- Email: [email protected]
Identifying Populations at Risk for Child Abuse: A Nationwide Analysis
YJPSU-59422; No of Pages 5 Journal of Pediatric Surgery xxx (xxxx) xxx
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Identifying Populations at Risk for Child Abuse: A Nationwide Analysis☆ Hallie J. Quiroz a,⁎, Joshua Parreco b, Lavanya Easwaran c, Brent Willobee d, Anthony Ferrantella d, Rishi Rattan b, Chad M. Thorson e, Juan E. Sola e, Eduardo A. Perez e a
Dewitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine, RMSB RM 1010, 1600 NW 10th Avenue, Miami, Florida 33136 Division of Trauma Surgery and Surgical Critical Care, Dewitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine c University of Miami Miller School of Medicine d Dewitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine e Division of Pediatric Surgery, Dewitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine b
a r t i c l e
i n f o
Article history: Received 7 September 2019 Accepted 29 September 2019 Available online xxxx Key words: Abuse Psychiatric diagnosis Traumatic injuries
a b s t r a c t Purpose: Child abuse is a national, often hidden, epidemic. The study objective was to determine at-risk populations that have been previously hospitalized prior to their admission for child abuse. Methods: The Nationwide Readmissions Database (NRD) was queried for all children hospitalized for abuse. Outcomes were previous admissions and diagnoses. χ2 analysis was used; significance equals p b 0.05. Results: 31,153 children were hospitalized for abuse (half owing to physical abuse) during the study period. 11% (n = 3487) of these children had previous admissions (one in three to a different hospital), while 3% (n = 1069) had multiple hospitalizations. 60% of prior admissions had chronic conditions, and 12% had traumatic injuries. Children with chronic conditions were more likely to have sexual abuse (89% vs. 57%, p b 0. 001) and emotional abuse (75% vs. 60%, p b 0. 01). 25% of chronic diagnoses were psychiatric, who were also more likely to have sexual and emotional abuse (47% vs. 5.5% and 10% vs. 1%, all p b 0. 001). Conclusion: This study uncovers a hidden population of children with past admissions for chronic conditions, especially psychiatric diagnoses that are significantly associated with certain types of abuse. Improved measures to accurately identify at-risk children must be developed to prevent future childhood abuse and trauma. Level of evidence: Level III. Type of study: Retrospective comparative study. © 2019 Elsevier Inc. All rights reserved.
Child abuse is a major public health concern that impacts around 650,000 children annually in the United States. It accounts for more than 1500 child deaths each year [1,2] and despite increasing efforts at surveillance and prevention, [3–6] this epidemic of child mistreatment continues to plague society. Over the last two decades, the estimated number of child protective services investigations has increased to 3.5 million per year [8] and 5 children die every day as a result of abuse [7]. The Federal Abuse and Treatment Act provides minimum standards for defining child maltreatment, however individual state laws have varying definitions of abuse with some states requiring “serious bodily injury” or “severe pain” to constitute abuse [9]. Clinical indicators of abuse can vary widely and as such, many children suffer child abuse,
☆ How this study should change care: This analysis suggests that certain children with past hospital admissions such as psychiatric diagnoses or traumatic injuries are at risk for future child abuse admissions. Further improvements in child abuse detection are needed to decrease future childhood abuse and trauma. ⁎ Corresponding author. Tel.: +1 316 253-8950. E-mail address: [email protected] (H.J. Quiroz).
hidden in plain sight, with little identifiable risk factors using current identification strategies [10]. Early identification of abuse is challenging, and physicians often miss the evidence of abuse that can lead to an appropriate intervention [1]. Successfully identifying abuse from particular injury patterns has been widely reported [11–13], yet despite this, abuse has been initially missed in 20%–30% of cases [12,14]. Consistent evidence has also shown that victims of child abuse have documented contact with the health care system, especially in the emergency department setting. These missed instances of abuse are typically only identified when a subsequent abusive injury occurs, [15] which occurs in 35% of cases with an estimated 5%–10% of these children who die as a result of an initial missed intervention [14]. Despite these alarming statistics and the notion that hospitals are a crucial entry point for many children in the healthcare system, there have been no national studies evaluating hospital encounters prior to an admission for child abuse. Previous studies of hospitalization for child abuse have been limited by an inability to track prior hospital encounters across different hospitals [15–17]. The purpose of this study was to identify the characteristics of nationwide hospital admissions
https://doi.org/10.1016/j.jpedsurg.2019.09.069 0022-3468/© 2019 Elsevier Inc. All rights reserved.
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069
2
H.J. Quiroz et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
occurring before an admission for child abuse, including admissions across different hospitals. We hypothesized that children with past hospitalizations might reveal hidden risk factors for child abuse, in hopes to improve current identification of childhood abuse.
weighted for national estimates according to HCUP standards [20]. The Institutional Review Board of the University of Miami deemed this retrospective study to be exempt from full review. Statistical analyses were performed using SPSS Statistics version 24, IBM Corporation, Armonk, New York.
1. Methods 2. Results The largest Federal-State-Industry partnership providing national encounter-level healthcare data in the US is the Healthcare Utilization Project (HCUP). The project provides researchers with a family of databases including the Nationwide Inpatient Sample (NIS) and the recently released, Nationwide Readmissions Database (NRD). The NRD extends the capability of the NIS by providing the unique capability of tracking patients across hospitalizations at different hospitals. The NRD data are also extensively processed to ensure that readmissions are counted as accurately as possible. This is especially important for trauma studies that utilize population-level data since it is possible that admissions related to the index injury could be miscoded as a new injury [18,19]. The NRD also greatly reduces the likelihood of this type of miscoding by collapsing multiple records into one if they involve a transfer or same-day event such as discharge and admission between different hospitals [20]. For this study, the NRD for 2010–2014 was queried for all admissions with an age less than 18 years and an International Classification of Diseases, Ninth Revision, (ICD-9) Clinical Modification diagnosis code corresponding to abuse (99,550, 99,551, 99,552, 99,553, 99,554, 99,555, 99,559, 99,580, 99,581, 99,582, 99,583, 99,584, 99,585, V7181) [21]. Admissions other than for birth (ICD-9 diagnosis code V3*) prior to the first admission for abuse were identified. The primary outcome was prior admission and the secondary outcomes included multiple (greater than one) prior admissions, prior admission at a different hospital, and primary diagnoses for past admissions. Analysis was performed for each outcome using all variables during the abuse admission. Categorical variables were compared using a chi-squared test and significance defined as a p-value less than 0.05. Results were
There were 31,153 children admitted for abuse during the study period with 11% (n = 3487) having a prior admission. Of all patients with prior admissions, 31% (n = 1069) had multiple prior admissions and 37% (n = 1290) had a prior admission at a different hospital. Median age for the study population was 0 with an interquartile range of 0–5. Those b1 year were most commonly affected by abuse (50%) followed by 1–6 years (28%). The patient characteristics overall and for each outcome are shown in Table 1. 2.1. Abuse types Physical Abuse: The most common type of abuse in this study was physical abuse (54%, n = 16,779). Upon review of past admissions, physically abused children had the highest rate of prior admissions at a different hospital (48%, p b 0.001, Table 1). When compared to older patients, children b 13 years old were more likely to be admitted for physical abuse (55% vs. 27%), p b 0.001, Table 2. Sexual Abuse: In children with past admissions, children ≥13 years comprised a larger portion of eventual sexual abuse when compared to children b13 years (76% vs. 24%), p b 0.001, Table 2. Neglect: Neglect is the second most common type of abuse identified in our study at 15% (n = 4601) and similarly had the second highest rate of prior admissions (17%, n = 759), p b 0.001, Table 1. Children with prior admissions and eventual neglect were more likely to be b 13 years old as opposed to those 13 or older (84% vs. 16%, p b 0.001 Table 2.) Emotional Abuse: Children with recorded emotional abuse had the highest rates of prior hospitalization (31%, n = 81, p b 0.001, Table 1). When analyzing the subset of children
Table 1 Overall patient characteristics for prior admission, multiple prior admissions, and prior admission at a different hospital. Characteristic
Total
Prior admission
Multiple prior admissions
n (%)
n (%)
n (%)
Total Female Male Abuse type
31,153 (100.0) 14,472 (46.5) 16,681 (53.5) 16,779 (53.9) 2644 (8.5) 4601 (14.8) 266 (0.9) 6863 (22.0) 5042 (16.2)
3487 (11.2) 1520 (10.5) 1967 (11.8) 1701 (10.1) 290 (11.0) 759 (16.5) 82 (30.8) 655 (9.5) 574 (11.4)
3639 (11.7)
502 (13.8)
301 (1.0) 250 (0.8) 21,920 (70.4) 15,633 (50.2) 8607 (27.6) 2621 (8.4) 4293 (13.8) 5213 (16.7) 24,042 (77.2) 1845 (5.9) 10,939 (35.1) 8528 (27.4) 6783 (21.8) 4245 (13.6) 658 (2.1)
38 (12.6) 38 (15.2) 2336 (10.7) 1724 (11.0) 622 (7.2) 386 (14.7) 755 (17.6) 885 (17.0) 2416 (10.1) 176 (9.5) 1292 (11.8) 971 (11.4) 520 (7.7) 662 (15.6) 43 (6.5)
Perpetrator
Age group (years)
Primary payer
Median household income quartile
Physical Sexual Neglect Emotional Other/unspecified Male Parental Figure (father, stepfather, or boyfriend) Female Parental Figure (mother, stepmother, or girlfriend) Sibling or child Grandparent Other/unspecified b1 1–6 7–12 13–18 Private insurance Medicaid/Medicare Self-pay/No charge/Other 1st ($0–40,999) 2nd ($41,000-50,999) 3rd ($51,000-66,999) 4th ($67,000 +) Unknown
p b0.001 b0.001
b0.001
1069 (30.7) 399 (26.3) 670 (34) 537 (31.6) 90 (31.0) 230 (30.3) 28 (34.2) 184 (28.1) 115 (20.0)
Prior admission at a different hospital p b0.001 0.519
b0.001
n (%) 1290 (37.0) 514 (33.8) 776 (39.5) 817 (48.0) 72 (24.8) 179 (23.6) 25 (30.5) 198 (30.2) 328 (57.1)
167 (33.3)
b0.001
*
b0.001
b0.001
* * 768 (32.9) 488 (28.3) 175 (28.1) 131 (33.9) 275 (36.4) 362 (40.9) 663 (27.4) 38 (21.6) 328 (25.4) 204 (21.0) 164 (31.5) 370 (55.9)
b0.001
0.001
129 (25.7)
* * b0.001
p
b0.001
b0.001
b0.001
817 (35.0) 843 (48.9) 143 (23.0) 115 (29.8) 189 (25.0) 563 (63.6) 673 (27.9) 50 (28.4) 394 (30.5) 413 (42.5) 118 (22.7) 350 (52.9) 15 (34.9)
b0.001
b0.001
b0.001
Cells marked with an asterisk (*) represent actual values censored from publication in accordance with the Healthcare Cost and Utilization Project Data Use Agreement.
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069
H.J. Quiroz et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
3
Table 2 Comparison between age & socioeconomic cohorts with type of abuse in patients who had previous hospitalizations. Age Group
Abuse type
Physical -Yes -No Sexual -Yes -No Neglect -Yes -No Emotional -Yes -No
Highest household income
b13 yrs n = 2732 n (%)
≥13 yrs n = 755 n (%)
1494(88) 1238(70)
206(12) 549(30)
87(24) 2645(85)
270(76) 485(16)
666(84) 2066(77)
131(16) 624(23)
42(41) 2690(80)
60(59) 695(21)
p-value
Yes n (%)
No n (%)
498(75) 164(25)
1203(43) 1623(57)
25(4) 637(96)
333(12) 2493(88)
72(11) 590(89)
726(26) 2100(74)
16(3) 645(97)
86(3) 2740(97)
b0.001
Lowest household income p-value
Yes n (%)
No n (%)
471(37) 821(64)
1230(56) 965(44)
189(15) 1103(85)
169(8) 2027(92)
379(29) 913(71)
419(19) 1777(81)
35(3) 1258(97)
68(3) 2128(97)
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
p-value
n.s.
n.s.
yrs: Years, Highest household income = $67,000+, Lowest household income $1–40,999.
with these past admissions, the ones with eventual emotional abuse tended to be older (13 years or older) when compared with children younger than 13 years (59% vs. 41%), p b 0.001, Table 2.
for chronic health conditions were for psychiatric conditions. Compared to those without psychiatric conditions, they were more likely to have sexual (47% vs. 6%) and emotional abuse (10% vs 2%), both p b 0.001, Table 5.
2.2. Socioeconomic status 3. Discussion One-third of patients (35%) were in the lowest median household income quartile ($1–$40,999). The wealthiest median household income quartile ($67,000 +) was found to have higher rates of prior admission (16%), multiple prior admissions (56%, p b 0.001), and prior admission at a different hospital (53%), all p b 0.001, Table 1. For children with previous hospitalizations, those in the highest median household income had higher rates of physical abuse compared to those in lower income (75% vs. 43%, p b 0.001), Table 2. Conversely children in the lowest median household income cohort had higher rates of sexual abuse (15% vs. 8%, p b 0.001) and neglect (11% vs. 3%, p b 0.001) compared to those in higher income households, Table 2. 2.3. Past hospitalizations Patients 13–18 years had the highest rate of prior admissions (18%, p b 0.001) and multiple prior admissions (36%, p b 0.001). The rate of prior admission at a different hospital was highest in those b 1 year (49%) when compared to older patients (23%–30%), p b 0.001. The rate of multiple prior admissions was not significantly different between the abuse types (p = 0.519). Patients with multiple prior admissions more often sustained abuse from a female parental figure (33%) compared to children who sustained abuse from a male parental figure (20%), p b 0.001. Those abused by a male parental figure however were more likely to have a prior admission to a different hospital (57%) when compared to those abused by a female parental figure (26%), p b 0.001, Table 1. Nearly two-thirds (60%, n = 2091) of prior admissions were for chronic health conditions and 12% (n = 399) were for traumatic injuries, Table 3. The most common chronic condition was psychiatric (26%) followed by gastrointestinal (13%), neonatal conditions (10%), failure to thrive/malnutrition (8%) and neurologic conditions (7%). Orthopedic trauma made up the majority (48%) of traumatic conditions followed by head trauma (31%), burns (13%), and poisoning/self-harm (5%). 70% of children with past traumatic injuries (n = 276) would eventually go on to be admitted for physical abuse, while 30% would later sustain other types of abuse (p b 0.001, Table 3). Of these children, 65% had previously sustained orthopedic fractures and 25% sustained closed-head trauma. Children with chronic health conditions were more likely to experience emotional abuse (75%) and sexual abuse (12%), both p b 0.01 (Table 4). A quarter (26%, n = 537) of admissions
This study represents the first nationwide evaluation of pediatric hospital admissions prior to abuse including admissions to a different hospital while exploring past admission diagnoses. The high rate of prior admission (11%) suggests that at least a portion of these prior
Table 3 Characteristics of past-admissions in children admitted for nonaccidental trauma.
Diagnoses at prior admission Chronic Conditions Acute Conditions Trauma Unknown Chronic Conditions groups Neonatal/Perinatal Conditions Psychiatric Hematologic / Malignancies Obstetrics / Gynecology Transplant Gastrointestinal disorders Postoperative conditions Heart conditions Respiratory conditions Neurological conditions Failure to Thrive / Malnutrition Endocrine Other / Miscellaneous Acute Conditions groups Infectious Dehydration/Electrolyte abnormalities Neonatal event Altered mental status Other / Miscellaneous Traumatic Conditions groups Burns Drowning / Near Drowning Poisoning / Self-harm Orthopedic trauma Head trauma Blunt trauma Sexual trauma
n
%
2091 989 399 *
59.9 28.3 11.5 b1
200 537 104 34 19 273 26 28 133 149 157 120 311
9.6 25.7 5.0 1.6 0.9 13.1 1.2 1.3 6.3 7.1 7.5 5.7 14.9
1047 47 48 *8 135
81.5 3.7 3.8 b1 10.5
51 * 20 190 125 * *
12.8 b1 5.0 47.6 31.4 b1.5 b1.5
Cells marked with an asterisk (*) represent actual values censored from publication in accordance with the Healthcare Cost and Utilization Project Data Use Agreement.
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069
4
H.J. Quiroz et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
Table 4 Comparison between prior admission diagnoses with types of eventual abuse. Acute n = 997
Abuse Type
Physical -Yes -No Sexual -Yes -No Neglect -Yes -No Emotional -Yes -No
Trauma n = 399
Yes n (%)
No n (%)
454(27) 543(30)
1247(73) 1244(70)
25(7) 972(31)
333(93) 2158(69)
288(36) 709(26)
510(64) 1981(74)
26(26) 970(29)
76(75) 2415(71)
p-value
Chronic n = 2091
Yes n (%)
No n (%)
276(16) 123(7)
1425(84) 1663(93)
16(4) 384(13)
342(96) 2746(87)
31(4) 368(14)
766(96) 2322(86)
*(b1) 399(12)
102(100) 2986(88)
p-value
Yes n (%)
No n (%)
965(57) 1126(63)
736(43) 661(37)
317(12) 1774(57)
41(88) 1356(43)
486(61) 1604(60)
311(39) 1086(40)
76(75) 2015(59)
26(26) 1371(41)
b0.001
0.02
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
n.s.
b0.001
n.s.
p-value
0.002
Cells marked with an asterisk (*) represent actual values censored from publication in accordance with the Healthcare Cost and Utilization Project Data Use Agreement.
admissions were because of unrecognized child abuse or for diagnoses that may be correlated with abuse. This potentially suggests that despite having past hospitalizations, either children are not undergoing proper screening for abuse or current screening tools are ineffective for certain subsets. Some groups have indicated a “child-abuse” checklist with further screening by child-abuse experts which could potentially limit the missed cases of child abuse seen by emergency department ED providers [24]. This study presents the novel finding of a 37% rate of prior admissions occurring at a different hospital. This is higher than the previously reported rates of readmission to a different hospital after trauma from the NRD (10%–26%) [22,23]. The rate of prior admission to a different hospital was much higher for physical abuse (48%). Most strikingly, those abused by a male parental figure were two times more likely to have a prior admission to a different hospital than those abused by females. These findings have important implications for prevention through identifying the most common patterns of abuse. This could indicate that the perpetrators are evading detection by presentation to a different hospital, particularly when certain injury patterns or frequency of injury will increase suspicion by healthcare workers. Further novel findings are that maternal perpetrators had a higher rate of multiple prior admissions (along with grandparents) when compared to other perpetrators. Previous studies of mothers of neglected children found that these women tended to believe that they had been caring for their children well and were surprised when their child became ill [25,26]. Unlike paternal physical abuse, there is no pressure to present to a different hospital if the abuse is unacknowledged by the perpetrator. These findings are consistent with previous reports that suggest that physical abuse is more often carried out by fathers while neglect is more often associated with mothers [25,27,28] Thus, the
pattern for negligent mothers appears to be multiple prior admissions at the same hospital, indicating that the “cry for help” was missed. It has been widely reported that both abuse and neglect are more prevalent with a lower socioeconomic status [29–31]. Despite this study's support of prior work regarding the overall rate of child abuse in relation to socioeconomic status, this is the first study to reveal the subtle differences in abuse category with regard to socioeconomic status. Namely, patients with prior admissions and highest socioeconomic status have higher rates of physical abuse than other socioeconomic strata. This could be because of either a higher likelihood to present for hospital admission or higher likelihood of detection and treatment for their injuries. Further studies are required to determine a causal relationship. Our study also aligns with previous reports that neglect is associated with lower socioeconomic status [29,32,33]; however, our data reveal a hidden association between lowest household income quartile with sexual abuse. Sexual abuse is a difficult diagnosis to make; it may not necessarily be the reason for seeking medical care and is often unreported. We suspect that the true incidence is underestimated in the sample. Clinicians should remain vigilant and always consider sexual abuse given these findings. To try to determine potential harbingers of eventual abuse, the indications for past hospitalizations were evaluated. Interestingly, a large portion of these admissions were for chronic conditions (60%) and past traumatic injuries (12%). The most common traumatic injuries identified on past admissions were orthopedic injuries, closed-head trauma, and burns. Additionally, having past traumatic injury was significantly associated with eventual hospitalizations for physical abuse, which indicates that a portion of the children with past abuse-related injuries [12,13,18,35] and subsequent hospitalization for abuse were
Table 5 Subgroup analysis of psychiatric conditions and types of eventual abuse. Psychiatric Diagnosis Yes Abuse Type
Physical -Yes -No Sexual -Yes -No Neglect -Yes -No Emotional -Yes -No
No
p-value b0.001
180(34) 354(66)
428(40) 632(60)
253(47) 281(53)
58(6) 1003(94)
46(9) 488(91)
366(35) 694(66)
51(10) 483(90)
15(1.5) 1045(98.5)
b0.001
b0.001
b0.001
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069
H.J. Quiroz et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
likely cases of missed abuse. This failure of our healthcare system to detect abuse and attempt prevention of further abuse reveals an area for improvement in definitive identification and management of children with abuse. Our study also reveals that one in four children with chronic conditions have past admissions for psychiatric conditions. Child sexual abuse is linked to an increase in depression, posttraumatic stress disorder, other psychiatric disorders, and suicidal ideations [36,37]. According to our data, children with past psychiatric admissions have higher rates of admissions for sexual abuse and emotional abuse. Thus, children presenting with psychiatric conditions represent an important population where careful consideration should be given to current or past child abuse and appropriate steps made for future prevention. Our data also revealed that the rates of readmission for abuse are much higher than other studies in trauma patients [19,34]. Eleven percent of patients treated for child abuse had prior admissions, and an astounding 37% of these children were admitted at a different hospital. The importance of this finding cannot be overstated — at least one in three children is treated for abuse at multiple hospitals. These at-risk patients are difficult to identify in real time owing to the lack of communications between hospitals and no centralized state or national medical record system. The limitations of this study include those of retrospectively collected administrative databases like the NRD such as errors in data sampling, collection measures, usage of the ICD-9 coding scheme, and possible administrative errors during data entry. Additionally, the NRD cannot follow patients across years or if the patients are readmitted across state lines. HCUP estimates that less than 5% of readmissions occur across state lines [20]. Despite these limitations the NRD is the largest database that allows for following readmissions across different hospitals in the US. This is particularly important in victims of child abuse who often present to nonindex hospitals [23,34]. 4. Conclusions Utilization of the NRD provides the unique capability of tracking patients across hospital admissions and reveals the pattern of risk factors that could lead to early identification and prevention of child abuse. More than 10% of patients admitted for child abuse had prior admissions, with some highly likely to have been prior abuse, representing an enormous opportunity for improvement on current identification and preventative measures. A large portion of these past admissions were at a different hospital, indicating the paramount need for a standardized medical record easily transferable between hospitals. This study also uncovered a hidden population of children with past admissions for past traumatic injuries and chronic conditions, especially psychiatric diagnoses that are at significant risk. Improved measures to accurately identify children at risk must be further studied to decrease future childhood abuse and trauma. 5. Discussion Identifying Populations at Risk for Child Abuse: A Nationwide Analysis. Presenter: Hallie Quiroz. Q. Jon Groner from Columbus, Ohio: Great work. My question is, are you able to look at ethnic backgrounds and race because there's other data that shows that white kids are a lot more likely to be missed actually? Hallie Quiroz: Right. So actually as opposed to the kid database where you can look at racial differences, the NRD actually does not breakdown racial components of that. So that would be a weakness of the NRD. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
5
Disclosure The authors have no relevant disclosures.
References [1] Schilling S, Christian CW. Child physical abuse and neglect. Child Adolesc Psychiatr Clin N Am 2014;2:309–19 (ix). [2] Saini SM, Hoffmann CR, Pantelis C. et. al. Systematic review and critical appraisal of child abuse measurement instruments. Psychiatry Res 2019:106–13. [3] Whitaker DJ, Lutzker JR, Shelley GA. Child maltreatment prevention priorities at the Centers for Disease Control and Prevention. Child Maltreat 2005(3):245–59. [4] M C. The Changing Focus of Child Maltreatment Research and Practice Within Psychology. J Soc Issues 2006;4:663–84. [5] Schnitzer PG, Covington TM, Wirtz SJ, et al. Public health surveillance of fatal child maltreatment: analysis of 3 state programs. Am J Public Health 2008(2):296–303. [6] Gilbert R, Kemp A, Thoburn J, et al. Recognising and responding to child maltreatment. Lancet 2009;9658:167–80. [7] Sanada F, Taniyama Y, Muratsu J, et al. Gene-Therapeutic Strategies Targeting Angiogenesis in Peripheral Artery Disease. Medicines (Basel, Switzerland) 2018:2. [8] Al-Damluji MS, Dharmarajan K, Zhang W, et al. Readmissions after carotid artery revascularization in the Medicare population. J Am Coll Cardiol 2015(14):1398–408. [9] Christian CW. Committee on Child A, Neglect AAoP. The evaluation of suspected child physical abuse. Pediatrics 2015;5:e1337–54. [10] Normandin PA. Hidden Child Abuse Identification in the Emergency Department. J Emerg Nurs 2016(2):181–2. [11] Dorfman MVMJ, Feldman KW, Farris R, et al. ExTRA Investigators Oral injuries and occult harm in children evaluated for abuse. Arch Dis Child 2018(8):47–52. [12] Lindberg DM, Beaty B, Juarez-Colunga E, et al. Testing for Abuse in Children With Sentinel Injuries. Pediatrics 2015(5):831–8. [13] Paine CW, Wood JN. Skeletal surveys in young, injured children: A systematic review. Child Abuse Negl 2018:237–49. [14] Ravichandiran N, Schuh S, Bejuk M, et al. Delayed identification of pediatric abuserelated fractures. Pediatrics 2010(1):60–6. [15] Jenny C, Hymel KP, Ritzen A, et al. Analysis of missed cases of abusive head trauma. JAMA 1999(7):621–6. [16] Leventhal JM, Martin KD, Asnes AG. Incidence of fractures attributable to abuse in young hospitalized children: results from analysis of a United States database. Pediatrics 2008(3):599–604. [17] Wood JN, Medina SP, Feudtner C, et al. Local macroeconomic trends and hospital admissions for child abuse, 2000-2009. Pediatrics 2012(2):e358–64. [18] Kaufman E, Rising K, Wiebe DJ, et al. Recurrent violent injury: magnitude, risk factors, and opportunities for intervention from a statewide analysis. Am J Emerg Med 2016(9):1823–30. [19] Parreco J, Rattan R. Machine Learning Models for Prediction of Reinjury After Penetrating Trauma. JAMA Surg 2018(2):184–6. [20] Quality AfHRa. Introduction to the HCUP Nationwide Readmissions. Database 2017. [21] Allison Russo C HM, M. SW, Owens PL. Hospital Stays Related to Child Maltreatment, https://www.hcup-us.ahrq.gov/reports/statbriefs/sb49.pdf; 2005 2019]. [22] Sittig JS, Uiterwaal CSPM, Moons KGM, et al. Value of systematic detection of physical child abuse at emergency rooms: a cross-sectional diagnostic accuracy study. BMJ Open 2016(3):e010788. [23] Parreco JBJ, Cortolillo N, Namias N, et al. Risk factors and costs associated with nationwide nonelective readmission after trauma. Eastern Association for the Surgery of Trauma Annual Scientific Assembly; 2017 Hollywood, FL. [24] Buicko JL, Parreco J, Willobee BA, et al. Risk factors for nonelective 30-day readmission in pediatric assault victims. J Pediatr Surg 2017(10):1628–32. [25] Golden MH, Samuels MP, Southall DP. How to distinguish between neglect and deprivational abuse. Arch Dis Child 2003;2:105–7. [26] Pignatelli AM, Wampers M, Loriedo C, et al. Childhood neglect in eating disorders: A systematic review and meta-analysis. J Trauma Dissociation 2017(1):100–15. [27] Lee SJ, Bellamy JL, Guterman NB. Fathers, physical child abuse, and neglect: advancing the knowledge base. Child Maltreat 2009(3):227–31. [28] Konishi A, Yoshimura B. Child abuse and neglect by mothers hospitalized for mental disorders. Arch Womens Ment Health 2015(6):833–4. [29] Eckenrode J, Smith EG, McCarthy ME, et al. Income inequality and child maltreatment in the United States. Pediatrics 2014(3):454–61. [30] Belsky J. Etiology of child maltreatment: a developmental-ecological analysis. Psychol Bull 1993;3:413–34. [31] Pelton LH. Child abuse and neglect: the myth of classlessness. Am J Orthop 1978(4): 608–17. [32] Eckenrode J, Smith EG, McCarthy ME, et al. Income Inequality and Child Maltreatment in the United States. Pediatrics 2014(3):454. [33] Kotch JB, Browne DC, Ringwalt CL, et al. Risk of child abuse or neglect in a cohort of low-income children. Child Abuse Negl 1995(9):1115–30. [34] Block RW. No Surprise: The Rate of Fatal Child Abuse and Neglect Fatalities Is Related to Poverty. Pediatrics 2017(5). [35] Pawlik MC, Kemp A, Maguire S. et. al. Children with burns referred for child abuse evaluation: Burn characteristics and co-existent injuries. Child Abuse Negl 2016:52–61. [36] O'Brien BS, Sher L. Child sexual abuse and the pathophysiology of suicide in adolescents and adults. International journal of adolescent medicine and health 2013(3):201–5. [37] Molnar BE, Berkman LF, Buka SL. Psychopathology, childhood sexual abuse and other childhood adversities: relative links to subsequent suicidal behaviour in the US. Psychol Med 2001(6):965–77.
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Identifying Populations at Risk for Child Abuse: A Nationwide Analysis☆ Hallie J. Quiroz a,⁎, Joshua Parreco b, Lavanya Easwaran c, Brent Willobee d, Anthony Ferrantella d, Rishi Rattan b, Chad M. Thorson e, Juan E. Sola e, Eduardo A. Perez e a
Dewitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine, RMSB RM 1010, 1600 NW 10th Avenue, Miami, Florida 33136 Division of Trauma Surgery and Surgical Critical Care, Dewitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine c University of Miami Miller School of Medicine d Dewitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine e Division of Pediatric Surgery, Dewitt Daughtry Family Department of Surgery, University of Miami Miller School of Medicine b
a r t i c l e
i n f o
Article history: Received 7 September 2019 Accepted 29 September 2019 Available online xxxx Key words: Abuse Psychiatric diagnosis Traumatic injuries
a b s t r a c t Purpose: Child abuse is a national, often hidden, epidemic. The study objective was to determine at-risk populations that have been previously hospitalized prior to their admission for child abuse. Methods: The Nationwide Readmissions Database (NRD) was queried for all children hospitalized for abuse. Outcomes were previous admissions and diagnoses. χ2 analysis was used; significance equals p b 0.05. Results: 31,153 children were hospitalized for abuse (half owing to physical abuse) during the study period. 11% (n = 3487) of these children had previous admissions (one in three to a different hospital), while 3% (n = 1069) had multiple hospitalizations. 60% of prior admissions had chronic conditions, and 12% had traumatic injuries. Children with chronic conditions were more likely to have sexual abuse (89% vs. 57%, p b 0. 001) and emotional abuse (75% vs. 60%, p b 0. 01). 25% of chronic diagnoses were psychiatric, who were also more likely to have sexual and emotional abuse (47% vs. 5.5% and 10% vs. 1%, all p b 0. 001). Conclusion: This study uncovers a hidden population of children with past admissions for chronic conditions, especially psychiatric diagnoses that are significantly associated with certain types of abuse. Improved measures to accurately identify at-risk children must be developed to prevent future childhood abuse and trauma. Level of evidence: Level III. Type of study: Retrospective comparative study. © 2019 Elsevier Inc. All rights reserved.
Child abuse is a major public health concern that impacts around 650,000 children annually in the United States. It accounts for more than 1500 child deaths each year [1,2] and despite increasing efforts at surveillance and prevention, [3–6] this epidemic of child mistreatment continues to plague society. Over the last two decades, the estimated number of child protective services investigations has increased to 3.5 million per year [8] and 5 children die every day as a result of abuse [7]. The Federal Abuse and Treatment Act provides minimum standards for defining child maltreatment, however individual state laws have varying definitions of abuse with some states requiring “serious bodily injury” or “severe pain” to constitute abuse [9]. Clinical indicators of abuse can vary widely and as such, many children suffer child abuse,
☆ How this study should change care: This analysis suggests that certain children with past hospital admissions such as psychiatric diagnoses or traumatic injuries are at risk for future child abuse admissions. Further improvements in child abuse detection are needed to decrease future childhood abuse and trauma. ⁎ Corresponding author. Tel.: +1 316 253-8950. E-mail address: [email protected] (H.J. Quiroz).
hidden in plain sight, with little identifiable risk factors using current identification strategies [10]. Early identification of abuse is challenging, and physicians often miss the evidence of abuse that can lead to an appropriate intervention [1]. Successfully identifying abuse from particular injury patterns has been widely reported [11–13], yet despite this, abuse has been initially missed in 20%–30% of cases [12,14]. Consistent evidence has also shown that victims of child abuse have documented contact with the health care system, especially in the emergency department setting. These missed instances of abuse are typically only identified when a subsequent abusive injury occurs, [15] which occurs in 35% of cases with an estimated 5%–10% of these children who die as a result of an initial missed intervention [14]. Despite these alarming statistics and the notion that hospitals are a crucial entry point for many children in the healthcare system, there have been no national studies evaluating hospital encounters prior to an admission for child abuse. Previous studies of hospitalization for child abuse have been limited by an inability to track prior hospital encounters across different hospitals [15–17]. The purpose of this study was to identify the characteristics of nationwide hospital admissions
https://doi.org/10.1016/j.jpedsurg.2019.09.069 0022-3468/© 2019 Elsevier Inc. All rights reserved.
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069
2
H.J. Quiroz et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
occurring before an admission for child abuse, including admissions across different hospitals. We hypothesized that children with past hospitalizations might reveal hidden risk factors for child abuse, in hopes to improve current identification of childhood abuse.
weighted for national estimates according to HCUP standards [20]. The Institutional Review Board of the University of Miami deemed this retrospective study to be exempt from full review. Statistical analyses were performed using SPSS Statistics version 24, IBM Corporation, Armonk, New York.
1. Methods 2. Results The largest Federal-State-Industry partnership providing national encounter-level healthcare data in the US is the Healthcare Utilization Project (HCUP). The project provides researchers with a family of databases including the Nationwide Inpatient Sample (NIS) and the recently released, Nationwide Readmissions Database (NRD). The NRD extends the capability of the NIS by providing the unique capability of tracking patients across hospitalizations at different hospitals. The NRD data are also extensively processed to ensure that readmissions are counted as accurately as possible. This is especially important for trauma studies that utilize population-level data since it is possible that admissions related to the index injury could be miscoded as a new injury [18,19]. The NRD also greatly reduces the likelihood of this type of miscoding by collapsing multiple records into one if they involve a transfer or same-day event such as discharge and admission between different hospitals [20]. For this study, the NRD for 2010–2014 was queried for all admissions with an age less than 18 years and an International Classification of Diseases, Ninth Revision, (ICD-9) Clinical Modification diagnosis code corresponding to abuse (99,550, 99,551, 99,552, 99,553, 99,554, 99,555, 99,559, 99,580, 99,581, 99,582, 99,583, 99,584, 99,585, V7181) [21]. Admissions other than for birth (ICD-9 diagnosis code V3*) prior to the first admission for abuse were identified. The primary outcome was prior admission and the secondary outcomes included multiple (greater than one) prior admissions, prior admission at a different hospital, and primary diagnoses for past admissions. Analysis was performed for each outcome using all variables during the abuse admission. Categorical variables were compared using a chi-squared test and significance defined as a p-value less than 0.05. Results were
There were 31,153 children admitted for abuse during the study period with 11% (n = 3487) having a prior admission. Of all patients with prior admissions, 31% (n = 1069) had multiple prior admissions and 37% (n = 1290) had a prior admission at a different hospital. Median age for the study population was 0 with an interquartile range of 0–5. Those b1 year were most commonly affected by abuse (50%) followed by 1–6 years (28%). The patient characteristics overall and for each outcome are shown in Table 1. 2.1. Abuse types Physical Abuse: The most common type of abuse in this study was physical abuse (54%, n = 16,779). Upon review of past admissions, physically abused children had the highest rate of prior admissions at a different hospital (48%, p b 0.001, Table 1). When compared to older patients, children b 13 years old were more likely to be admitted for physical abuse (55% vs. 27%), p b 0.001, Table 2. Sexual Abuse: In children with past admissions, children ≥13 years comprised a larger portion of eventual sexual abuse when compared to children b13 years (76% vs. 24%), p b 0.001, Table 2. Neglect: Neglect is the second most common type of abuse identified in our study at 15% (n = 4601) and similarly had the second highest rate of prior admissions (17%, n = 759), p b 0.001, Table 1. Children with prior admissions and eventual neglect were more likely to be b 13 years old as opposed to those 13 or older (84% vs. 16%, p b 0.001 Table 2.) Emotional Abuse: Children with recorded emotional abuse had the highest rates of prior hospitalization (31%, n = 81, p b 0.001, Table 1). When analyzing the subset of children
Table 1 Overall patient characteristics for prior admission, multiple prior admissions, and prior admission at a different hospital. Characteristic
Total
Prior admission
Multiple prior admissions
n (%)
n (%)
n (%)
Total Female Male Abuse type
31,153 (100.0) 14,472 (46.5) 16,681 (53.5) 16,779 (53.9) 2644 (8.5) 4601 (14.8) 266 (0.9) 6863 (22.0) 5042 (16.2)
3487 (11.2) 1520 (10.5) 1967 (11.8) 1701 (10.1) 290 (11.0) 759 (16.5) 82 (30.8) 655 (9.5) 574 (11.4)
3639 (11.7)
502 (13.8)
301 (1.0) 250 (0.8) 21,920 (70.4) 15,633 (50.2) 8607 (27.6) 2621 (8.4) 4293 (13.8) 5213 (16.7) 24,042 (77.2) 1845 (5.9) 10,939 (35.1) 8528 (27.4) 6783 (21.8) 4245 (13.6) 658 (2.1)
38 (12.6) 38 (15.2) 2336 (10.7) 1724 (11.0) 622 (7.2) 386 (14.7) 755 (17.6) 885 (17.0) 2416 (10.1) 176 (9.5) 1292 (11.8) 971 (11.4) 520 (7.7) 662 (15.6) 43 (6.5)
Perpetrator
Age group (years)
Primary payer
Median household income quartile
Physical Sexual Neglect Emotional Other/unspecified Male Parental Figure (father, stepfather, or boyfriend) Female Parental Figure (mother, stepmother, or girlfriend) Sibling or child Grandparent Other/unspecified b1 1–6 7–12 13–18 Private insurance Medicaid/Medicare Self-pay/No charge/Other 1st ($0–40,999) 2nd ($41,000-50,999) 3rd ($51,000-66,999) 4th ($67,000 +) Unknown
p b0.001 b0.001
b0.001
1069 (30.7) 399 (26.3) 670 (34) 537 (31.6) 90 (31.0) 230 (30.3) 28 (34.2) 184 (28.1) 115 (20.0)
Prior admission at a different hospital p b0.001 0.519
b0.001
n (%) 1290 (37.0) 514 (33.8) 776 (39.5) 817 (48.0) 72 (24.8) 179 (23.6) 25 (30.5) 198 (30.2) 328 (57.1)
167 (33.3)
b0.001
*
b0.001
b0.001
* * 768 (32.9) 488 (28.3) 175 (28.1) 131 (33.9) 275 (36.4) 362 (40.9) 663 (27.4) 38 (21.6) 328 (25.4) 204 (21.0) 164 (31.5) 370 (55.9)
b0.001
0.001
129 (25.7)
* * b0.001
p
b0.001
b0.001
b0.001
817 (35.0) 843 (48.9) 143 (23.0) 115 (29.8) 189 (25.0) 563 (63.6) 673 (27.9) 50 (28.4) 394 (30.5) 413 (42.5) 118 (22.7) 350 (52.9) 15 (34.9)
b0.001
b0.001
b0.001
Cells marked with an asterisk (*) represent actual values censored from publication in accordance with the Healthcare Cost and Utilization Project Data Use Agreement.
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069
H.J. Quiroz et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
3
Table 2 Comparison between age & socioeconomic cohorts with type of abuse in patients who had previous hospitalizations. Age Group
Abuse type
Physical -Yes -No Sexual -Yes -No Neglect -Yes -No Emotional -Yes -No
Highest household income
b13 yrs n = 2732 n (%)
≥13 yrs n = 755 n (%)
1494(88) 1238(70)
206(12) 549(30)
87(24) 2645(85)
270(76) 485(16)
666(84) 2066(77)
131(16) 624(23)
42(41) 2690(80)
60(59) 695(21)
p-value
Yes n (%)
No n (%)
498(75) 164(25)
1203(43) 1623(57)
25(4) 637(96)
333(12) 2493(88)
72(11) 590(89)
726(26) 2100(74)
16(3) 645(97)
86(3) 2740(97)
b0.001
Lowest household income p-value
Yes n (%)
No n (%)
471(37) 821(64)
1230(56) 965(44)
189(15) 1103(85)
169(8) 2027(92)
379(29) 913(71)
419(19) 1777(81)
35(3) 1258(97)
68(3) 2128(97)
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
p-value
n.s.
n.s.
yrs: Years, Highest household income = $67,000+, Lowest household income $1–40,999.
with these past admissions, the ones with eventual emotional abuse tended to be older (13 years or older) when compared with children younger than 13 years (59% vs. 41%), p b 0.001, Table 2.
for chronic health conditions were for psychiatric conditions. Compared to those without psychiatric conditions, they were more likely to have sexual (47% vs. 6%) and emotional abuse (10% vs 2%), both p b 0.001, Table 5.
2.2. Socioeconomic status 3. Discussion One-third of patients (35%) were in the lowest median household income quartile ($1–$40,999). The wealthiest median household income quartile ($67,000 +) was found to have higher rates of prior admission (16%), multiple prior admissions (56%, p b 0.001), and prior admission at a different hospital (53%), all p b 0.001, Table 1. For children with previous hospitalizations, those in the highest median household income had higher rates of physical abuse compared to those in lower income (75% vs. 43%, p b 0.001), Table 2. Conversely children in the lowest median household income cohort had higher rates of sexual abuse (15% vs. 8%, p b 0.001) and neglect (11% vs. 3%, p b 0.001) compared to those in higher income households, Table 2. 2.3. Past hospitalizations Patients 13–18 years had the highest rate of prior admissions (18%, p b 0.001) and multiple prior admissions (36%, p b 0.001). The rate of prior admission at a different hospital was highest in those b 1 year (49%) when compared to older patients (23%–30%), p b 0.001. The rate of multiple prior admissions was not significantly different between the abuse types (p = 0.519). Patients with multiple prior admissions more often sustained abuse from a female parental figure (33%) compared to children who sustained abuse from a male parental figure (20%), p b 0.001. Those abused by a male parental figure however were more likely to have a prior admission to a different hospital (57%) when compared to those abused by a female parental figure (26%), p b 0.001, Table 1. Nearly two-thirds (60%, n = 2091) of prior admissions were for chronic health conditions and 12% (n = 399) were for traumatic injuries, Table 3. The most common chronic condition was psychiatric (26%) followed by gastrointestinal (13%), neonatal conditions (10%), failure to thrive/malnutrition (8%) and neurologic conditions (7%). Orthopedic trauma made up the majority (48%) of traumatic conditions followed by head trauma (31%), burns (13%), and poisoning/self-harm (5%). 70% of children with past traumatic injuries (n = 276) would eventually go on to be admitted for physical abuse, while 30% would later sustain other types of abuse (p b 0.001, Table 3). Of these children, 65% had previously sustained orthopedic fractures and 25% sustained closed-head trauma. Children with chronic health conditions were more likely to experience emotional abuse (75%) and sexual abuse (12%), both p b 0.01 (Table 4). A quarter (26%, n = 537) of admissions
This study represents the first nationwide evaluation of pediatric hospital admissions prior to abuse including admissions to a different hospital while exploring past admission diagnoses. The high rate of prior admission (11%) suggests that at least a portion of these prior
Table 3 Characteristics of past-admissions in children admitted for nonaccidental trauma.
Diagnoses at prior admission Chronic Conditions Acute Conditions Trauma Unknown Chronic Conditions groups Neonatal/Perinatal Conditions Psychiatric Hematologic / Malignancies Obstetrics / Gynecology Transplant Gastrointestinal disorders Postoperative conditions Heart conditions Respiratory conditions Neurological conditions Failure to Thrive / Malnutrition Endocrine Other / Miscellaneous Acute Conditions groups Infectious Dehydration/Electrolyte abnormalities Neonatal event Altered mental status Other / Miscellaneous Traumatic Conditions groups Burns Drowning / Near Drowning Poisoning / Self-harm Orthopedic trauma Head trauma Blunt trauma Sexual trauma
n
%
2091 989 399 *
59.9 28.3 11.5 b1
200 537 104 34 19 273 26 28 133 149 157 120 311
9.6 25.7 5.0 1.6 0.9 13.1 1.2 1.3 6.3 7.1 7.5 5.7 14.9
1047 47 48 *8 135
81.5 3.7 3.8 b1 10.5
51 * 20 190 125 * *
12.8 b1 5.0 47.6 31.4 b1.5 b1.5
Cells marked with an asterisk (*) represent actual values censored from publication in accordance with the Healthcare Cost and Utilization Project Data Use Agreement.
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069
4
H.J. Quiroz et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
Table 4 Comparison between prior admission diagnoses with types of eventual abuse. Acute n = 997
Abuse Type
Physical -Yes -No Sexual -Yes -No Neglect -Yes -No Emotional -Yes -No
Trauma n = 399
Yes n (%)
No n (%)
454(27) 543(30)
1247(73) 1244(70)
25(7) 972(31)
333(93) 2158(69)
288(36) 709(26)
510(64) 1981(74)
26(26) 970(29)
76(75) 2415(71)
p-value
Chronic n = 2091
Yes n (%)
No n (%)
276(16) 123(7)
1425(84) 1663(93)
16(4) 384(13)
342(96) 2746(87)
31(4) 368(14)
766(96) 2322(86)
*(b1) 399(12)
102(100) 2986(88)
p-value
Yes n (%)
No n (%)
965(57) 1126(63)
736(43) 661(37)
317(12) 1774(57)
41(88) 1356(43)
486(61) 1604(60)
311(39) 1086(40)
76(75) 2015(59)
26(26) 1371(41)
b0.001
0.02
b0.001
b0.001
b0.001
b0.001
b0.001
b0.001
n.s.
b0.001
n.s.
p-value
0.002
Cells marked with an asterisk (*) represent actual values censored from publication in accordance with the Healthcare Cost and Utilization Project Data Use Agreement.
admissions were because of unrecognized child abuse or for diagnoses that may be correlated with abuse. This potentially suggests that despite having past hospitalizations, either children are not undergoing proper screening for abuse or current screening tools are ineffective for certain subsets. Some groups have indicated a “child-abuse” checklist with further screening by child-abuse experts which could potentially limit the missed cases of child abuse seen by emergency department ED providers [24]. This study presents the novel finding of a 37% rate of prior admissions occurring at a different hospital. This is higher than the previously reported rates of readmission to a different hospital after trauma from the NRD (10%–26%) [22,23]. The rate of prior admission to a different hospital was much higher for physical abuse (48%). Most strikingly, those abused by a male parental figure were two times more likely to have a prior admission to a different hospital than those abused by females. These findings have important implications for prevention through identifying the most common patterns of abuse. This could indicate that the perpetrators are evading detection by presentation to a different hospital, particularly when certain injury patterns or frequency of injury will increase suspicion by healthcare workers. Further novel findings are that maternal perpetrators had a higher rate of multiple prior admissions (along with grandparents) when compared to other perpetrators. Previous studies of mothers of neglected children found that these women tended to believe that they had been caring for their children well and were surprised when their child became ill [25,26]. Unlike paternal physical abuse, there is no pressure to present to a different hospital if the abuse is unacknowledged by the perpetrator. These findings are consistent with previous reports that suggest that physical abuse is more often carried out by fathers while neglect is more often associated with mothers [25,27,28] Thus, the
pattern for negligent mothers appears to be multiple prior admissions at the same hospital, indicating that the “cry for help” was missed. It has been widely reported that both abuse and neglect are more prevalent with a lower socioeconomic status [29–31]. Despite this study's support of prior work regarding the overall rate of child abuse in relation to socioeconomic status, this is the first study to reveal the subtle differences in abuse category with regard to socioeconomic status. Namely, patients with prior admissions and highest socioeconomic status have higher rates of physical abuse than other socioeconomic strata. This could be because of either a higher likelihood to present for hospital admission or higher likelihood of detection and treatment for their injuries. Further studies are required to determine a causal relationship. Our study also aligns with previous reports that neglect is associated with lower socioeconomic status [29,32,33]; however, our data reveal a hidden association between lowest household income quartile with sexual abuse. Sexual abuse is a difficult diagnosis to make; it may not necessarily be the reason for seeking medical care and is often unreported. We suspect that the true incidence is underestimated in the sample. Clinicians should remain vigilant and always consider sexual abuse given these findings. To try to determine potential harbingers of eventual abuse, the indications for past hospitalizations were evaluated. Interestingly, a large portion of these admissions were for chronic conditions (60%) and past traumatic injuries (12%). The most common traumatic injuries identified on past admissions were orthopedic injuries, closed-head trauma, and burns. Additionally, having past traumatic injury was significantly associated with eventual hospitalizations for physical abuse, which indicates that a portion of the children with past abuse-related injuries [12,13,18,35] and subsequent hospitalization for abuse were
Table 5 Subgroup analysis of psychiatric conditions and types of eventual abuse. Psychiatric Diagnosis Yes Abuse Type
Physical -Yes -No Sexual -Yes -No Neglect -Yes -No Emotional -Yes -No
No
p-value b0.001
180(34) 354(66)
428(40) 632(60)
253(47) 281(53)
58(6) 1003(94)
46(9) 488(91)
366(35) 694(66)
51(10) 483(90)
15(1.5) 1045(98.5)
b0.001
b0.001
b0.001
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069
H.J. Quiroz et al. / Journal of Pediatric Surgery xxx (xxxx) xxx
likely cases of missed abuse. This failure of our healthcare system to detect abuse and attempt prevention of further abuse reveals an area for improvement in definitive identification and management of children with abuse. Our study also reveals that one in four children with chronic conditions have past admissions for psychiatric conditions. Child sexual abuse is linked to an increase in depression, posttraumatic stress disorder, other psychiatric disorders, and suicidal ideations [36,37]. According to our data, children with past psychiatric admissions have higher rates of admissions for sexual abuse and emotional abuse. Thus, children presenting with psychiatric conditions represent an important population where careful consideration should be given to current or past child abuse and appropriate steps made for future prevention. Our data also revealed that the rates of readmission for abuse are much higher than other studies in trauma patients [19,34]. Eleven percent of patients treated for child abuse had prior admissions, and an astounding 37% of these children were admitted at a different hospital. The importance of this finding cannot be overstated — at least one in three children is treated for abuse at multiple hospitals. These at-risk patients are difficult to identify in real time owing to the lack of communications between hospitals and no centralized state or national medical record system. The limitations of this study include those of retrospectively collected administrative databases like the NRD such as errors in data sampling, collection measures, usage of the ICD-9 coding scheme, and possible administrative errors during data entry. Additionally, the NRD cannot follow patients across years or if the patients are readmitted across state lines. HCUP estimates that less than 5% of readmissions occur across state lines [20]. Despite these limitations the NRD is the largest database that allows for following readmissions across different hospitals in the US. This is particularly important in victims of child abuse who often present to nonindex hospitals [23,34]. 4. Conclusions Utilization of the NRD provides the unique capability of tracking patients across hospital admissions and reveals the pattern of risk factors that could lead to early identification and prevention of child abuse. More than 10% of patients admitted for child abuse had prior admissions, with some highly likely to have been prior abuse, representing an enormous opportunity for improvement on current identification and preventative measures. A large portion of these past admissions were at a different hospital, indicating the paramount need for a standardized medical record easily transferable between hospitals. This study also uncovered a hidden population of children with past admissions for past traumatic injuries and chronic conditions, especially psychiatric diagnoses that are at significant risk. Improved measures to accurately identify children at risk must be further studied to decrease future childhood abuse and trauma. 5. Discussion Identifying Populations at Risk for Child Abuse: A Nationwide Analysis. Presenter: Hallie Quiroz. Q. Jon Groner from Columbus, Ohio: Great work. My question is, are you able to look at ethnic backgrounds and race because there's other data that shows that white kids are a lot more likely to be missed actually? Hallie Quiroz: Right. So actually as opposed to the kid database where you can look at racial differences, the NRD actually does not breakdown racial components of that. So that would be a weakness of the NRD. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
5
Disclosure The authors have no relevant disclosures.
References [1] Schilling S, Christian CW. Child physical abuse and neglect. Child Adolesc Psychiatr Clin N Am 2014;2:309–19 (ix). [2] Saini SM, Hoffmann CR, Pantelis C. et. al. Systematic review and critical appraisal of child abuse measurement instruments. Psychiatry Res 2019:106–13. [3] Whitaker DJ, Lutzker JR, Shelley GA. Child maltreatment prevention priorities at the Centers for Disease Control and Prevention. Child Maltreat 2005(3):245–59. [4] M C. The Changing Focus of Child Maltreatment Research and Practice Within Psychology. J Soc Issues 2006;4:663–84. [5] Schnitzer PG, Covington TM, Wirtz SJ, et al. Public health surveillance of fatal child maltreatment: analysis of 3 state programs. Am J Public Health 2008(2):296–303. [6] Gilbert R, Kemp A, Thoburn J, et al. Recognising and responding to child maltreatment. Lancet 2009;9658:167–80. [7] Sanada F, Taniyama Y, Muratsu J, et al. Gene-Therapeutic Strategies Targeting Angiogenesis in Peripheral Artery Disease. Medicines (Basel, Switzerland) 2018:2. [8] Al-Damluji MS, Dharmarajan K, Zhang W, et al. Readmissions after carotid artery revascularization in the Medicare population. J Am Coll Cardiol 2015(14):1398–408. [9] Christian CW. Committee on Child A, Neglect AAoP. The evaluation of suspected child physical abuse. Pediatrics 2015;5:e1337–54. [10] Normandin PA. Hidden Child Abuse Identification in the Emergency Department. J Emerg Nurs 2016(2):181–2. [11] Dorfman MVMJ, Feldman KW, Farris R, et al. ExTRA Investigators Oral injuries and occult harm in children evaluated for abuse. Arch Dis Child 2018(8):47–52. [12] Lindberg DM, Beaty B, Juarez-Colunga E, et al. Testing for Abuse in Children With Sentinel Injuries. Pediatrics 2015(5):831–8. [13] Paine CW, Wood JN. Skeletal surveys in young, injured children: A systematic review. Child Abuse Negl 2018:237–49. [14] Ravichandiran N, Schuh S, Bejuk M, et al. Delayed identification of pediatric abuserelated fractures. Pediatrics 2010(1):60–6. [15] Jenny C, Hymel KP, Ritzen A, et al. Analysis of missed cases of abusive head trauma. JAMA 1999(7):621–6. [16] Leventhal JM, Martin KD, Asnes AG. Incidence of fractures attributable to abuse in young hospitalized children: results from analysis of a United States database. Pediatrics 2008(3):599–604. [17] Wood JN, Medina SP, Feudtner C, et al. Local macroeconomic trends and hospital admissions for child abuse, 2000-2009. Pediatrics 2012(2):e358–64. [18] Kaufman E, Rising K, Wiebe DJ, et al. Recurrent violent injury: magnitude, risk factors, and opportunities for intervention from a statewide analysis. Am J Emerg Med 2016(9):1823–30. [19] Parreco J, Rattan R. Machine Learning Models for Prediction of Reinjury After Penetrating Trauma. JAMA Surg 2018(2):184–6. [20] Quality AfHRa. Introduction to the HCUP Nationwide Readmissions. Database 2017. [21] Allison Russo C HM, M. SW, Owens PL. Hospital Stays Related to Child Maltreatment, https://www.hcup-us.ahrq.gov/reports/statbriefs/sb49.pdf; 2005 2019]. [22] Sittig JS, Uiterwaal CSPM, Moons KGM, et al. Value of systematic detection of physical child abuse at emergency rooms: a cross-sectional diagnostic accuracy study. BMJ Open 2016(3):e010788. [23] Parreco JBJ, Cortolillo N, Namias N, et al. Risk factors and costs associated with nationwide nonelective readmission after trauma. Eastern Association for the Surgery of Trauma Annual Scientific Assembly; 2017 Hollywood, FL. [24] Buicko JL, Parreco J, Willobee BA, et al. Risk factors for nonelective 30-day readmission in pediatric assault victims. J Pediatr Surg 2017(10):1628–32. [25] Golden MH, Samuels MP, Southall DP. How to distinguish between neglect and deprivational abuse. Arch Dis Child 2003;2:105–7. [26] Pignatelli AM, Wampers M, Loriedo C, et al. Childhood neglect in eating disorders: A systematic review and meta-analysis. J Trauma Dissociation 2017(1):100–15. [27] Lee SJ, Bellamy JL, Guterman NB. Fathers, physical child abuse, and neglect: advancing the knowledge base. Child Maltreat 2009(3):227–31. [28] Konishi A, Yoshimura B. Child abuse and neglect by mothers hospitalized for mental disorders. Arch Womens Ment Health 2015(6):833–4. [29] Eckenrode J, Smith EG, McCarthy ME, et al. Income inequality and child maltreatment in the United States. Pediatrics 2014(3):454–61. [30] Belsky J. Etiology of child maltreatment: a developmental-ecological analysis. Psychol Bull 1993;3:413–34. [31] Pelton LH. Child abuse and neglect: the myth of classlessness. Am J Orthop 1978(4): 608–17. [32] Eckenrode J, Smith EG, McCarthy ME, et al. Income Inequality and Child Maltreatment in the United States. Pediatrics 2014(3):454. [33] Kotch JB, Browne DC, Ringwalt CL, et al. Risk of child abuse or neglect in a cohort of low-income children. Child Abuse Negl 1995(9):1115–30. [34] Block RW. No Surprise: The Rate of Fatal Child Abuse and Neglect Fatalities Is Related to Poverty. Pediatrics 2017(5). [35] Pawlik MC, Kemp A, Maguire S. et. al. Children with burns referred for child abuse evaluation: Burn characteristics and co-existent injuries. Child Abuse Negl 2016:52–61. [36] O'Brien BS, Sher L. Child sexual abuse and the pathophysiology of suicide in adolescents and adults. International journal of adolescent medicine and health 2013(3):201–5. [37] Molnar BE, Berkman LF, Buka SL. Psychopathology, childhood sexual abuse and other childhood adversities: relative links to subsequent suicidal behaviour in the US. Psychol Med 2001(6):965–77.
Please cite this article as: H.J. Quiroz, J. Parreco, L. Easwaran, et al., Identifying Populations at Risk for Child Abuse: A Nationwide Analysis, Journal of Pediatric Surgery, https://doi.org/10.1016/j.jpedsurg.2019.09.069