OBJECTIVE:The reported prevalence of intracranial injuries (ICI) in asymptomatic infants undergoing physical abuse evaluations varies widely. To inform neuroimaging decisions, we sought to 1) describe ICI prevalence among asymptomatic infants evaluated for abuse, 2) assess the impact of different definitions of "asymptomatic" on ICI prevalence across different age groups, and 3) assess whether hospital-level neuroimaging rates impact prevalence of identified ICI. METHODS:We conducted a cross-sectional study of infants <12 months without seizures or altered mental status (AMS) undergoing subspecialty evaluations for physical abuse from February 2021 through December 2022 within CAPNET, a multicenter child abuse research network. We report ICI prevalence among infants without seizures or AMS and across 3 subpopulations with progressively stricter definitions of "asymptomatic." RESULTS:Of 1947 infants without seizures or AMS, 1513 (77.7%) underwent neuroimaging. Of these infants, 13.7% had ICI. After excluding infants with signs of blunt head injury, ICI prevalence decreased to 5.4%. After additionally excluding infants with neurologic signs or symptoms, ICI prevalence was 5.0%. With additional exclusion of infants with nonspecific symptoms, ICI prevalence was 4.7%. CONCLUSIONS:ICI prevalence in asymptomatic infants depends upon the definition of "asymptomatic". When strictly defined, asymptomatic ICI occurs in approximately 1 in 21 infants undergoing evaluations for abuse. Among infants without signs of blunt head trauma, infants <6 months have a higher risk of ICI compared to older infants.

BACKGROUND:Many children who experience abusive head trauma (AHT) have long-term, multidisciplinary healthcare needs. However, little is known about longitudinal follow-up or characteristics associated with medical retention. Prior longitudinal studies of outcomes have been hampered by low retention, raising questions about barriers to follow-up after the diagnosis of AHT. In order to facilitate follow-up care, we created a comprehensive, multidisciplinary clinic for children with AHT. OBJECTIVE:Describe patient retention in a multidisciplinary clinic for children with AHT and examine associations between injury characteristics and child placement outcomes with retention metrics. PARTICIPANTS AND SETTING:Data were abstracted from the medical charts of all children diagnosed with AHT from 6/1/2011 to 12/31/2020 at a single center and referred to a dedicated follow-up AHT clinic (Non-Accidental Brain Injury Care Clinic; NABICC). METHOD:We calculated the number of follow-up visits over 24 months. We used multivariable regression to determine the association between demographics, injury characteristics, and child placement outcomes with clinic retention. RESULTS:Over the study period, 407 patients were diagnosed with AHT and 365 (90%) children survived to hospital discharge. Eighty percent attended at least one NABICC follow-up visit, with an average of 4.5 visits over 24 months. In univariate analysis, parenchymal injuries, retinal hemorrhage, increasing PICU days, and removal from the home were statistically correlated with increased number of follow-up visits. In multivariable regression analyses, subdural hemorrhage, parenchymal injuries, PICU days, and out-of-home placement were all significantly associated with increased visits. Younger children had more visits compared to older children (≥ 24 months). CONCLUSION:Retention was high in this longitudinal, multi-disciplinary AHT clinic. Retention is associated with severe injury and out-of-home placement. This model of care has potential to advance clinical care and outcomes research.

OBJECTIVES:Retinal hemorrhages (RHs) are reported as rare in children with concerns for physical abuse without intracranial injury (ICI). The current performance of retinal examinations in this population is unknown. We sought to assess factors associated with performing a retinal examination in young children without ICI and describe the yield of these examinations. METHODS:We performed a retrospective, multicenter cross-sectional study of children younger than 2 years undergoing subspecialty physical abuse evaluations between February 2021 and August 2023. Children were excluded if they had ICI, lacked neuroimaging, had RHs prompting evaluation for abuse, or died. Outcomes were (1) performance of a retinal examination and (2) yield of these examinations. We used descriptive statistics to report frequencies and proportions, and χ 2 or Fisher exact tests to assess unadjusted associations between clinical and consult characteristics with performance of retinal examinations. RESULTS:Among 2591 children without ICI, retinal examinations were obtained in 10.5% (272). Performance of a retinal examination was significantly associated with younger age, in-person CAP evaluation, transfer status, CAPNET site, presence of high-risk bruising, presence of high-risk fractures, presence of subconjunctival hemorrhage, evidence of head trauma, and ill appearance. Four (1.5%) of the 272 children receiving retinal examinations had RHs on examination, all of which had potential alternate explanations for the RHs. CONCLUSION:Retinal examinations rarely reveal RHs in young children without ICI who are undergoing abuse evaluations, suggesting that these examinations may be safely deferred in this population. Retinal examinations may be warranted in cases of suspected strangulation or other signs of ocular trauma.

BACKGROUND:While subconjunctival hemorrhages (SCHs) have been suggested by some to be sentinel injuries which should prompt further evaluation for abuse, the risk of abuse in children with isolated SCHs is not currently known. OBJECTIVE:To determine if infants and young children with isolated SCHs have current or subsequent concerns for physical abuse. PARTICIPANTS AND SETTING:Children 1 to <48 months old with SCHs diagnosed between 2011 and 2015 within a large, integrated healthcare system were included and followed until their fifth birthdays. METHODS:Retrospective cohort study. Children with isolated SCHs had no other injuries identified within 7 days before or after the SCHs. Child physical abuse before the child's fifth birthday was identified by diagnosis codes and free-text review of medical notes and radiology reports. RESULTS:We identified 1281 patients with SCHs (mean [SD] age 22.5 [14] months), 603 (47.1%) were female. Only 12 children (0.9%) had physical abuse. Among children with isolated SCHs, 0.4% had physical abuse, compared to 3.0% of those with non-isolated SCHs (p < 0.001). For infants <12 months old, 0.6% with isolated SCHs had physical abuse, compared to 10.0% of those with non-isolated SCHs. CONCLUSIONS:Most children with SCHs in our study did not have child physical abuse before their fifth birthday. Children with isolated SCHs had a lower risk of future abuse compared to children with non-isolated SCHs.

OBJECTIVE:Prior research on confessions of physical abuse (PA) has explored mechanisms of injury. Little is known about conditions supporting a confession of abuse or how confessions may influence case outcomes. METHODS:This cross-sectional study of suspected PA in children aged <10 years at 10 centers participating in CAPNET, a multicenter child PA research network, included children with in-person Child Abuse Pediatrics (CAP) consultations between February 2021 and December 2022 and excluded children without injury. Our focus was a confession of inflicted injury known to the CAP during clinical involvement. We compared child and clinical characteristics, out-of-home (OOH) placements, and arrests between cases with and without confessions. Multivariable models using generalized estimating equations (GEE) produced adjusted predicted probabilities (APP) clustering by CAPNET site. RESULTS:Confessions were known to CAPs in 115/4297 (2.7%) cases, with significant site variability. In a multivariable GEE model, confessions were more likely in cases with injuries with high specificity for abuse (APP 6.0% vs 1.6%, P < .001), near-fatality (APP 5.6% vs 2.3%, P < .001), and older children (APP 4.8% vs 2.5%, P = .025). OOH placements and arrests were more common in cases with confessions. CAP awareness of confessions, OOH placements, and arrests did not follow racial or ethnic patterns seen in the US child welfare system. Lower socioeconomic status was associated with OOH placements and arrests but not with CAP awareness of confessions. CONCLUSIONS:Our findings highlight differences in case characteristics and outcomes where a confession is known to the CAP and raise questions about agency response based on child characteristics.

OBJECTIVE:To determine whether community advantage moderated rates of hospital encounters for child physical abuse during the COVID-19 pandemic. METHODS:This was a population-based retrospective study of 12 states during 2019 to 2021. The Child Opportunity Index 3.0 (COI) measured ZIP code-level community advantage. Average monthly rates of hospital encounters (ie, ED visits and hospitalizations) for physical abuse per 100,000 children aged <5 years were calculated for each pandemic period overall and stratified by COI quintile and type of injury. Pandemic periods were defined a priori as pre-pandemic (Jan 2019 to Feb 2020), early-pandemic (Apr 2020 to Aug 2020), and mid-pandemic (Sept 2020 to Dec 2021). Poisson regression models compared rates between pandemic periods. RESULTS:Compared to pre-pandemic, the overall rate of hospital encounters for physical abuse was 14.0% lower during the early-pandemic period (P<.0001) and statistically unchanged during the mid-pandemic period. Compared to pre-pandemic, rates of physical abuse during the early-pandemic period were 20.3% lower in ZIP codes with very low COI (P = .001) and 17.8% lower in ZIP codes with low COI (P = .01), without significant changes among other COI quintiles. Compared to pre-pandemic, rates of isolated skin injuries were 19.3% lower in the early-pandemic period (P<.0001) and 7.5% lower in the mid-pandemic period (P = .02), without significant changes in abusive traumatic brain injury or abusive fractures. CONCLUSIONS:Lower-severity abusive injuries were less often diagnosed early in the pandemic, specifically among children from disadvantaged communities. These results hint at a complex interplay of potential prevention via social policy and factors relating to detection.

BACKGROUND:Siblings and household contacts of physically abused children are at increased risk for physical abuse and may have injuries that are not clinically apparent. However, injury testing recommendations for contact children remain debated. OBJECTIVE:Our objective was to describe testing practices and yield for contacts of physically abused children. PARTICIPANTS AND SETTINGS:This is a cross-sectional study of children <10 years old evaluated for physical abuse at 10 US pediatric referral centers in the CAPNET research network. METHODS:We report the proportion of contact children who had a skeletal survey (SS) or neuroimaging and describe the proportion of tests that identified injuries. RESULTS:Among 6256 index evaluations, 4863 contacts were identified. Among 794 contacts with physical examinations, 530 had no findings on physical examination and 694 had no history of trauma. Among 316 contacts <24 months old where the index child had high concern for abuse, 196 (62%) had a SS and 20 had fractures identified by the SS. SS completion was more common for twins, contacts <12 months old, and contacts with findings on physical exam. Among 123 contacts <12 months old where there was high concern for abuse in the index child, 49 (40%) had neuroimaging. Of these, 5 had intracranial hemorrhage; all were twins. CONCLUSIONS:When skeletal survey and neuroimaging were completed in contacts of abused children, injuries were more common in twins and younger contacts.

BACKGROUND AND PURPOSE:Traumatic brain injury (TBI) is a frequent indication for neuroimaging in young children. Head CT is widely available and fast, but it exposes children to ionizing radiation. Routine MRI avoids radiation but is longer and frequently requires sedation. Fast MRI without sedation is increasingly used instead of head CT in young children with head trauma, but the intracranial findings it may miss relative to routine MRI are not well defined. MATERIALS AND METHODS:We retrospectively studied children 6 years of age or younger who underwent fast brain MRI for evaluation of TBI and subsequent routine brain MRI within 7 days from 2014 through 2024. Reports were abstracted for traumatic intracranial findings. To address selection bias of the paired cohort, we also identified children who underwent fast MRI as their initial neuroimaging study for evaluation of TBI who did not undergo a subsequent routine MRI within 7 days. We reviewed any return emergency department or inpatient visits within 1 year of the initial fast MRI for evidence of a missed traumatic brain injury. RESULTS:The selected paired cohort included 69 children; 68 had traumatic intracranial findings on routine MRI. Fast MRI detected at least one traumatic intracranial abnormality in all 68 routine-MRI-positive cases. Compared with routine MRI, fast MRI was less sensitive for SAH (18/24, 75.0%) and parenchymal injury (20/31, 64.5%). Among 2471 children with an initially negative fast MRI, 1,122 re-presented to the emergency department or inpatient setting within one year. Four had traumatic findings on follow-up neuroimaging, all attributed to new trauma rather than delayed recognition. CONCLUSIONS:In this selected cohort, fast MRI detected traumatic brain injury consistently, while selected components were apparent only on routine MRI. Fast MRI may serve as an initial screening and triage examination in hemodynamically stable children when skull fracture detection or complete forensic characterization is not needed.

BACKGROUND AND OBJECTIVES:Intimate partner violence (IPV) and child physical abuse often co-occur, but the frequency and risk factors for abusive injuries in IPV-exposed children remain unknown. Our objectives were to (1) determine frequency of skeletal surveys (SSs) and neuroimaging use because of IPV and their diagnostic value in children aged less than 24 months evaluated by a child abuse pediatrics specialist (CAP) and (2) identify factors associated with performance of and findings on SS and neuroimaging. METHODS:We conducted a retrospective study of children aged less than 24 months undergoing CAP consultations because of IPV exposure from February 2021 to February 2024 in a multicenter child abuse research network. We analyzed use of SS and neuroimaging and associations between imaging use and sociodemographic factors and associations between imaging positivity and clinical factors (age, injury-related chief complaint, presenting trauma history, injury on physical examination). RESULTS:Among 326 children, 305 (93.6%) underwent SSs, with 29 (9.5%) positive for a new fracture. Among 267 children without signs/symptoms of head trauma, 110 (41.2%) had neuroimaging, with 7 (6.4%) positive for intracranial injury. There was no disproportionality in SS use, but race/ethnicity and insurance type were associated with the use of neuroimaging. The presence of an injury on physical examination finding was associated with newly identified fractures (15.7% vs 6.4%; P = .009). CONCLUSIONS:Current evidence supports use of SSs in children aged less than 24 months undergoing medical evaluations after exposure to IPV. Further research is needed to assess risk factors for positive findings on SSs and neuroimaging with larger sample sizes.

BACKGROUND AND OBJECTIVES:Physical abuse is a common cause of morbidity in infants but is often missed in those with medically minor injuries, increasing the risk for escalating violence. Our objective was to compare the prevalence of abuse-specific injuries among infants with and without prior injuries. METHODS:This is a retrospective, multicenter study of infants (aged <12 months) undergoing child abuse pediatrics provider (CAP) consultation for suspected physical abuse between 2/2021 and 12/2022 at 10 participating US centers. Encounters were dichotomized into those with and without history of prior injury. The groups were compared for rates of diagnostic screening and prevalence of abuse-specific injuries. The association of prior injuries with identification of abuse-specific injury was assessed using logistic regression controlling for demographics and presence of prior injury. RESULTS:We identified 2352 eligible encounters (median age: 4.2 months), of which 13% had a history of prior injury. In infants who underwent diagnostic testing, injury identification was significantly higher in those with prior injury (45% vs 22%, P < .05). Adjusting for patient characteristics, the presence of prior injury in infants was associated with higher odds of at least 1 type of abuse-specific injury being identified, compared with those without prior injury (adjusted odds ratio = 2.81, 95% CI = 2.10-3.76). CONCLUSION:Infants with prior injury undergoing CAP evaluation for suspected physical abuse have higher odds of abuse-specific injuries compared with those without prior injury. This study emphasizes the importance of standardized history and evaluation in infants presenting with concerning injuries.

OBJECTIVES:To assess the frequency and yield of retinal examination in children below 2 years old undergoing abuse evaluations in the setting of skull fracture(s) and small underlying intracranial hemorrhage. METHODS:This cross-sectional study used CAPNET, a multicenter child physical abuse network, to identify children below 2 years with a skull fracture(s) and intracranial injury limited to an underlying small focal intracranial hemorrhage undergoing subspeciality child abuse evaluations. Our outcomes of interest were (1) the performance of a retinal examination, (2) the identification of retinal hemorrhages, and (3) associations of clinical factors and CAPNET site with the performance of retinal examinations. We hypothesized that retinal hemorrhages would be identified in <5% of patients. RESULTS:Of 242 children who met inclusion criteria, the majority (189, 78.1%) presented with a reported history of accidental trauma, and most (211, 87.2%) lacked additional injuries. Only 9 (3.7%) had loss of consciousness and/or seizures/seizure-like activity. The majority (201, 83.1%) had low concern for abuse. Overall, 104 (43.0%) children underwent retinal examinations, of which 0 had retinal hemorrhages (one-sided 95% CI: 0-2.8%). Children without a reported accidental mechanism of injury ( P =0.004), those with intermediate/high concern for abuse ( P <0.001), and children with occipital fractures ( P =0.008) were more likely than their counterparts to undergo retinal examination. The proportion of children undergoing retinal examination varied by CAPNET site ( P <0.001). CONCLUSIONS:Our findings suggest that it may be reasonable to forgo retinal examinations in children below 2 years of age with skull fracture(s) and intracranial injury limited to an underlying small focal hemorrhage who are overall neurologically well-appearing.

BACKGROUND:Cytotoxic Edema (CE) is a form of brain parenchymal injury hypothesized to be associated with abusive mechanism and poor outcome for young children with traumatic brain injury. CE is reliably identified by magnetic resonance imaging but not by computed tomography. OBJECTIVE:We sought to test the association of CE with injury severity, mechanism (abusive or non-abusive) and functional outcome in a large cohort of young children who all had magnetic resonance imaging. PARTICIPANTS AND SETTING:Retrospective cohort study of children <6 years old admitted for traumatic brain injury between January 2011 - June 2020 and who had MR. METHODS:Two pediatric neuroradiologists determined CE presence; mechanism of injury was determined by reviewing child protection team notes. Injury severity was measured using the injury severity score (ISS); functional outcome at hospital discharge and 6-12 months later was measured using the functional status score (FSS). RESULTS:We identified 431 eligible children with traumatic brain injury. CE was associated with injury severity (Median [IQR] ISS for those with CE was 18 [17-26] vs. 17 [11-18] for those with no CE) and functional status at discharge (Median [IQR] FSS 9 [7-11] vs 6 [6-8] without CE) and 6-12 months from injury (Median [IQR] FSS 8 [6-10] vs 6 [6-8] without CE). CE was not associated with injury mechanism (OR 0.78; 95%CI 0.52-1.17). CONCLUSION:While CE was associated with injury severity and outcome, it was not associated with injury mechanism. Prior associations may reflect increased injury severity for abused children.

OBJECTIVES:Using a multicenter child abuse pediatrics research network (CAPNET), the study objectives were to characterize the adherence to American Academy of Pediatrics (AAP) recommendations for bleeding disorder testing and the frequency of bleeding disorder identification. METHODS:We performed a descriptive study of bleeding disorder evaluations among children younger than 10 years of age who presented with bruising and/or intracranial hemorrhage (ICH) concerning for abuse from February 2021-May 2022 utilizing CAPNET. Cases were classified as bruising without ICH or ICH with or without bruising. Based on AAP guidance, testing was not recommended in cases with concomitant suspicious injuries, inflicted injury history, patterned injury, and low clinical concern for abuse. The primary outcome was adherence to AAP recommendations for testing by study group. We also calculated the frequency of bleeding disorder identification. RESULTS:A total of 2491 children presented with bruising and/or ICH concerning for abuse during the study period. 25.1 % of subjects (468 with bruising and 157 with ICH) were recommended to have testing by AAP criteria. Of these, 10.5 % cases of bruising and 33.1 % cases of ICH completed all AAP recommended testing. Among CAPNET centers, AAP recommended testing completion ranged from 0 to 34.1 % of bruising cases and 4.8-100 % of ICH cases. In total, 0.7 % (18/2491) cases had a newly identified bleeding disorder, with more identified at one center with higher testing rates. CONCLUSIONS:Testing for bleeding disorders based on AAP recommendations frequently did not occur and varied by CAPNET center. Although rare, bleeding disorders were present among cases with testing.

ABSTRACTBackgroundAn emergency department (ED) visit or hospitalization provides an opportunity to identify elder mistreatment and initiate intervention, but this seldom occurs. To address this, we developed the Vulnerable Elder Protection Team (VEPT), a novel interdisciplinary consultation service. We explored the long‐term trajectories of patients receiving VEPT evaluation and intervention.MethodsWe followed up at multiple intervals for 12 months older adults seen by VEPT from 9/1/2020–3/27/2023 with high or moderate concern for mistreatment who were discharged to the community, an elder abuse shelter, or rehabilitation facilities. We collected information through telephone calls to the older adult and others involved. We also analyzed separately cases in which the patient re‐presented to the ED/hospital with VEPT consultation during the follow‐up period.ResultsA total of 157 older adults met criteria for follow‐up, and 30 of these (16.4%) died within 12 months. At 1 month, elder mistreatment was no longer occurring in 47.5% and still occurring but reduced in 20.3%, with 29.7% having no contact with the perpetrator and 17.8% having reduced contact. At 12 months, elder mistreatment was no longer occurring in 60.9% and still occurring but reduced in 14.5%, with 34.8% having no contact with the perpetrator and 17.4% having reduced contact. During the 12‐month follow‐up period, 16 (10.2%) patients re‐presented to the ED with VEPT consultation, with 12 having persistent concern for ongoing elder mistreatment. Reasons included older adults/caregivers not accepting intervention or being willing to separate as well as VEPT reliance on community‐based agencies and programs after discharge.ConclusionsWe observed improved post‐discharge safety for elder mistreatment victims who engaged with the VEPT program, with this increased safety durable over 1 year. Re‐presentations highlighted the complexity of elder mistreatment intervention. Overall, these findings demonstrate the potential value of an ED/hospital‐based elder mistreatment response team, a promising new geriatric care model.

BACKGROUND:There is a lack of consensus on how best to measure injury severity in abusive head trauma in order to predict long-term neurodevelopmental outcomes. OBJECTIVE:We hypothesized that a constellation of injury-related variables along with child and family variables would more accurately predict outcomes in children who have sustained an AHT than the Glasgow Coma Scale (GCS) alone. PARTICIPANTS AND SETTING:In 2012-2020, we enrolled 270 patients (median age 4.6 months) treated for AHT at a large tertiary care children's hospital who survived their injuries and came to a multi-disciplinary follow-up clinic. METHODS:Exploratory analyses examined bivariate relationships of injury severity and child and family variables with neurodevelopmental outcomes, as measured by the Bayley Scales of Infant and Toddler Development, using Pearson correlations, independent samples t-tests, and one-way ANOVAs. These exploratory analyses informed the selection of variables for stepwise multivariate regressions predicting neurodevelopmental outcomes. RESULTS:Stepwise regression revealed that a constellation of injury-related variables including cytotoxic edema, length of intensive care stay, neurosurgical intervention, seizures, intubation, eye injuries, and abnormal spine imaging explained significantly more variance in Bayley scores than GCS alone (14-22 %, all p-values < .01). The largest effect sizes were for measures of hospital course (length of intensive care stay, neurosurgical intervention, seizures, and intubation). Including child and family variables explained an additional 6-10 % of the variance (all p-values < .05). CONCLUSIONS:A constellation of injury-related variables, especially those related to hospital course, was more predictive of neurodevelopment than solely GCS for children with AHT.

ObjectiveCommunity context influences children’s risk for injury. We aimed to measure the explanatory capacity of two ZIP code-level measures—the Child Opportunity Index V.3.0 (COI) and median household income (MHHI)—for rates of paediatric injury hospitalisations.MethodsThis was a retrospective cross-sectional population-based study of children living in 19 US states in 2017. We examined injury hospitalisation rates for three categories: physical abuse among children <5 years, injuries suspicious for abuse among infants <12 months and unintentional injuries among children <18 years. Hospitalisation counts were obtained from the Healthcare Cost and Utilization Project and population data from the US Census. The COI is a multidimensional measure of communities’ education, health and environment and social and economic characteristics. We used pseudo R2values from Poisson regression models to describe the per cent of variance in rates of each injury category explained by the COI and MHHI.ResultsThe COI explained 75.4% of the variability in rates of physical abuse, representing a 13.5% improvement over MHHI. The COI explained 58.5% of the variability in injuries suspicious for abuse, a 20.7% improvement over MHHI. The COI and MHHI explained 85.7% and 85.8% of the variability in unintentional injuries, respectively; results differed when unintentional injuries were stratified by mechanism and age.ImplicationsThe COI had superior explanatory capacity for physical abuse and injuries suspicious for abuse compared with MHHI and was similar for unintentional injury hospitalisations. COI represents a means of accounting for community advantage in paediatric injury data, research and prevention.

BACKGROUND:The association between child abuse pediatric (CAP) assessments and child welfare outcomes is unknown. OBJECTIVE:To determine the association between a CAP determination of the likelihood of physical abuse and change in child placement. We hypothesized that child race would be associated with CAP determination of abuse likelihood and child welfare outcomes. PARTICIPANTS AND SETTING:Children under age ten years with in-person CAP consultation and referral to child protective services for suspected physical abuse at a U.S. pediatric referral center participating in CAPNET, a CAP research network, from 02/2021 to 01/2023. METHODS:We created a series of generalized estimating equations clustered by site, adding covariate blocks representing child characteristics, clinical case factors, and social risk indicators to understand the probability of placement change after CAP consultation. RESULTS:Of 3732 eligible children, 950 (25.5 %) experienced a placement change around a CAP consultation for physical abuse. Adjusting for site, placement change was 28.7 % (25.7-31.8 %) more likely for children with a CAP determination of a high v. lower likelihood of abuse and 6.7 % (3.1-10.4 %) more likely for children of Black/Indigenous v. other race (p < 0.001). These differences persisted with attenuation in fully adjusted models. There was no significant association between CAP determination and child race. CONCLUSIONS:CAP assessment of physical abuse likelihood is strongly associated with the probability of change in child placement. While not associated with CAP assessment of abuse likelihood, Black or Indigenous race is associated with increased probability of placement change even after adjusting for child, case, and social risk factors.