OBJECTIVE: Siblings and other children who share a home with a physically abused child are thought to be at high risk for abuse, but rates of injury in these contact children are unknown and screening of contacts is highly variable. Our objective was to determine the prevalence of abusive injuries identified by a common screening protocol among contacts of physically abused children.
METHODS: This is an observational, multicenter cross-sectional study of children evaluated for physical abuse, and their contacts, by 20 US child abuse teams who used a common screening protocol for the contacts of physically abused children with serious injuries. Contacts underwent physical examination if they were <5 years old, physical examination and skeletal survey (SS) if they were <24 months old, and physical examination, SS, and neuroimaging if they were <6 months old.
RESULTS: Protocol-indicated SS identified at least 1 abusive fracture in 16 of 134 contacts (11.9%, 95% confidence interval [CI] 7.5–18.5) <24 months of age. None of these fractures had associated findings on physical examination. No injuries were identified by neuroimaging in 19 of 25 eligible contacts (0.0%, 95% CI 0.0–13.7). Twins were at substantially increased risk of fracture relative to nontwin contacts (odds ratio 20.1, 95% CI 5.8–69.9).
CONCLUSIONS: SS should be obtained in the contacts of injured, abused children for contacts who are <24 months old, regardless of physical examination findings. Twins are at higher risk of abusive fractures relative to nontwin contacts.
- CI —
- confidence interval
- CPS —
- child protective services
- SS —
- skeletal survey
What’s Known on This Subject:
Siblings and other contacts of abused children, especially twins, are thought to be at higher risk for abuse than other children. However, the rate at which screening tests identify injuries in contacts is currently unknown.
What This Study Adds:
Contacts of abused children with serious injuries have fractures identified on skeletal survey at significant rates. Twins are at substantially increased risk for fracture. Physical examination findings were not sensitive for fractures.
Physical abuse is an important, and often overlooked, source of morbidity and mortality in childhood.1,2 Early detection can protect children because abuse is often an escalating phenomenon.3–5 Many abusive injuries are clinically occult, and serious injuries can be missed even with careful examination.6,7 Screening high-risk populations can increase detection of abuse,8,9 and siblings and other contacts who share a home with an abused child may represent 1 such population.10,11
Violence often affects an entire household. Child abuse commonly coexists with intimate partner violence,12–15 and abuse of pets has been linked with both child abuse and intimate partner violence.16 Contacts of abused children are victimized at high rates in the years after an initial report of abuse, but the number of contacts who have injuries at the time of the initial referral is unknown.17,18 Case reports suggest that twins of abused children are at especially high risk for abuse, but this has never been tested rigorously.19,20
Child abuse physicians are subspecialty consultants who recommend screening when there is concern for physical abuse, although they rely on cooperation from child protective services (CPS) and others to complete testing.21 CPS agencies are governmental organizations charged with, among other things, investigating reports of abuse and protecting children. In index children, several organizations support guidelines including skeletal survey (SS) for all children with concern for abuse <24 months old.22,23 However, in contact children, in the absence of data about injury prevalence, disagreement about the need for screening is common between medical and CPS professionals,11,24 and rates of screening (physical examination, SS, and neuroimaging) show high variability.10,24
This was a prospective, observational, cross-sectional study involving 20 child abuse teams in the United States who shared a common, minimum standard of care for screening the contacts of abused children. Each team and the coordinating center obtained approval from their respective institutional review board with waiver of informed consent. All screening for abuse was undertaken as part of routine care, and no testing was done for research purposes. Each site established an independent, prospective method for tracking the census of eligible patients, and we tracked enrollment with monthly audits. Missed patients were entered retrospectively. On the basis of monthly censuses, all participating centers completed enrollment for >90% of eligible patients.
We defined index children as children <120 months (10 years) old who were evaluated by a child abuse physician for concerns of physical abuse. Contact children were defined as children <120 months (10 years) old who, in the previous month, were known to share the same household or other care environment where abuse was suspected in an index child. We included contacts from in-home day cares but excluded contacts from commercial day cares because of the potential for large numbers of contacts and the very low rates of abuse at commercial day cares.25 When multiple children from 1 household presented simultaneously with concerns of abuse, all were designated as index children.
We defined index children as “physically abused” if they had both a high likelihood of abuse and at least 1 serious injury (Fig 1). Our screening protocol was used for the contacts of such physically abused index children. The likelihood of abuse was described by the child abuse physician using a previously published 7-point scale. A score of 6 or 7 on this scale was considered “high likelihood” (Appendix).26 This scale uses example cases for each category to improve interrater reliability but ultimately relies on the opinion of the responsible child abuse physician. For example, patterned bruises or burns are listed as an example of category 7, definite inflicted injury, but a child abuse physician could choose to describe a child differently if, for example, the pattern was consistent with an accidental injury or if the child had a bleeding diathesis. Serious injury was defined as fracture(s), burns of >5% total body surface area, traumatic brain injury, intra-abdominal or intrathoracic injury, ICU admission, or death.
Screening Protocol for Contacts
The common screening protocol was used for contacts of physically abused index children as defined earlier. For these contacts, the protocol determined the minimum screening recommendations by age. Child abuse physicians were expected to recommend in-person examination by CPS personnel or physical examination by a medical provider for contacts <5 years old. For contacts <24 months of age, the protocol directed child abuse physicians to recommend SS and physical examination by a medical provider. Contacts <6 months old were expected to undergo neuroimaging in addition to SS and physical examination.
The protocol articulated a common minimum standard for contact screening, and child abuse physicians were free to recommend additional screening. Although the protocol defined the recommendations to be made by child abuse physicians, it could not ensure that screening was ultimately completed, because this may have depended on cooperation from CPS, other medical providers, law enforcement, and others.
All participating centers conducted SS according to guidelines published by the American Academy of Pediatrics and the American College of Radiology.22,27 SS were interpreted by attending radiologists with experience reading SS as part of their clinical duties at each participating center. Physical examinations were conducted according to the normal practices of the examiner. Neuroimaging could consist of CT or MRI, but not ultrasound.
The main outcome measure was the proportion of protocol-indicated SS that identified an abusive fracture in a contact child. Secondary outcomes included the proportions of protocol-indicated physical examinations and neuroimaging that identified injuries in contacts.
Our sample size was determined by using the presumption that clinicians would routinely order SS for a group in which the rate of abusive fractures identified was >5%. To have 80% power at the 2-sided .05 significance level to exclude a rate of 5% if the true rate of abusive fractures is ≥9.3%, we would need to recruit 250 contacts who meet protocol criteria for SS. A pilot study of 8 participating child abuse teams suggested that we would enroll 250 such contacts if we enrolled at least 2500 index subjects, and our funding was budgeted to enroll this number. However, we were only able to enroll 134 contacts meeting protocol criteria for SS when our funding limit was reached, despite enrolling 2890 index children.
Data were entered via a secure, web-based data entry form (Quickbase, Intuit, Waltham, MA). Data included the initial history of trauma or chief complaint in the index child, demographic information, screening tests performed, and all injuries identified for index and contact children. Race and ethnicity of index children were collected to determine if these characteristics were related to the decision to undertake screening. Race was recorded by the child abuse physician according to the information reported at hospital registration by the patient or their caregiver.
We calculated proportions and their associated confidence intervals (CIs) by using intercept-only generalized estimating equations, which account for the correlation of observations from contacts with the same index child. Odds ratios were calculated with logistic generalized estimating equations assuming a compound symmetric working correlation structure to account for correlation of observations from contacts with the same index child. We used SAS 9.2 (The SAS Institute, Cary, NC) for all analyses.
Between January 15, 2010, and April 30, 2011, we evaluated 2901 children for concerns of physical abuse (index children). These index children had 1927 identified contacts. We excluded 11 consultations in which contact children were initially coded as index children. In these cases, data were not available for the true index child because the child died before presentation or was evaluated at a nonparticipating center (Fig 1). Among these index children, 627 (21.7%) met our research definition of “physically abused,” and these children had 479 contacts, of whom 134 were <24 months old.
Index children had as many as 9 contacts identified (Table 1). Level of concern for abuse and injuries identified among index children are shown in Table 2. As with other large groups of children evaluated for physical abuse, index children were predominantly infants, and there was a slight majority of boys.2,9 Index and contact children included 80 pairs of twins and 2 sets of triplets. Most contacts (1141, 59.2%) were biological siblings of the index child, with others being half-siblings (470, 24.3%), other relatives such as cousins (130, 6.7%), adopted or step-siblings (51, 2.6%), and in-home daycare contacts (51, 2.6%).
The protocol indicated SS for 134 contacts. The SS was recommended by the child abuse physician in 122 (91.0%) of these cases and was completed in 101 (75.4%). SS identified ≥1 abusive fracture in 16 of 134 contacts (11.9%, 95% CI 7.5–18.5).
Among contacts who had SS, 9 children with fractures were identified among the 22 who were <6 months old (40.9%), compared with 4 of 16 (25.0%) contacts 6 to 12 months old and 3 of 63 (4.8%) contacts 12 to 24 months old. Contacts 0 to 12 months old were significantly more likely to have fractures than contacts 12 to 24 months old (odds ratio 10.4, 95% CI 2.5–50.8). Eight children had isolated fractures and 8 had multiple fractures, with 51 fractures identified in all (Table 3). Seven contacts were noted to have at least 1 fracture with evidence of healing. None of the fractures had associated signs or symptoms such as bruising, swelling, or tenderness on physical examination, and none were suspected clinically before the SS.
Among the 134 contacts meeting protocol criteria for SS, 16 (11.9%) were twins, and 9 of these (56.3%) had fractures on SS. Twins were significantly more likely than nontwin contacts to have a fracture identified on SS (odds ratio 20.1, 95% CI 5.8–69.9).
There were 25 contact children who met criteria to undergo neuroimaging. Neuroimaging was recommended in 22 (88.0%) and completed in 19 (76.0%). No study demonstrated an intracranial injury (0%, 95% CI 0–13.7). One head CT identified a skull fracture that had previously been identified by SS.
There were 355 contacts who met criteria for physical examination. Physical examination was recommended for 343 (96.6%) of these children and completed in 259 (73.0%). Injuries were identified in 22 contacts (6.2% 95% CI 4.1–9.3) with ages ranging from 2.5 to 59 months (median 26.4 months). Injuries included 19 children with bruises and abrasions, 2 with burns, and 1 with an upper labial frenulum tear.
Child abuse physicians described 6 of the contacts with bruises as having injuries that were patterned or otherwise concerning for abuse. In 6 other contacts, bruises were described as consistent with accidental injury. No information was given about the specificity of the remaining injuries for abuse. Failure to thrive was not categorized as an identified injury but was noted in 4 additional contacts.
Contacts Not Screened
For each of these modalities, approximately one-quarter of patients did not undergo screening indicated by the protocol, either because child abuse physicians did not recommend testing or because recommended tests were not completed. We did not detect a difference in the age, gender, or type of insurance (a surrogate marker for socioeconomic status) of contacts who were not screened relative to those who were. However, contacts were more likely to undergo screening if the associated index child was of nonwhite race or Hispanic ethnicity (Table 4).
To our knowledge, this is the first study to prospectively evaluate screening for physical abuse in contacts of abused children. Nearly half of all index children had at least 1 contact child who shared the same potentially harmful environment. We identified abusive injuries in a significant proportion of contacts who underwent protocol-indicated screening, even though contacts were almost always asymptomatic. These data support obtaining a SS in all children <24 months old who share a household with a physically abused child with a serious injury. The rate of fractures in this group (11.9%) is similar to that of <24-month-old index children being evaluated with concerns for abuse, in whom American Academy of Pediatrics considers the SS to be “mandatory.”22,28,29
Physical examination identified some children with abusive injuries such as patterned bruising and a frenulum tear, but there were no indications of the identified fractures on physical examination. Some injuries identified by physical examination were not specific for abuse and similar injuries are likely to exist in similarly aged children with no risk for abuse.30,31 We recommend physical examination of the young contacts of physically abused children because physical examinations are inexpensive, safe, and may identify specific indications of abuse. However, we wish to underscore that the absence of injury on physical examination does not exclude the potential for other abusive injuries.6,7,9,32
Although neuroimaging did not identify any injuries, our small sample size is not sufficient to support a conclusion about neuroimaging. Because cases of unsuspected abusive head trauma in asymptomatic contacts have been reported,20 it seems reasonable to continue to screen some contacts with neuroimaging until additional data can be obtained.
Because some forms of abuse (shaking, suffocation, punching) can occur without resulting in visible injury, and because injuries could heal before the child is brought for evaluation, the true rate of abuse to contacts is almost certainly higher than the rate of injuries identified in our study.33,34 Because abuse is frequently a progressive process with serious and irreversible outcomes,3,4 identification of subtle abusive injuries is an important opportunity for secondary prevention. Although the injuries we identified in contacts were rarely a significant source of morbidity or mortality, these injuries demonstrate that children who share a home with an abused child are at high risk and should be included when planning out-of-home placement or other protective interventions.
Our data show that twins are at higher risk for abuse relative to other contact children. Twins may be at increased risk because of the increased stress of caring for 2 children simultaneously or because rates of abuse are higher in children with a history of prematurity.35 Twins are also more likely to share other, unmeasured factors, such as their biological relationship to the perpetrator, which may make them more likely to be concurrently abused.
These data do not include results of testing outside our protocol, at the discretion of the child abuse physician, and should not be interpreted to discourage testing of children beyond the protocol.
There exists no gold-standard diagnostic test for abuse, and ratings of abuse likelihood are highly variable between child abuse physicians.27,36 Our definition of whether a child was “physically abused” contained subjective elements26 and excluded children with a high likelihood of abuse but without serious injury (eg, a child with witnessed assault that results only in bruises). If clinicians use a lower threshold to determine that a patient is physically abused, the proportion of injuries identified by the protocol is likely to be lower.
Although each participating center considered the protocol to be a minimum standard of care, as with all protocols, consultants occasionally deviated from the standard. In other cases, testing was recommended by the child abuse physicians but never performed, perhaps as the result of disagreements between CPS and child abuse physicians.11 In total, approximately one-quarter of eligible contacts did not have indicated testing for each modality. Screening was more likely to be completed in contacts who were of nonwhite race or Hispanic ethnicity, a finding consistent with other studies of the effects of race in abuse screening.3,37 Our estimate of the rate of injuries identified by the protocol should therefore be considered a lower limit because contacts who were not tested were counted as if they had no injury.
The SS findings that are most concerning for abuse are subtle and require an experienced radiologist to exclude findings that can mimic abusive fractures.22,38,39 Although all participating centers had enough experience in abuse evaluation to support a dedicated child protection team, and all conducted SS according to published guidelines, we did not review the primary imaging data for children with or without identified fractures. Therefore, some fractures may have been missed that would have been identified by other radiologists or findings that were considered fractures on the clinical interpretation might not have been confirmed by outside review. However, these methods reflect the real-world circumstances likely to face clinicians who cannot routinely refer SS for outside review. We did not measure the time between initial presentation of the index child and completion of the SS in the contact child. However, as fractures show evidence of healing for several months or more,40 we feel it is not likely that a significant number of fractures were missed because of delayed SS.
Similarly, because physical examination or in-person evaluation by CPS was performed according to the normal practices of the examiner, it is possible that abusive injuries may have been missed when examinations did not include specific components such as dedicated retinal or genital examination. It is therefore possible that an even more aggressive protocol of examination would identify a greater number of injuries.
Our conclusion that routine SS is warranted rests on the assumption that the injuries we identified would not have been detected if child abuse physicians had recommended contact screening on a case-by-case basis, according to other measured and unmeasured factors such as parental age or behavior or social risk factors such as known alcoholism or drug use in the home. A study that prospectively asked child abuse physicians to report whether they felt testing was warranted could determine whether such a case-by-case approach would eliminate unnecessary testing, or would miss injuries or increase potential for testing bias.41 We are conducting further analysis of these data to determine whether other measured factors can identify subsets of contacts at higher or lower risk for injury.
Young contacts of physically abused children are at high risk for physical abuse, with risk to twins being relatively increased relative to other contacts. A SS should be performed in contacts <24 months old when the index child is physically abused with serious injury, regardless of physical examination findings.
The Examining Siblings To Recognize Abuse (ExSTRA) investigators were as follows: Jayme Coffman, MD, Cook Children’s Hospital (Ft Worth, TX); Deb Bretl, APNP, Children’s Hospital Wisconsin (Wauwatosa, WI); Nancy Harper, MD, Driscoll Children’s Hospital (Corpus Cristi, TX); Katherine Deye, MD, Children’s National Medical Center (Washington, DC); Tara Harris, MD, Riley Hospital for Children (Indianapolis, IN); Yolanda Duralde, MD, Mary Bridge Children’s Health Center (Tacoma, WA); Marcella Donaruma-Kwoh, MD, Texas Children’s Hospital (Houston, TX); Daryl Steiner, DO, Akron Children’s Hospital (Akron, OH); Ken Feldman, MD, Seattle Children’s Hospital (Seattle, WA); Kimberly Schwartz, MD, University of Massachusetts Medical Center (Worcester, MA); Mary Greiner, MD, Cincinnati Children’s Hospital Medical Center (Cincinnati, OH); Alice Newton, MD, Boston Children’s Hospital (Boston, MA); 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 & Medical Center (Omaha, NE); Alicia Pekarsky, MD, SUNY Upstate Medical University (Syracuse, NY); Andrea Asnes, MD, Yale-New Haven Children’s Hospital (New Haven, CT); Paul McPherson, MD, Akron Children’s Hospital (Youngstown, OH); Neha Mehta, MD, Sunrise Children’s Hospital (Las Vegas, NV); and Gwendolyn Gladstone, MD, Exeter Pediatric Associates (Exeter, NH). Mithat Gönen, PhD provided initial advice about statistical analysis.
We thank Kristine Campbell for her important contributions to early versions of the project protocol.
- Accepted March 27, 2012.
- Address correspondence to Daniel M. Lindberg, MD, Department of Emergency Medicine, Brigham & Women’s Hospital, 75 Francis St—Neville House, Boston, MA 02132. E-mail:
FINANCIAL DISCLOSURE: Drs Lindberg, Shapiro, and Laskey have provided paid expert testimony for prosecution and defense in cases of alleged child physical abuse; and Drs Pallin, Blood, and Berger declare they have no financial relationships relevant to this article to disclose.
FUNDING: This project was supported by a grant from the Health Resources and Services Administration/Maternal and Child Health Bureau, Emergency Medical Services for Children Program (grant H34MC19346-01-02). Dr Lindberg’s effort is funded in part by an Eleanor and Miles Shore Fellowship for Scholars in Medicine from Harvard Medical School. The funders played no role in design and conduct of the study; in collection, management, analysis, and interpretation of the data; and in preparation, review, or approval of the manuscript.
- ↵Sedlak AJ, Mettenburg J, Besena M, et al. Fourth National Incidence Study of Child Abuse and Neglect (NIS-4): Report to Congress. Washington, DC: US Department of Health & Human Services; 2010
- ↵US Department of Health and Human Services. Child Maltreatment 2010. Washington, DC: Administration for Children & Families, Administration on Children, Youth & Families, Children’s Bureau; 2011
- Rubin DM,
- Christian CW,
- Bilaniuk LT,
- Zazyczny KA,
- Durbin DR
- Kellogg ND,
- American Academy of Pediatrics Committee on Child Abuse and Neglect
- Lindberg D,
- Makoroff K,
- Harper N,
- et al.,
- ULTRA Investigators
- Vitale MA,
- Squires J,
- Zuckerbraun NS,
- Berger RP
- Berkowitz CD
- Edelson JL
- Cahill L,
- Sherman P
- Ellaway BA,
- Payne EH,
- Rolfe K,
- et al
- Groothuis JR,
- Altemeier WA,
- Robarge JP,
- et al
- Kleinman PK,
- Di Pietro MA,
- Brody AS,
- et al.,
- Section on Radiology,
- American Academy of Pediatrics
- Chadwick DL,
- Bertocci G,
- Castillo E,
- et al
- Lindberg DM,
- Lindsell CJ,
- Shapiro RA
- ↵ACR Appropriateness Criteria: Suspected Physical Abuse—Child. Reston, VA: American College of Radiology; 2009
- Duffy SO,
- Squires J,
- Fromkin JB,
- Berger RP
- Carpenter RF
- Southall DP,
- Plunkett MC,
- Banks MW,
- Falkov AF,
- Samuels MP
- Hymel KP,
- American Academy of Pediatrics,
- Committee on Child Abuse and Neglect,
- National Association of Medical Examiners
- Kleinman PK
- ↵Gupta M, Schriger DL, Hiatt JR, Cryer HG, Tillou A, Hoffman JR, et al. Selective use of computed tomography compared with routine whole body imaging in patients with blunt trauma. Ann Emerg Med 2011;58(5):407–416.e15.
- Copyright © 2012 by the American Academy of Pediatrics