BEAM, Circuit Judge.
A federal indictment accused Edward F. Balfany of sexually abusing a girl under the age of twelve on four different occasions. A jury ultimately found Balfany guilty of four counts of aggravated sexual assault in violation of 18 U.S.C. §§ 1153, 2241(c) & 2245, and the district court sentenced him to 168 months in prison. Balfa-ny appeals both his conviction and his sentence. Balfany contends, among other things, that the district court admitted prejudicial hearsay, excluded relevant evidence, and improperly enhanced his offense level based on the child’s age and a finding that he threatened the victim. The government cross-appeals, asserting that the district court erred in failing to enhance Balfany’s offense level based on a finding that the child was in his custody when the abuse occurred. We affirm on all issues raised in Balfany’s appeal, but reverse and remand for resentencing on the issue raised in the government’s cross-appeal.1
I. BACKGROUND
The indictment against Balfany accused him of sexually abusing S.N., then an eight-year old girl, on four separate occasions. All four incidents occurred in a house located on the Rosebud Indian Reservation near Mission, South Dakota. Balfa-ny lived in the house with Angela Night-pipe, his common-law wife. Also living in the house were Balfany’s three children from an earlier marriage, Nightpipe’s two children, including S.N., and an infant conceived during the Balfany and Nightpipe relationship. Although Balfany and Night-pipe generally disciplined their own children, they essentially lived together as husband and wife.
S.N. testified at trial, using language appropriate for her age. According to her testimony, Balfany had sexually assaulted her on four different occasions in 1988, before or shortly after the end of the school year. Three of the assaults occurred at night in S.N.’s bedroom. In essence, S.N. testified that on each occasion Balfany entered the room dressed only in a T-shirt and underwear, lifted up her nightgown, and pulled down her panties. Balfa-ny then inserted his penis or fingers into S.N.’s vagina and began to move up and down. The fourth assault occurred during the day on a couch in the living-room. Bal-fany was home alone with S.N. and ordered her to pull down her pants and lay on the couch. Balfany again inserted his penis into S.N.’s vagina and moved up and down. On at least one occasion, Balfany ejaculated, and during one of the assaults, bit S.N. on the shoulder, piercing her skin. S.N. also testified that she experienced great pain during the assaults, but did not shout or scream because she was afraid. Balfa-ny had threatened to whip her with his belt if she told anyone about the assaults.
[578]*578After the school year had ended in 1988, S.N. went to Oregon to live with her aunt, Ilona Bordeaux, and her grandmother. S.N. had spent the previous three summers in Oregon, and this year she stayed in Oregon at the end of the summer and attended school. Bordeaux became concerned about S.N. during the summer. S.N. was often depressed, had trouble eating, and suffered from nightmares. Bordeaux also noticed that S.N. had a bite mark on her shoulder when she first came to Oregon, complained of vaginal pain, and had a vaginal discharge.
Bordeaux took S.N. to see a doctor shortly before the school year began in 1988 because the Oregon public school system required a yearly physical for enrollment. On September 8, 1988, S.N. was examined by Dr. Mark Nurre, a general practitioner who had examined less than ten children for signs of sexual abuse. Bordeaux told Dr. Nurre about her suspicions that S.N. may have been sexually abused. After finishing a routine physical, Dr. Nurre conducted an examination of S.N.’s genital area without the aid of a culdoscope.2 Dr. Nurre was unable to find any evidence of sexual abuse. According to Bordeaux, when Dr. Nurre asked S.N. if she had been abused, S.N. merely hung her head without responding.
While at school in Oregon, S.N. attended presentations addressing good touching and bad touching. Shortly after these presentations, S.N. told Bordeaux "I have to talk to you, I’ve got a secret.” S.N. then revealed to Bordeaux that Balfany had sexually abused her and described the various assaults in detail. Bordeaux, however, did not report the abuse to any authority because she was afraid S.N. would be in danger if Balfany learned that S.N. had revealed the abuse.
In early 1989, S.N. also told a school counselor that she had been sexually abused. The counselor reported the matter to the Children’s Services Division of the Oregon Department of Social Services, and the case was assigned to Glenda Berger, an emergency response worker for the department. In addition to a bachelor’s degree and master’s work in psychology, Berger has received extensive training in conducting interviews with children suspected of being sexually abused.
Berger conducted an interview with S.N. on February 14,1989. A detective with the Oregon State Police Department was also present, but his direct participation in the interview was minimal. Berger used no anatomical dolls or leading questions during the interview. At the beginning of the interview, Berger told S.N. that she and the detective were trying to help her. Berger described S.N. as being very anxious, tense and fearful during the interview. S.N.’s voice was shaky, and she cried throughout the interview. S.N. discussed the various assaults, using age-appropriate language, and identified Balfany as the assailant.
After the interview, Berger arranged for a physician, Dr. Scott Halpert, to conduct a thorough examination of S.N. for medical evidence of sexual abuse. Dr. Halpert is a board certified pediatrician, has attended several training programs on sexual abuse examination and has performed over 150 examinations with a culdoscope. Dr. Hal-pert examined S.N. on February 22, 1989. He began by taking a medical history from S.N., including information about the abuse. According to Dr. Halpert, the patient’s medical history is important because it helps the physician determine what to concentrate on during the examination. Dr. Halpert also stated that the identity of the assailant in particular is very important because it affects his recommendation for counseling. During her conversation with Dr. Halpert, S.N. again described the various incidents of abuse, using age-appropriate language, and identified Balfany as the abuser.
In addition to a small scar on S.N.’s shoulder, Dr. Halpert found positive signs of sexual abuse, including an increased [579]*579opening of the hymen, a significant rounding of part of the hymen, and a scar from a healed tear of the hymen. He testified that these findings were consistent with either penile or digital penetration. Although he could not verify whether S.N.’s hymenal injuries were present at the time Dr. Nurre examined her, Dr. Halpert testified that Dr. Nurre’s findings were not necessarily inconsistent with his. Dr. Hal-pert explained that Dr. Nurre easily could have missed the injuries because he did not have a culdoscope and did not have sufficient experience to distinguish an abnormal hymen from a normal one.
In February 1989, Patricia O’Brien Wal-row (who uses the name O’Brien professionally), a family sexual abuse treatment specialist with the Children’s Services Division, became responsible for S.N.’s treatment. O’Brien has a bachelor’s degree in psychology and a master’s degree in social work. She also has received extensive training in interviewing children suspected of being sexually abused and has dealt with over 1,800 sexually abused children.
O’Brien placed S.N. in a group therapy program with other abused children, which started about April 1989. Prior to the group sessions, however, O’Brien met with S.N. individually five or six times because she believed S.N. needed special attention. At the beginning of these individual sessions, O’Brien explained to S.N. that she needed to know exactly what happened to her. O’Brien asked S.N. about the assaults without using leading questions or anatomical dolls. S.N. described the various assaults to O’Brien, using age-appropriate language, and identified Balfany as the assailant.
Before trial, the government filed a notice with the court stating that it intended to introduce S.N.’s hearsay statements to Bordeaux, Berger, and O’Brien under Fed. R.Evid. 803(24) and 804(5), and S.N.’s statements to Dr. Halpert under Rule 803(4). At trial, Balfany objected to the hearsay testimony of Bordeaux, Berger, and O’Brien on confrontation and hearsay grounds. The court overruled the objections, holding that no confrontation issue existed because S.N. testified at trial and that the hearsay was sufficiently trustworthy for admission under Rule 803(24). Bal-fany did not object to Dr. Halpert’s hearsay testimony at trial.
The hearsay testimony of the four witnesses was mainly the same as S.N.’s testimony. There were a few distinctions, however, some of which were significant. Bordeaux, for example, testified on cross-examination that S.N. said Balfany had threatened to kill her mother and sister if she ever told anyone about the assaults.
II. DISCUSSION
A. Balfany’s Appeal
1. Admission of Hearsay
Balfany contends that the district court committed prejudicial error in permitting Dr. Halpert, O’Brien, Berger, and Bordeaux to testify about the hearsay statements S.N. made to them. We agree that the district court erred in admitting Bordeaux’s hearsay testimony, but we find this to be harmless error.
Initially, we hold that the district court did not err in admitting Dr. Halpert’s hearsay testimony under Rule 803(4). Balfany did not object to this portion of Dr. Halpert’s testimony at trial, so we review the hearsay’s admission for plain error only. E.g., United States v. Miller, 929 F.2d 364, 365 (8th Cir.), cert. denied, — U.S. -, 112 S.Ct. 139, 116 L.Ed.2d 105 (1991). We see no such error. This court previously has held that statements made by an abused child to a physician, including identification of the abuser, are admissible under Rule 803(4) where such statements were reasonably pertinent to the diagnosis and treatment of the child’s physical or psychological injuries. United States v. Shaw, 824 F.2d 601, 608 (8th Cir.1987), cert. denied, 484 U.S. 1068, 108 S.Ct. 1033, 98 L.Ed.2d 997 (1988); United States v. Renville, 779 F.2d 430, 438 (8th Cir.1985); cf. White v. Illinois, — U.S. -, 112 S.Ct. 736, 743-44, 116 L.Ed.2d 848 (1992) (confrontation clause does not preclude admission of child’s out-of-court statements under “firmly rooted” hearsay exception [580]*580for medical treatment). In the present case, Dr. Halpert explained why S.N.’s history, including the identity of her assailant, was pertinent to his diagnosis and recommended treatment. As such, the hearsay statements S.N. made to Dr. Halpert were admissible.
Balfany urges us to conclude that the statements were not admissible under Rule 803(4) because they were not truly pertinent to medical diagnosis or treatment. In particular, Balfany argues that because Dr. Nurre had already performed a medical examination and because S.N. had made several previous statements, both Dr. Halpert’s examination and his interview of S.N. were unnecessary for diagnosis or treatment and were simply conducted to generate incriminating evidence against Balfany. We find this argument to have little merit. Dr. Nurre admitted that he had little experience examining children for signs of sexual abuse and did not have a culdoscope, which is necessary for a thorough examination. Dr. Halpert’s examination, therefore, was not simply designed to generate evidence for trial. Furthermore, we refuse to read a requirement into Rule 803(4) that statements made for the purpose of medical diagnosis or treatment are admissible only if the doctor has no alternative source for the information other than direct questioning of the patient. We consider it eminently prudent and reasonable for a doctor to seek such information directly from the patient as opposed to relying on second-hand sources.
Balfany further argues that the Supreme Court’s decision in Idaho v. Wright, 497 U.S. 805, 110 S.Ct. 3139, 111 L.Ed.2d 638 (1990), compels us to find Dr. Halpert’s testimony inadmissible. In Wright, the Court held that statements made by an abused child to a physician during a suggestive interview are not sufficiently trustworthy for admission under Rule 803(24) over a defendant’s objection that admission violates the confrontation clause. Id. 110 S.Ct. at 3152-53. Wright, however, is inap-posite to Balfany’s case. The Wright Court was concerned with the admission of hearsay under the residual exception— which hearsay is presumptively inadmissible—where the declarant does not testify at trial. See id. Here, in contrast, the hearsay was admissible under Rule 803(4), a firmly established exception, and the de-clarant, S.N., testified at trial and was subject to meaningful cross-examination, avoiding any potential violation of the confrontation clause. See White, 112 S.Ct. at 742-43 & n. 8; United States v. Spotted War Bonnet, 933 F.2d 1471, 1473-74 (8th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 1187, 117 L.Ed.2d 429 (1992).
More importantly, Balfany has not shown that Dr. Halpert’s interview of S.N. was untrustworthy. Although Dr. Halpert knew the substance of S.N.’s claims, such knowledge alone does not render an interview untrustworthy. Balfany has not presented any evidence indicating that Dr. Halpert, like the doctor in Wright, conducted the interview in a suggestive or otherwise improper manner. Instead, Balfany implies that either direct or indirect prompting occurred in the previous interviews of S.N. and thereby rendered Dr. Halpert’s interview with S.N. untrustworthy. Although the Wright Court noted that “ ‘[i]f there is evidence of prior interrogation, prompting, or manipulation by adults, spontaneity may be an inaccurate indicator of trustworthiness,’ ” the Court was addressing prompting by the interviewing doctor, not other adults. See Wright, 110 S.Ct. at 3152 (quoting State v. Robinson, 153 Ariz. 191, 735 P.2d 801, 811 (1987)). Moreover, the fundamental holding of Wright is that only circumstances surrounding the hearsay statements when made are relevant in determining the trustworthiness of the statements. See id. Balfany, therefore, must show that Dr. Halpert conducted the interview in a improper manner. He has not.
We also reject Balfany's argument that the district court erred in permitting O’Brien and Berger to testify about the statements S.N. made to them during their conversations. The court admitted the statements under Rule 803(24) and we review its decision only for clear abuse of discretion. E.g., United States v. DeNoy[581]*581er, 811 F.2d 436, 438 (8th Cir.1987); Renville, 779 F.2d at 439. Not only do we believe that the hearsay was admissible, we believe the district court did not need to resort to the residual exception to the hearsay rule.
Like Dr. Halpert’s testimony, O’Brien’s and Berger’s was admissible under Rule 803(4), an established hearsay exception. This court has established that statements about abuse, including the identity of the abuser, made by a child to a trained social worker or psychologist pursuant to diagnosis or treatment for emotional or psychological injuries are admissible under Rule 803(4). United States v. Provost, 875 F.2d 172, 176-77 (8th Cir.), cert. denied, 493 U.S. 859, 110 S.Ct. 170, 107 L.Ed.2d 127 (1989); DeNoyer, 811 F.2d at 438; cf. White, 112 S.Ct. at 743-44 (confrontation clause does not preclude admission of child’s out-of-court statements under “firmly rooted” hearsay exception for medical treatment). O’Brien is a trained psychologist and Berger is a trained social worker. Both explained that information they acquired from S.N., including the identity of S.N.’s abuser, was pertinent to their diagnosis and treatment of S.N.’s emotional and psychological injuries. S.N.’s statements to O’Brien and Berger, therefore, were admissible under Rule 803(4).
Balfany counters that S.N.’s statements to O’Brien and Berger lack the trustworthiness sufficient for admission and again urges us to follow Idaho v. Wright. We find Wright inapposite here for essentially the same reasons as we did in regard to Dr. Halpert’s testimony. Although O’Brien and Berger knew about S.N.’s claims before they interviewed her, Balfany has not shown any reason to believe that the interviews were conducted in an improper manner.
Balfany additionally argues that O’Brien’s testimony — as well as S.N.’s — is untrustworthy because as part of S.N.’s treatment, O’Brien had S.N. participate with other abused children in a mock grand jury and trial. O’Brien testified that the role-playing games were designed to familiarize children with the court process and to reduce their fear of appearing in court. We see nothing wrong with this practice on its face. Balfany has only shown that the children were educated about the court process in a manner suitable for their age. Prospective witnesses commonly are educated about the trial process before testifying. Balfany must show the games affected the substance of S.N.’s statements to O’Brien or S.N.’s testimony. Balfany must show, for example, that O’Brien coached S.N. to say certain things. He has not done this.
Bordeaux’s testimony, however, presents a more difficult question. The district court permitted Bordeaux to testify about S.N.’s statements to her under Rule 803(24) and we see no other potential basis for admission. To be admissible under the residual hearsay exception, a statement must satisfy the following criteria:
(1) The statement must have circumstantial guarantees of trustworthiness equivalent to the twenty-three specified exceptions listed in rule 803;
(2) The statement must be offered as evidence of a material fact;
(3) The statement must be more probative on the point for which it is offered than any other evidence the proponent can procure through reasonable efforts;
(4) The general purposes of the Federal Rules and the interest of justice must best be served by admission of the statement into evidence;
(5) The proponent of the evidence must give the adverse party the notice specified within the rule.
E.g., Renville, 779 F.2d at 439 (citations omitted). Balfany argues that Bordeaux’s hearsay testimony was neither more probative than other available evidence nor sufficiently trustworthy. We agree with both arguments.
In certain cases where we have approved the admission under Rule 803(24) of an adult’s testimony concerning statements that an abused child made to the adult, we have noted that the child did not testify at trial. See, e.g., United States v. Dorian, 803 F.2d 1439, 1443-46 (8th Cir.1986). Our decisions in United States v. Azure, 845 [582]*582F.2d 1503 (8th Cir,1988), and United States v. Shaw, 824 F.2d 601 (8th Cir.1987), however, indicate that where a child does testify, the child’s testimony may be more probative than the adult’s. In Azure, we held that an adult’s testimony about a child’s hearsay identification of her abuser was not more probative than the child’s in-court identification. Azure, 845 F.2d at 1506-07.3 In Shaw, in contrast, we held that an adult’s testimony concerning a child’s hearsay description of sexual abuse was more probative than the child’s in-court testimony. The adult’s testimony reflected the child’s first comments to anyone about the abuse, the child made the statements just days after the abuse, and the hearsay included information the child was unable to provide at trial. Shaw, 824 F.2d at 610.
The question, therefore, is whether Bordeaux’s hearsay testimony provided information of a kind or to a degree that S.N.’s live testimony or the testimony of the other hearsay witnesses could not. We do not believe this is so. The additional probative value of Bordeaux’s hearsay testimony is very questionable. S.N.’s trial testimony was very comprehensive and included approximate times and details about each incident of abuse. Furthermore, the testimony of the three other hearsay witnesses was also extensive. Although Bordeaux’s hearsay testimony included some facts that the testimony of S.N. and the other witnesses did not, the district court could have (and probably should have) required the prosecution to establish that such information was unavailable from any other witness before permitting Bordeaux to testify. On the other hand, Bordeaux was the first person to whom S.N. revealed the abuse. This factor alone, however, is not sufficient to show that Bordeaux’s testimony was more probative than any other available evidence. See id. (hearsay witness’s status as first person to hear child’s statements only one of several factors noted).
Nonetheless, even if Bordeaux’s testimony had been more probative, we believe the prosecution failed to show that S.N.’s statements to Bordeaux contained sufficient circumstantial guarantees of trustworthiness for admission under Rule 803(24). The Supreme Court has instructed that “ ‘particularized guarantees of trustworthiness’ must be shown from the totality of the circumstances ... that surround the making of the statement and that render the declarant particularly worthy of belief.” Wright, 110 S.Ct. at 3148. Unlike the other hearsay witnesses, Bordeaux did not state that she had any training or background in interviewing children. Although she testified that S.N. appeared very disturbed and used age-appropriate language during their conversations, Bordeaux’s testimony never established that she questioned S.N. in a nonsuggestive manner. If anything, her testimony indicated that she used leading or suggestive questions. On this record, we cannot conclude that the circumstances surrounding S.N.’s statements to Bordeaux contained sufficient guarantees of trustworthiness. See id. at 3152-53 (holding doctor’s use of suggestive questions during interview rendered child’s statements untrustworthy); cf. Dorian, 803 F.2d at 1444 & n. 10; Dana v. Department of Corrections, 958 F.2d 237, 239 (8th Cir.1992). The district court, therefore, erred in admitting Bordeaux’s hearsay testimony.
Balfany, however, is not entitled to reversal of his conviction. We find that the admission of Bordeaux’s hearsay testimony was harmless. Nearly all of Bordeaux’s testimony was either essentially the same as the testimony of S.N. and the other hearsay witnesses, or distinguishable only in minor respects. To this extent, Bordeaux’s hearsay testimony was merely cumulative and, as such, harmless. The one significant difference between Bordeaux’s testimony and that of the other witnesses occurred during Balfany’s cross-examination of Bordeaux. This is when Bordeaux testified that S.N. said Balfany had threatened to kill her mother and sister if she ever told anyone about the assaults.
[583]*583Although this portion of Bordeaux’s hearsay testimony was potentially prejudicial, Balfany waived any objection to it because he elicited the damaging testimony through his own questioning of Bordeaux. We acknowledge that where a defendant objects to certain testimony on direct examination as inadmissible, the defendant may cross-examine the witness concerning the subject of that testimony without waiving an objection to the admissibility of the witness’s testimony on direct or cross-examination. United States v. Straughan, 453 F.2d 422, 428 (8th Cir.1972). The present situation, however, is distinguishable. Balfany did not elicit Bordeaux’s testimony about the death threat while cross-examining her about the hearsay testimony she gave on direct examination over his objection. Instead, Balfany elicited the damaging testimony while asking Bordeaux why she failed to report S.N.’s claims of abuse to anyone until after S.N. had spoken with the school counselor in early 1989:
Q. But you knew about [the abuse] before January 25th, 1989, didn’t you?
A. Yes.
Q. And you didn’t report it to any authorities after [S.N.] told you all of those things, did you?
A. For her and her mother’s protection, no, I didn’t.
Q. That was the only reason; is that right?
A. Because she said that Ed had threatened to kill her mother and her sister if she was to tell anybody. And I was afraid for their lives.
Trial Transcript at 93. Although Bordeaux’s answer may have contained more information than Balfany expected, Balfa-ny did not object to it on any grounds. By eliciting Bordeaux’s damaging testimony through questioning on a subject other than one to which he had previously objected and then failing to object to Bordeaux’s answer, Balfany waived any objection on appeal. Cf. United States v. White, 794 F.2d 367, 369 (8th Cir.1986) (defendant waived objection to admission of written confession where defense counsel, not the government, introduced statement through cross-examination of government witness).4
2. Exclusion of Evidence
Balfany also complains that the district court erred in excluding, on the grounds of relevance, a document that Balfany had sought to use to impeach Bordeaux’s testimony.5 The document was a one-page form that Bordeaux had filled out when she requested medical services for S.N. in the fall of 1988. Balfany argues that the document impeached part of Bordeaux’s testimony, in particular, that S.N.’s grandmother cared for her in the summers before 1988 and that Bordeaux suspected child abuse in 1988. Bordeaux wrote on the form that she had taken care of S.N. during the previous three summers and never expressly noted any concerns about child abuse. Bordeaux, however, did note on the form that S.N. appeared to have a vaginal discharge.
Decisions concerning relevance belong within the sound discretion of the district court. E.g., United States v. Bayless, 940 F.2d 300, 304 (8th Cir.1991). Although we believe Balfany’s argument for admission of the medical form is strong, we conclude that any error in excluding the document was harmless. The district court permitted Balfany to cross-examine Bordeaux about the contents of the document and neither of the points Balfany emphasized were [584]*584complex or difficult to understand. As such, we do not believe Balfany was prejudiced by the jury not having the document during deliberations.
We have considered Balfany’s other claims on appeal concerning the district court’s exclusion of evidence and its restrictions on Balfany’s opening statement and closing argument. We find them to be meritless.
3. Calculation of Offense Level
Balfany also asserts that the district court erred in calculating his offense level. The district court applied U.S.S.G. § 2A3.1, which governs criminal sexual abuse. See U.S.S.G. § 2A3.1 commentary. The district court noted that the base level for Balfany’s offense was 27 and then enhanced the offense level by four points based on a finding that Balfany had used threats to commit the assaults and by four more points based on a finding that S.N. was under the age of twelve at the time of the assaults.
Balfany first contends that the district court committed improper double counting in enhancing his offense level for the victim’s age. See id. § 2A3.1(b)(2). He notes that one of the elements necessary for conviction of aggravated sexual assault under 18 U.S.C. § 2241(c) is that the victim be less than twelve years old. See 18 U.S.C. § 2241(c). Balfany thus argues that his base offense level of 27 includes the victim’s age and the district court’s subsequent enhancement amounted to double counting.
We find this argument specious. Section 2A3.1 applies to both aggravated sexual abuse in violation of 18 U.S.C. § 2241 and simple sexual assault in violation of 18 U.S.C. § 2242. See U.S.S.G. § 2A3.1 commentary. The application notes to section 2A3.1 clearly state that “[t]he base offense level represents sexual abuse as set forth in 18 U.S.C. § 2242.” Id. commentary. The elements for a violation of section 2242 do not include that the victim be less than twelve years old. See 18 U.S.C. § 2242. The Sentencing Commission obviously intended that the age of the victim and other elements of aggravated sexual assault be addressed through enhancements of the base offense level. See U.S.S.G. § 2A3.1(b) & commentary. As such, the district court did not commit improper double counting in enhancing Balfany’s offense level by four points for the victim’s age.
Balfany next contends that the district court erred in enhancing his offense level based on a finding that he committed the assaults by the means set forth in 18 U.S.C. § 2241(a) or (b). See id. § 2A3.1(b)(l). In particular, Balfany challenges the court’s finding that he caused S.N. to engage in a sexual act by threatening or placing S.N. in fear that any person will be subjected to “death, serious bodily injury, or kidnapping.” See 18 U.S.C. § 2241(a)(2). Congress has established that in reviewing a district court’s sentencing decision, we “shall accept the findings of fact of the district court unless they are clearly erroneous and shall give due deference to the district court’s application of the guidelines to the facts.” 18 U.S.C. § 3742(e) (1988); see also, e.g., United States v. Cornelius, 931 F.2d 490, 493 (8th Cir.1991). We do not believe this finding of the district court was clearly erroneous.
Although the district court noted Bordeaux’s inadmissible hearsay during sentencing, the court neither relied solely on this evidence nor needed to do so. Even without Bordeaux’s hearsay statement that S.N. said Balfany had threatened to kill her mother and sister if she told anyone about the abuse, sufficient evidence existed to show that Balfany committed the assaults against S.N. by the means set forth in 18 U.S.C. § 2241(a)(2). S.N. testified that Bal-fany threatened to beat her with his belt if she told anyone. Although this may not appear to be a threat of serious bodily harm on its face, the threat must be examined in context. The threat was not made to an adult; instead, it was made to an eight-year-old girl. In this context, we believe it is reasonable to conclude that Balfa-ny threatened S.N. with serious bodily injury. Moreover, it is reasonable to infer that Balfany made the threat not only to prevent detection of his wrongdoing, but also [585]*585to facilitate future assaults. This case did not involve a threat made after an isolated incident of abuse; on the contrary, the case involved a series of assaults. By threatening S.N. with a beating, Balfany facilitated the commission of multiple assaults.
B. Government’s Cross-Appeal
The government raises only one issue on cross-appeal. It contends that the district court erred in refusing to enhance Balfany’s offense level by an additional two points based on a finding that S.N. was in the “custody, care, or supervisory control” of Balfany. See U.S.S.G. § 2A3.1(b)(3). In reaching its conclusion that the prosecution had failed to prove facts sufficient to justify the enhancement, the district court noted that the question was “very close” and that “it wouldn’t be very hard to go the other way.” Sentencing Transcript at 4. As noted above, we review the district court’s findings of fact under a clearly erroneous standard and accord due deference to its application of the guidelines to the facts. 18 U.S.C. § 3742(e) (1988). We believe the district court erred in its findings of fact and application of the guidelines.
The application notes to section 2A3.1 state that the enhancement for custody of the victim applies because, “[wjhether the custodial relationship is temporary or permanent, the defendant in such a case is a person the victim trusts or to whom the victim is entrusted.” Id. commentary. The purpose behind the enhancement is that such a situation “represents the potential for greater and prolonged psychological damage.” Id.
We believe the evidence shows Balfany had at least temporary custody or care over S.N. Balfany and Nightpipe lived together as husband and wife. Although they often disciplined their own children, testimony at sentencing showed that they shared many household responsibilities, including caring for the children at times. On at least one occasion, the assault on the couch, S.N. was clearly in Balfany’s sole custody and care. Moreover, the purposes underlying the enhancement apply in this case. Balfany was not a complete stranger to S.N. On the contrary, he was a member of the same household. Balfany effectively was S.N.’s stepfather, someone she should have been able to trust. Balfany not only abused S.N.; he abused his relationship to S.N. and her trust. Consequently, the potential for greater and prolonged psychological damage to the abused child exists here. We believe the present circumstances fall directly within the scope of the Sentencing Guidelines. The district court’s failure to enhance Balfany’s offense level by two points is error even with deference to the court in its application of the guideline to the facts adduced. We remand for resentencing.
III. CONCLUSION
We affirm the district court on all the issues Balfany raises in his appeal. We reverse on the issue the government raises in its cross-appeal, however, and remand for resentencing consistent with this opinion.