Secondary Logo

Use of Imaging in Children With Witnessed Physical Abuse

Melville, John D., MD*; Hertz, Stephanie K., DO; Steiner, R. Daryl, DO; Lindberg, Daniel M., MD§ for the ExSTRA Investigators

doi: 10.1097/PEC.0000000000001096
Original Articles

Objective Physicians are occasionally asked to evaluate children who are reported to have been victims of witnessed abuse, but who have no injuries noted on examination. The rate of injury in these patients is presently unknown. This is important because abuse allegations are brought for both altruistic and other reasons. This study compares the use of skeletal survey and neuroimaging in well-appearing and clearly injured children reported to be victims of witnessed child abuse.

Methods Retrospectively planned secondary analysis of the Examination of Siblings to Recognize Abuse cohort of children referred to a child abuse pediatrician with concerns for physical abuse. Children were selected who presented to a medical provider with a history of witnessed child abuse including shaking. Rates of radiographically evident injuries are noted among children with and without injuries noted on physical examination.

Results Among 2890 children evaluated by a child abuse pediatrician, 90 children (3.1%) presented with a history of witnessed abuse. Among these, 51 children (57%) had injuries noted on physical examination; 9 (29%) of 31 skeletal surveys and 9 (35%) of 26 neuroimaging studies revealed injuries. Of 39 children (43%) with witnessed abuse and normal examination, 3 (10%) of 30 skeletal surveys and 2 (8%) of 25 neuroimaging studies revealed an injury.

Conclusions A significant minority of children evaluated for allegations of witnessed abuse will have occult injuries identified radiographically. Absence of injury on examination should not deter physicians from obtaining otherwise indicated skeletal surveys and neuroimaging in children reported to have experienced witnessed abuse.

From the *Division of Child Abuse Pediatrics, Medical University of South Carolina, Charleston, SC;

Akron Children's Hospital,

CARE Center, Akron Children's Hospital, Akron, OH;

§Department of Emergency Medicine, Kempe Center for the Prevention and Treatment of Child Abuse, University of Colorado School of Medicine, Aurora, CO.

Disclosure: The authors J.D.M., R.D.S., and D.M.L. have provided paid expert testimony in cases of alleged child abuse. This project was supported by a grant from the Health Resources and Services Administration/Maternal and Child Health Bureau, Emergency Services for Children Program (H34MC19346-01-02). D.M.L.'s effort is funded in part by an Eleanor and Miles Shore Fellowship for Scholars in Medicine from Harvard Medical School. Funders did not participate in writing the paper or approving its publication.

Reprints: John Donald Melville, MD, Child Advocacy Center, Akron Children's Hospital, 6505 Market St, Bldg C, Suite 3100, Boardman, OH 44512 (e-mail:

Early diagnosis of child physical abuse is frequently difficult, especially in very young children. A physician might receive no history or multiple histories from multiple historians, each with a unique assortment of interests and biases. The patient is often too young to give a meaningful history. Physical examination findings range from clearly diagnostic to completely absent. In a recent series,1 55 (27.5%) of 200 abused infants had previous concerning injuries, which had not been recognized as abuse. Failure to promptly diagnose 1 kind of physical abuse, abusive head trauma (AHT), has resulted in further injury or death.2

Confirming a diagnosis of physical abuse and AHT in infants as well as toddlers rests largely on radiology. In 37% of children suspected to be abused with facial injuries, rib fracture, multiple fractures, or aged younger than 6 months computed tomography (CT) shows a previously unsuspected intracranial injury.3 In another series,4 skeletal survey identified additional injuries in 11% of children being evaluated for suspected physical abuse.

Screening skeletal survey and neuroimaging are recommended for a variety of indications. Among neurologically normal children referred to 2 child abuse teams,5 24 (47%) of 51 skeletal surveys and 11 (29%) of 38 neuroimaging revealed an injury. Skeletal survey revealed occult fractures in 16 (11.9%) of 134 siblings6 of children thought to be abused. Among 150 children aged younger than 18 months with a skull fracture, 9 children (6%) had additional fractures on skeletal survey.7 In 2 studies of children with abusive burns, skeletal survey revealed fractures in 18 (19%) of 978 and 5 (14%) of 369 of the cases.

In appropriate children, skeletal survey and head CT are standard of care for the evaluation of suspected physical abuse.6 Information derived from skeletal survey and follow-up skeletal surveys frequently changes the physician's assessment of child abuse cases.10

Occasionally, physicians are asked to evaluate children who were reportedly victims of witnessed child abuse but in whom no injury was found upon examination. Possible explanations include public misinformation of the amount of force needed to cause AHT,11 witness misperceptions of the alleged abuse, accurate reports of trauma that did not result in observable injury, or false/malicious reporting. Although exact data are not available, deliberate, false reporting of child abuse to physicians is thought to be rare. Among reports made to child protective services, data from the National Child Abuse and Neglect Data System12 indicates that 0.15% of reports are closed as “intentionally false” and 24.2% as “not alleging abuse” in the 4 states that report both dispositions.

In children who have documented injuries, witnessed abuse is strong evidence that physical abuse has occurred. Numerous studies13–15 have used witnessed abuse as part of a composite criterion standard for the diagnosis of physical abuse. However, concerns regarding cost, inconvenience, and the safety of ionizing radiation in children16,17 raise a question as to whether an allegation of witnessed abuse without an observable injury is sufficient to justify radiologic examination. This study specifically addresses the use of skeletal survey and neuroimaging in the evaluation of asymptomatic children in whom witnesses report observing physical abuse.

Back to Top | Article Outline


This is a retrospectively planned secondary analysis of the Examining Siblings to Recognize Abuse (ExSTRA) study. ExSTRA is a prospective, multicenter, observational study of children referred for a child abuse evaluation and their siblings.18 The ExSTRA network enrolled children aged younger than 10 years who were referred to a hospital child protection team with concerns for physical abuse from January 15, 2010 through April 30, 2011. Although the ExSTRA network also studied siblings and contacts of the children referred for abuse, this analysis considered only the children who were initially suspected of having been abused. Each child was evaluated by a child abuse pediatrician (CAP). The ExSTRA network did not direct the CAP with regard to how the children referred for suspected abuse should be evaluated.

Cases for this secondary analysis were selected from children referred to the child protection team for suspected abuse if, at the time of presentation, a witness reported observing an abusive act. Cases were identified in 1 of the 3 ways hereafter: (1) the participating CAP checked a box indicating that the injury event was witnessed by an independent observer; (2) missed cases were identified by a full-text search of the entire database for records containing the terms “shak*” or “witness*.” Full text searches for additional terms including “hit*,” “slap*,” or “saw” did not yield additional cases; (3) subjects who were coded as having been injured in an assault, and inspection of the history of present illness field in the ExSTRA database revealed a witnessed incident.

The ExSTRA data set included injuries where the history was given of an accidental injury, and some of these accidental injuries were witnessed. For each of the cases identified as being a witnessed event, a single reviewer (J.D.M.) assessed the history of present illness field in the ExSTRA data set and identified cases where the witnessed event was not an abusive injury. An independent witness was any person other than the patient and the alleged perpetrator who reported witnessing an abusive event. Children who were reported to have been injured during an altercation between 2 or more adults, but in whom the altercation or injury was not otherwise witnessed, were not included.

The same reviewer (J.D.M.) read the complete ExSTRA data collection record to determine whether the initial evaluation identified clinical signs that, at the time of the initial presentation, could indicate injury. Fields that were reviewed included the CAP's summary of physical examination findings and specific queries regarding bruises, burns, mental status changes, pattern injuries, tenderness, and swelling that were noted on the initial physical examination.

Descriptive statistics were used to summarize the usage and yield of skeletal survey, neuroimaging, and follow-up skeletal survey. The 2 population test for difference in a proportion was used to compare the yield of imaging interventions among children with and without injuries on examination.

In addition to yield of completed examinations, the diagnostic yield was also computed, assuming that all indicated but not completed studies were normal. Based on the data that was previously mentioned and on the guidelines from the American Academy of Pediatrics5 and the American College of Radiology,19 we determined that an initial and follow-up skeletal survey was indicated for a child aged younger than 24 months, and that neuroimaging was indicated for children aged younger than 6 months.

Back to Top | Article Outline


Figure 1 details the selection of cases for this retrospective review. Of 2890 index patients, 195 were noted to have a witnessed injury event. The reviewer determined that the witnessed event was abusive in 90 cases. Physical examination corroborated a history of trauma in 51 cases, leaving 39 children with a witnessed, abusive history and a normal examination.



Table 1 describes the perpetrators and witnessed events among 90 children with witnessed abuse. Parents perpetrated most of the abuse. Hitting, kicking, punching, or slapping was the most common abusive mechanism. Table 2 indicates the witnesses who observed either abuse or nonabusive injury.





Most of the children assigned to the “normal examination” group had no documented physical examination findings other than a notation of a normal examination. A total of 3 cases with findings unrelated to abuse, diaper rash, stigmata of congenital syndrome, and a “sleepy child,” were included in the normal examination group. In the case of the sleepy child, the CAP who evaluated the 12-month-old infant documented both “no injury” on physical examination and a “normal” mental status.

Table 3 displays demographic characteristics of children who are reported to have experienced witnessed abuse, compared with the entire cohort of children referred to a CAP. Among children with witnessed abuse, those with normal examinations are significantly younger than children with clinical signs of injury (P = 0.010.)



As shown in Table 3, the yield of skeletal survey, follow-up skeletal survey, and neuroimaging among children with reported witnessed abuse, a normal examination, and completed imaging studies were 10%, 10%, and 8%, respectively. One skeletal survey and 2 neuroimaging studies were not completed that would have been indicated based on the patients' age. Compliance with follow-up skeletal survey, which requires the patient to return approximately 2 weeks after the evaluation, was only 33%. Assuming that all the studies indicated by the patient's age but not performed had been normal, the yield of skeletal survey, follow-up skeletal survey, and neuroimaging would be 8%, 3%, and 5%, respectively.

Back to Top | Article Outline


This study is aimed at a specific clinical situation, an allegation of witnessed physical abuse in a well-appearing child with a normal physical examination. This presentation, even though unusual, is not rare. On average, participating ExSTRA sites encountered this situation more than once a year.

Our data suggest that children alleged to be victims of witnessed abuse should have a radiologic workup for occult injury. Currently, we recommend skeletal survey for children aged younger than 2 years.5 Noncontrast head CT is indicated for children with multiple fractures, rib fractures, injuries to the face, or who are aged younger than 6 months.3 In this cohort, failure to complete the radiologic workup would have missed findings in 4 (10%) of 39 patients with normal examinations. One child had findings on both skeletal survey and neuroimaging.

The yield of skeletal survey in children with witnessed abuse and a normal examination (10%) is similar to the yield of other recommended imaging tests. A previous analysis of the ExSTRA data set found that 11.9% of siblings and contacts of abused children have positive skeletal surveys.14 In a recent sample of 703 consecutive skeletal surveys obtained because of concern for abuse, 10.8% showed positive results.4

This recommendation that reports of witnessed abuse prompt a radiologic workup is consistent with the actual practice of the ExSTRA investigators. Of 57 patients aged younger than 24 months, 53 patients (93%) were evaluated with a skeletal survey. Of 36 patients aged younger than 6 months, 31 patients (86%) underwent neuroimaging.

As mentioned previously, witnessed assault is frequently used as a criterion standard for the diagnosis of abuse in children with traumatic injuries. The data presented here do not suggest that reports of witnessed abuse are frequently unreliable. In our sample, 52 (57%) of 91 children with witnessed abuse had physical examination findings consistent with that history. Skeletal survey and neuroimaging identified injuries in 30% and 36% of these children, respectively. An additional of 4 children without injuries noted on examination had radiographic evidence of trauma.

The yield of skeletal survey and neuroimaging is lower in children without injuries noted on examination than those with injuries. This finding could be explained by reasoning that abuse that does not cause visible injury might be less severe than abuse that does. Thus, abuse that did not leave cutaneous injury might be less likely to cause skeletal injury as well. However, we cannot exclude the possibility that in some cases, no skeletal injury was found because the histories were fabricated or incorrect.

Strengths of this study include that it was a multicenter, study in which data was obtained prospectively, and in which regular audits ensured that all eligible children were enrolled. Weaknesses include that few children had experienced witnessed abuse with no physical examination findings.

This was a retrospective, observational study. Some information, such as the specific diagnoses on the positive imaging studies, was not uniformly recorded. Most of the patients entered the study because the CAP checked a box indicating that the event was “witnessed.” The accompanying history frequently did not delineate who witnessed or who committed the abuse. This prevents us from analyzing if some types of witnesses were more reliable than others.

This study addresses only children who were referred to the hospital's child protection team for evaluation. Anecdotally, many of the participating centers report a policy of referring all reported child abuse to the child protection team. Emergency physicians and other front-line providers, however, may have elected not to refer a child to the child protection team despite a report of witnessed child abuse if they performed all the tests, and these were negative. If this happened, it would decrease the yield of testing. These results are therefore most applicable to cases in which consultation occurs, and not to the overall population of children seen in an emergency department.

Back to Top | Article Outline


Physicians are occasionally asked to evaluate children who were reportedly the victims of witnessed abuse but who seem healthy and have a normal examination. This situation leads to concerns that false reporting may lead to excessive use of ionizing radiation for child abuse workups. The results of this multicenter observational study suggest that a minority of these children will have significant finding on skeletal survey or neuroimaging. Given the potentially dire consequences of missing child abuse, especially in this younger population, skeletal survey and head CT are indicated for children aged younger than 2 years and 6 months, respectively.

Back to Top | Article Outline


The ExSTRA investigators are: Jayme Coffman, MD (Cook Children's Hospital, Fort Worth, Tex), Deb Bretl, APNP (Children's Hospital Wisconsin, Wauwatosa, Wis), Nancy Harper, MD (Driscoll Children's Hospital, Corpus Christi, Tex), Katherine Deye, MD (Children's National Medical Center, Washington, DC), Antoinette L. Laskey, MD, and Tara Harris, MD (Riley Hospital for Children, Indianapolis, Ind), Yolanda Duralde, MD (Mary Bridge Children's Health Center, Tacoma, Wash), Marcella Donaruma-Kwoh, MD (Texas Children's Hospital, Houston, Tex), Daryl Steiner, DO (Akron Children's Hospital, Akron, Ohio), Ken Feldman, MD (Seattle Children's Hospital, Seattle, Wash), Kimberly Schwartz, MD (University of Massachusetts Medical Center, Worcester, Mass), Robert A. Shapiro, MD, and Mary Greiner, MD (Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio), Alice Newton, MD (Boston Children's Hospital, Boston, Mass), Rachel Berger, MD, MPH, and Ivone Kim, MD (Children's Hospital Pittsburgh of University of Pittsburgh Medical Center, Pittsburgh, Pa), Kent Hymel, MD (Dartmouth-Hitchcock Medical Center, Lebanon, NH), Suzanne Haney, MD (Children's Hospital and Medical Center, Omaha, Neb), Alicia Pekarsky, MD (SUNY Upstate Medical University, Syracuse, NY), Andrea Asnes, MD (Yale-New Haven Children's Hospital, New Haven, Conn), Paul McPherson, MD (Akron Children's Hospital, Youngstown, Ohio), Neha Mehta, MD (Sunrise Children's Hospital, Las Vegas, Nev), and Gwendolyn Gladstone, MD (Exeter Pediatric Associates, Exeter, NH).

Back to Top | Article Outline


1. Sheets LK, Leach ME, Koszewski IJ, et al. Sentinel injuries in infants evaluated for child physical abuse. Pediatrics. 2013;131:701–707.
2. Jenny C, Hymel KP, Ritzen A, et al. Analysis of missed cases of abusive head trauma. JAMA. 1999;281:621–626.
3. Rubin DM, Christian CW, Bilaniuk LT, et al. Occult head injury in high-risk abused children. Pediatrics. 2003;111:1382–1386.
4. Duffy SO, Squires J, Fromkin JB, et al. Use of skeletal surveys to evaluate for physical abuse: analysis of 703 consecutive skeletal surveys. Pediatrics. 2011;127:e47–e52.
5. Laskey A, Holsti M, Runyan D, et al. Occult head trauma in young suspected victims of physical abuse. J Pediatr. 2004;144:719–722.
6. Christian CW and the Committee on Child Abuse and Neglect. The evaluation of suspected child physical abuse. Pediatrics. 2015;135:e1337–e1354.
7. Laskey AL, Stump TE, Hicks RA, et al. Yield of skeletal surveys in children ≤ 18 months of age presenting with isolated skull fractures. J Pediatr. 2013;162:86–89.
8. Degraw M, Hicks RA, Lindberg D. Incidence of fractures among children with burns with concern regarding abuse. Pediatrics. 2010;125:e295–e299.
9. Hicks RA, Stolfi A. Skeletal surveys in children with burns caused by child abuse. Pediatr Emerg Care. 2007;23:308–313. J Pediatr. 2004; 144(6):719–722.
10. Harper NS, Eddleman S, Lindberg DM, et al. The utility of follow-up skeletal surveys in child abuse. Pediatrics. 2013;131:e672–e678.
11. National Center on Shaken Baby Syndrome. Can tossing or rough play cause SBS/AHT. Available at: Accessed September 4, 2014.
12. U.S. Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau. Child maltreatment 2012. Available at: Accessed October 7, 2014.
13. Hymel KP, Wilson DF, Boos SC, et al. Derivation of a clinical prediction rule for pediatric abusive head trauma. Pediatr Crit Care Med. 2013;14:210–220.
14. Lindberg DM, Lindsell CJ, Shapiro RA. Variability in expert assessments of child physical abuse likelihood. Pediatrics. 2008;121:e945–e953.
15. Kemp AM, Dunstan F, Harrison S, et al. Patterns of skeletal fractures in child abuse: systematic review. BMJ. 2008;337:a1518.
16. Mathews JD, Forsythe AV, Brady Z, et al. Cancer risk in 680 000 people exposed to computed tomography scans in childhood or adolescence: data linkage study of 11 million Australians. BMJ. 2013;346:f2360.
17. Pearce MS, Salotti JA, Little MP, et al. Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study. Lancet. 2012;380:499–505. Brenner DJ, Hall EJ. Computed tomography–an increasing source of radiation exposure. The New England journal of medicine. Nov 29 2007;357(22):2277–2284.
18. Lindberg DM, Shapiro RA, Laskey AL, et al. Prevalence of abusive injuries in siblings and household contacts of physically abused children. Pediatrics. 2012;130:193–201.
19. Meyer JS, Gunderman R, Coley BD, et al. ACR Appropriateness Criteria(®) on suspected physical abuse-child. J Am Coll Radiol. 2011;8:87–94.

child abuse; witnessed; shaking; skeletal survey; imaging

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.