3) the medical expert actually relies on the physical evidence in making a diagnosis of sexual abuse.”
253 Or App at 630. Here, we conclude that the reported physical evidence satisfies those criteria. Although White herself did not make a diagnosis of sexual abuse, she explained that her observations of K’s labia and clitoris — “increased redness and swelling of the clitoris *** and at the labial folds on the upper portion” — were consistent with the type of contact that K had reported. White reported that her observations of K’s hymen — increased redness in the upper and lower parts — were consistent with some type of rubbing or blunt trauma. Thus, White’s physical findings corroborated the alleged type of abuse.
Additionally, the significance of the physical findings — that they indicated abuse rather than other nonabusive contact — was a complex factual determination that an expert was better able to make than a layperson. Kyriakos testified that the abrasions described by White were caused by vigorous rubbing or by something with a sharp edge or a fingernail that would be unusual for a person to self-inflict through bathing or wiping.
Finally, although Kyriakos did not explicitly state that the diagnosis of sexual abuse was based on White’s physical findings, it is undisputed that Kyriakos had read White’s report before making the diagnosis. She also testified that she had taken note of White’s physical findings and that she had expected that the “acute injury” would have healed by the time of her own physical examination of K, some six weeks after White’s examination. Kyriakos testified, further, that the physical observations made by White were consistent with the abuse that K had described. We conclude that the evidence supports a reasonable inference that Kyriakos actually relied on White’s physical findings in making a diagnosis of sexual abuse.
Having determined that the physical evidence of abuse was sufficiently corroborative as described in Ovendale, we conclude that the trial court did not err in admitting Kyriakos’s diagnosis that K had been sexually abused.
[845]*845We turn to the assignment of error in which defendant contends that the trial court erred in denying his motion in limine to exclude the testimony of Zancanella, Glesne, and Kyriakos concerning what they look for in evaluating a child for sexual abuse, coupled with their descriptions of K’s characteristics in this case. Defendant contends that, by describing characteristics of a truth-teller and then testifying that K exhibited those characteristics, the testimony constituted inappropriate “indirect” vouching and should therefore have been excluded under Lupoli. Defendant cites Zancanella’s testimony that “I know that if [the child] can spontaneously produce [information] from what’s called recall memory, I’m going to get more details and I’m going to get more accurate information.” Defendant also refers to Zancanella’s testimony that “the more [a child] can provide those kinds of sensory details,” “that tells us that there’s some additional validity.” Defendant cites Glesne’s testimony that the core details of information provided by K to KIDS Center were consistent with what she had told White and the police department and with what had been reported by another child. Glesne also testified that K had used multiple media — “words, her body, a drawing” — to describe what had happened to her. Kyriakos described K’s comments about stinging to be important “because when we get those kinds of sensory motor details * * * it weighs into our decision-making, because * * * it’s always likely the kids will describe a specific sensation if they’ve actually experienced themselves.”
The state acknowledges that, under Lupoli, when a diagnosis of sexual abuse is inadmissible, testimony concerning the “subsidiary principles” on which the diagnosis is based — that is, general statements derived from social science or other sources — is inextricably bound up with that inadmissible diagnosis. When, however, as here, the diagnosis is admissible and supported by physical evidence of abuse, the state contends that Lupoli allows an expert witness to testify not only about subsidiary principles in general, but about the victim’s specific characteristics relative to those subsidiary principles, including those that reflect on the validity of the complaint. As the state understands Lupoli, then, short of a direct comment on credibility, [846]*846credibility-directed subsidiary principles must be excluded only when there has been a Southard-type error, i.e., only when a diagnosis of sexual abuse has been admitted without corroborating physical evidence of abuse.
Although we acknowledge that the Supreme Court’s treatment of the relationship between unduly prejudicial testimony in Southard and inadmissible vouching in Lupoli is obscure, we nonetheless agree with the state’s understanding oí Lupoli. In that case, the defendant did not object to the admissibility of the diagnosis of sexual abuse, even though there was no physical evidence. The defendant did, however, object to the testimony of experts concerning the characteristics of truthful and untruthful children generally and to testimony that the children in question displayed the characteristics of truthful children. The defendant also objected to other testimony concerning the bases for the diagnosis of sexual abuse — that is, the “subsidiary principles.”
In agreeing with the defendant that the evidence was inadmissible, the court explained that, under the circumstances of that case — in the absence of physical evidence of abuse — the diagnosis of sexual abuse was based “ultimately and only” on whether the expert believed the child. Lupoli, 348 Or at 361. The court explained that “ordinarily,” an expert witness may explain the basis for the expert’s diagnosis, as long as the diagnosis is itself admissible. The court further observed that the portions of the challenged expert testimony in Lupoli “might be admissible in many circumstances, and perhaps in this case.” Id. at 362. As examples, the court identified the expert’s assessment of (1) whether the particular victim’s statements are developmentally appropriate; (2) the particular victim’s demeanor and changes in demeanor; (3) the particular victim’s disclosure as spontaneous and including descriptive details; and (4) the particular circumstances that point to a child’s suggestibility or the possibility that the child has been coached. Id. Such assessments, the court explained, are “the kind of expert opinion that can assist a jury,” and are not impermissible vouching on their own. Id. However, the court also noted that the challenged statements “must be analyzed with due regard for the context in which they were made, even though defendant does not now complain that the diagnosis itself is [847]*847inadmissible.” Id. at 361. The court went on to conclude that, in the absence of physical evidence of abuse, those otherwise permissible or potentially permissible assessments were not admissible, because they were “inextricably bound up with portions that consisted of clear ‘vouching,’” and could not be “meaningfully separated from the context of which it was apart.” Id. at 362. Lupoli, then, strongly suggests that, when physical evidence of abuse is present, expert opinion is not “ultimately and only” based on credibility, id., and those types of assessments that are offered in support of a diagnosis and that do not constitute clear and direct “vouching” are “ordinarily” admissible — that is, they are admissible if there is physical evidence of abuse.
Citing State v. Preuitt, 255 Or App 215, 223, 296 P3d 648, rev den, 353 Or 868 (2013), defendant asserts that, even if the underlying diagnosis of sexual abuse is admissible, in explaining that diagnosis, an expert may not make direct or indirect comments on the credibility of a witness. In Preuitt, the child victim had been diagnosed with post-traumatic stress disorder (PTSD) as a result of having been sexually abused, but there was no physical evidence of sexual abuse and no diagnosis of sexual abuse. On cross-examination concerning the bases for the diagnosis of PTSD, the expert testified that she had no “concerns or red flags” that the victim was telling a “story that she adopted from somewhere else.” Although the witness was merely responding to and rejecting defense counsel’s suggestions, we concluded that the expert’s statement was a comment on credibility that was not admissible even though it was offered as part of a discussion of a medical diagnosis. Id. at 223.
Preuitt did not involve a diagnosis of sexual abuse and therefore did not implicate the precise issues addressed in Southard or Lupoli. However, we agree with defendant that Preuitt stands for the rule that, when there is an admissible medical diagnosis, although it is permissible for an expert to provide information regarding the bases for the diagnosis, including the general circumstances showing whether a child is or is not suggestible, testimony directly commenting on the truthfulness of the particular child is not admissible.
[848]*848Focusing on the particular diagnosis of sexual abuse involved in this case, we reiterate that in Lupoli, the court held that an expert witness can ordinarily describe the subsidiary principles underlying the diagnosis. Those subsidiary principles include the general characteristics that the expert looks for in examining a child, and necessarily include characteristics that allow the expert to assess the validity of the complaint. When the diagnosis is supported by physical evidence of abuse, short of a direct comment on the child’s credibility, the expert’s testimony explaining the diagnosis can also describe the characteristics of the particular child that led to the diagnosis of sexual abuse, which will assist the jury in evaluating the child’s credibility. See State v. Middleton, 294 Or 427, 435, 657 P2d 1215 (1983) (“It is true that if the jurors believed the experts’ testimony, they would be more likely to believe the victim’s account. Neither of the experts directly expressed an opinion on the truth of the victim’s testimony. Much expert testimony will tend to show that another witness either is or is not telling the truth. See State v. Stringer, 292 Or 388, 639 P2d 1264 (1982). This, by itself, will not render evidence inadmissible.”).
We emphasize that, although an expert’s diagnosis of sexual abuse necessarily includes an assessment of the validity of the child’s complaint, we agree with defendant that the expert witness may not testify as to the expert’s own assessment of the victim’s truthfulness. We recognize that there is a fine line between diagnosing sexual abuse and vouching for a sexual abuse victim’s credibility. But such a line there must be, or the Supreme Court could not have indicated that, in some circumstances where there is physical evidence of abuse, a diagnosis of sexual abuse is admissible. In this case, there was a diagnosis from Kyriakos, not direct testimony of credibility or testimony tantamount to a direct comment. Kyriakos rendered a diagnosis based, as noted above, on physical evidence and subsidiary principles, but she did not directly testify that K was credible. The other witnesses testified as to the characteristics that bear on an evaluation of a child for sexual abuse — consistency of core details, information given in more than one media form, spontaneity, and sensory details. Further, they testified that K was spontaneous and consistent in her disclosures, [849]*849that she used words, her body, and a drawing to describe the abuse, and that she was able to communicate “sensory motor details.” We conclude that those are explanations for the admissible diagnosis of sexual abuse that are not tantamount to comments on truthfulness and that would assist the jury in determining whether K’s complaints were well founded, and that the trial court did not err in admitting the testimony.
As previously noted, defendant was charged with one count of sexual abuse in the first degree, relating to K, and with two counts of attempted sexual abuse in the first degree, relating to J. Defendant moved to sever the charges, arguing that he was substantially prejudiced by a joint trial. The trial court denied the motion, and the charges were tried jointly. During the state’s case in chief, the trial court refused to allow a state witness to be called to testify regarding the state’s theory of the case with respect to J, and the state then moved to dismiss the counts involving J. But the jury had already heard some evidence of the charges pertaining to J. On appeal, defendant contends that the trial court erred in denying his motion to sever the charges for trial, because evidence relating to the charges concerning J was not separately admissible in the trial of the charges relating to K, and ultimately became completely irrelevant when the charges relating to J were dismissed. Defendant contends, further, that the trial court’s failure to sever the charges caused him substantial prejudice.
Under ORS 132.560(3), offenses that have been charged in a single instrument may be severed upon motion, if the defendant is “substantially prejudiced” by a joinder of the offenses.2 Defendant, as the party objecting to joinder, had the burden of showing prejudice. State v. Miller, 327 Or [850]*850622, 626, 969 P2d 1006 (1998). The determination whether joinder results in substantial prejudice involves a case-by-case assessment, State v. Dimmick, 248 Or App 167, 178, 273 P3d 212 (2012), and the mere assertion that evidence relating to some charges will influence the jury’s consideration of other charges is insufficient to establish substantial prejudice, State v. Thompson, 328 Or 248, 257, 971 P2d 879, cert den, 527 US 1042 (1999). We review for legal error the trial court’s determination that defendant failed to demonstrate that joining the charges against him would result in “substantial prejudice[]” under ORS 132.560(3). State v. Luers, 211 Or App 34, 43, 153 P3d 688, adh’d to as modified on recons, 213 Or App 389, 160 P3d 1013 (2007).
Defendant contends on appeal that he was substantially prejudiced by joinder in this case because evidence of the charges pertaining to J was inadmissible propensity evidence under OEC 404(3)* *3 with respect to the charge pertaining to K. State v. Johns, 301 Or 535, 548, 725 P3d 312 (1986) (OEC 404(3) forbids prior crime evidence only when the evidence is offered solely to prove the character of a person and that the person acted in conformity therewith); Luers, 211 Or App at 43 (a defendant may be able to show substantial prejudice by showing that joinder deprived him of protections of the Oregon Evidence Code against the admission of evidence).
The state responds that, when charges have been joined for trial in accordance with ORS 132.560(1), evidence [851]*851of counts relating to separate victims is regarded as non-propensity evidence with respect to the count or counts to which the evidence pertains and does not offend OEC 404(3) as to the other counts. The state is correct. In Miller, 327 Or at 632, the Supreme Court said:
“Because the charges against defendant were joined lawfully for trial under ORS 132.560(1), the evidence of defendant’s criminal behavior during each incident was relevant to prove that defendant had perpetrated the particular offenses to which that evidence pertained. That qualifies as a valid noncharacter purpose that supports admission of the state’s evidence of defendant’s criminal acts. Because the admission of that evidence would not violate OEC 404(3), the trial court correctly determined that defendant failed to show that he would suffer prejudice under ORS 132.560(3) from the admission of that evidence.”
At the time the evidence pertaining to J was admitted in this case, the cases were joined for trial. Under Miller, the evidence was admissible for the nonpropensity purpose of proving the charges relating to J and, contrary to defendant’s contention, did not run afoul of OEC 404(3). The fact that the evidence subsequently became irrelevant after the state’s dismissal of the charges relating to J is not sufficient to demonstrate prejudice in this case.
In his fourth assignment of error, defendant contends that, after the state moved to dismiss the two counts of the indictment pertaining to J, the trial court erred in denying his motion for mistrial. For the same reasons that we reject defendant’s argument that the court should have granted his motion to sever, we conclude that the trial court did not abuse its discretion in denying defendant’s motion for a mistrial on that same issue. State v. Henderson, 242 Or App 357, 364, 255 P3d 661, rev den, 350 Or 571 (2011).
Affirmed.