Emergency. 2017; 5 (1): e8 OR I G I N A L RE S E A RC H Potential Child Abuse Screening in Emergency Depart- ment; a Diagnostic Accuracy Study Hossein Dinpanah1, Abazar Akbarzadeh Pasha2∗, Mojtaba Sanji3 1. Emergency Department, Shahid Beheshti Hospital, Babol University of Medical Sciences, Babol, Iran. 2. Urology Department, Shahid Beheshti Hospital, Babol University of Medical Sciences, Babol, Iran. 3. Emergency Department, Ali ebne Abitaleb Hospital, Rafsanjan University of Medical Sciences, Rafsanjan, Iran. Received: March 2016; Accepted: May 2016; Published online: 8 January 2017 Abstract: Introduction: Designing a tool that can differentiate those at risk of child abuse with great diagnostic accuracy is of great interest. The present study was designed to evaluate the diagnostic accuracy of Escape instrument in triage of at risk cases of child abuse presenting to emergency department (ED). Methods: The present di- agnostic accuracy study performed on 6120 of the children under 16 years old presented to ED during 3 years, using convenience sampling. Confirmation by the child abuse team (pediatrician, a social worker, and a forensic physician) was considered as the gold standard. Screening performance characteristics of Escape were calcu- lated using STATA 21. Results: 6120 children with the mean age of 2.19 ± 1.12 years were screened (52.7% girls). 137 children were suspected victims of child abuse. Based on child abuse team opinion, 35 (0.5%) children were confirmed victims of child abuse. Sensitivity, specificity, positive and negative likelihood ratio and positive and negative predictive values of this test with 95% CI were 100 (87.6 – 100), 98.3 (97.9 – 98.6), 25.5 (18.6 – 33.8), 100 (99.9 – 100), 0.34 (0.25 – 0.46), and 0 (0 – NAN), respectively. Area under the ROC curve was 99.2 (98.9 – 99.4). Conclusion: It seems that Escape is a suitable screening instrument for detection of at risk cases of child abuse presenting to ED. Based on the results of the present study, the accuracy of this screening tool is 99.2%, which is in the excellent range. Keywords: Child abuse; diagnosis; emergency service, hospital; risk assessment; decision support techniques © Copyright (2017) Shahid Beheshti University of Medical Sciences Cite this article as: Dinpanah H, Akbarzadeh Pasha A, Mojtaba Sanji. Potential Child Abuse Screening in Emergency Department; a Diag- nostic Accuracy Study. Emergency. 2017; 5 (1): e8. 1. Introduction Non-accidental physical, mental, emotional or sexual abuse, or neglect of children under 18 years of age, which endangers the child’s health, comfort, and education, is defined as child abuse (1). Regardless of the culture and beliefs of a society, mistreating children may be a major health problem that requires attention from the governments and health care systems due to its wide range of long term effects. It may seem like a personal problem at first sight, however consid- ering its probable side effects such as depression, borderline personality disorder, multiple personality disorder, atten- tion deficit disorder, drug and alcohol abuse, prostitution, ∗Corresponding Author: Abazar Akbarzadeh Pasha; Urology Depart- ment, Shahid Beheshti Hospital, Keshvari Square, Babol, Iran. Tel/Fax: 00989111112231; Email: aapash812@yahoo.com running away from home, antisocial and criminal behavior, and sexual crimes, it is considered a social and multidi- mensional phenomenon (2). Child abuse was first assessed as a problem that may affect the present and future life of a person in 1962 with publication of an article titled “the beaten child syndrome”, which became a stepping-stone for future studies (3). According to statistics, during 1976 to 1983 more than 50000 children were killed by their parents as a result of child abuse in the United States, and more than 25 million children were subject to abuse and anger (4). In 1995, more than 3 million children were referred to child support centers in the United States due to abuse and neglect. Death or sickness of a family member, financial problems and dissatisfaction with marriage, have been introduced as child abuse risk factors (5). In Iran, most cases of child abuse belong to physical abuse of boys and factors such as parents low educational level, low economic status, populated family, and mental and physical illnesses are This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com H. Dinpanah et al. 2 identified risk factors (6, 7). Potential child abuse screening in those presenting to emergency department (ED) can help identify effective factors in abuse incidence and move toward reducing its prevalence by proper intervention. Designing a tool that can differentiate those at risk of child abuse with great diagnostic accuracy is of great interest for emergency physicians. Although in recent years, child abuse screening tools have significantly helped emergency physicians, the accuracy of these tools is still a matter of question (8-11). Therefore, the present study was designed to evaluate the diagnostic accuracy of Escape instrument in triage of at risk cases of child abuse presenting to ED. 2. Methods 2.1. Study design and setting The present study is a prospective diagnostic accuracy study performed on children presented to ED of Shahid Beheshti and Amir Kola Hospitals, Babol, Mazandaran, Iran, during 2011 to 2014. The aim of this study was evaluating the ac- curacy of Escape tool in screening children at risk of child abuse. The study was approved by the ethics committee of Babol University of Medical Sciences. The researchers ad- hered to the principles of Helsinki Declaration and keep- ing patient information confidential at all stages during the study. The patients or their relatives were assured that their personal data will be confidential and only used for the pur- pose of the study and written informed consent was obtained from them. 2.2. Participants 6120 of the children (under 16 years old) presented to ED dur- ing the study period were triaged and enrolled using conve- nience sampling. Inclusion criteria were consent for partici- pation, cooperation in filling out the questionnaire, and sta- ble clinical and hemodynamic status. Cases of suicide injury, poisoning, and those who had introduced their case as child abuse or were injured by their peers were excluded. 2.3. Data gathering On admission to ED, demographic data of all children (age, sex, place of living) as well as their hydration status were recorded and Escape questionnaire for potential child abuse screening (appendix 1) was filled for them by asking ques- tions from the child or the guardians (11). Triage was done by trained nurses. In cases of one or more abnormal answer to the questions, the screening result was considered posi- tive. After admission to ED, standard treatment (based on the reason for admission) was initiated and a trained emergency medicine specialist, blind to results of screening, accurately examined the child and recorded the history regarding child abuse. In cases that were diagnosed as child abuse, the child was re-examined by the hospital’s child abuse team includ- ing a pediatrician, a social worker, and a forensic physician to confirm diagnosis. Emergency physician and all members of child abuse team were blind to the results of screening. Confirmation by the mentioned team was considered as the gold standard for identifying the patient as a victim of child abuse. 2 emergency medicine specialists passed three 2-hour educational courses with the hospital child abuse team and were responsible for initial evaluation of the patients on ad- mission. In-charge triage nurses in this study also underwent training for a few sessions to learn about filling the question- naire. The validity and reliability of the questionnaire were confirmed in a previous study (11). 2.4. Statistical analysis Minimum sample size required for the present study was calculated to be 2696 cases, considering 2.3% prevalence of child abuse (11), 80% sensitivity, 95% confidence interval (CI), desired precision (d = 0.1). Data were analyzed using STATA 11.0. Quantitative variables were reported as mean and standard deviation (SD) and qualitative ones were shown as frequency and percentage. To calculate the accuracy of the tool, sensitivity, specificity, positive and negative likelihood ratios, and positive and negative predictive values and area under the receiver operating characteristic (ROC) curve were calculated with 95% CI. 3. Results: 6120 children with the mean age of 2.19 ± 1.12 years were screened regarding potential child abuse (52% girls). Figure 1 shows the frequency of patients in each age group. 4376 (71.5%) of the participants resided in cities. Table 1 shows the frequency of positive answer to each of the 6 questions as well as their screening performance characteristics. Based on the results obtained from the screening questionnaire, 137 children were suspected victims of child abuse, 120 (2%) of which had 1 positive answer, 4 (0.1%) had 2 positive answers, 1 (0.01%) had 3 positive answers, and 1 (0.01%) had 4 pos- itive answers. Finally, based on child abuse team opinion (the gold standard), 35 (0.5%) children were victims of child abuse. Sensitivity, specificity, positive and negative likeli- hood ratio and positive and negative predictive values of this test with 95% CI were 100 (87.6 – 100), 98.3 (97.9 – 98.6), 25.5 (18.6 – 33.8), 100 (99.9 – 100), 0.34 (0.25 – 0.46), and 0 (0 – NAN), respectively. Figure 2 shows the area under the ROC curve for the studied instrument. Area under the ROC curve was 99.2 (98.9 – 99.4). This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com 3 Emergency. 2017; 5 (1): e8 Appendix 1: Escape questionnaire for screening child abuse 1. Is the history consistent? Yes No 2. Was seeking medical help unnecessarily delayed? Yes No 3. Does the onset of the injury fit with the development level of the child? Yes No 4. Is the behavior of the child, his or her care givers and their interaction appropriate? Yes No 5. Are findings of the head-to- toe examination in accordance with the history? Yes No 6. Are there signals that make you doubt the safety of the child or other family members? Yes No *if Yes describe the signals in the box Other comments below. Other comments Table 1: Screening performance characteristics of the child abuse questionnaire with 95% confidence interval in prediction of at risk children presented to emergency department Question Number (%) Sensitivity Specificity PPV NPV 1 14 (0.2) 11.4 (0.03 – 0.27) 99.8 (99.7 – 99.9) 28.5 (0.09 – 0.57) 99.5 (99.3 – 99.6) 2 26 (0.4) 20 (9 – 40) 99.7 (99.5 – 99.8) 26.9 (12.4 – 48.0) 99.5 (99.3 – 99.6) 3 38 (0.6) 14.2 (5.3 – 31.0) 99.4 (99.2 – 99.6) 13.1 (4.9 – 28.8) 99.5 (99.2 – 99.6) 4 27 (0.4) 17.1 (7.1 – 34.2) 99.6 (99.4 – 99.7) 22.2 (9.3 – 42.7) 99.5 (99.3 – 99.6) 5 12 (0.2) 11.4 (3.7 – 27.6) 99.8 (99.7 – 99.9) 33.3 (11.2 – 64.5) 99.4 (99.2 – 99.6) 6 18 (0.3) 14.2 (5.3 – 31.0) 99.7 (99.6 – 99.8) 27.7 (10.7 – 53.5) 99.5 (99.2 – 99.6) PPV: positive predictive value; NPV: negative predictive value. Figure 1: Age distribution of the studied children. 4. Discussion: Based on the findings of this study, Escape screening instru- ment has high sensitivity and specificity in identifying poten- tial child abuse cases presented to ED. Area under the ROC curve of 99.2 indicates the high accuracy of the test in this regard. As mentioned before, child abuse is a mental and health problem in every society, which is directly related to mental and physical health of the next generation. Based on the statistics reported by world health organization (WHO) about 3 million children are maltreated around the world each year and 31000 cases of murder have been reported in children under 15 years old in 2002 alone (12). Since a large Figure 2: Area under the ROC curve of the child abuse question- naire. number of children with various injuries are presented to ED daily, timely identification and evaluation of those suffering from or at risk of child abuse plays a significant role in pre- venting further damages. Child abuse rate reported in var- ious studies carried out in EDs has ranged from 2% to 10% (13-19). Using a standard tool that can accurately determine true cases is a challenge for physicians in supporting chil- dren’s rights. Protocols designed for this purpose should be able to guide the physicians toward a comprehensive answer with few questions. In 2012, Louwers et al. used Escape screening instrument in 3 health centers for the first time. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com H. Dinpanah et al. 4 In that study, Escape was used to evaluate potential risk of child abuse in children (aged 0 to 18 years) presented to the EDs. Using this instrument, screening rate increased from 20% in February 2008 to 67% in December 2009. Detection rate in the screened children was 5 times higher than those not screened. Therefore, it seems that Escape tool is effective in increasing detection of potential child abuse (20). Pless et al. studied the Accident- Suspected child abuse and neglect (A-SCAN) method, a checklist with 10 questions for assessing the risk of child abuse. The results of this checklist correlated with physical examination results reported by the physician. No significant increase in detection of child abuse was seen after introduction of this method. This could mean that ED staff were already doing well and the method used was not efficient (19). In another study to assess child abuse by con- sulting the child protection register, a flowchart with 4 ques- tions was included in the patient’s file. Results showed that inclusion of a flowchart improved awareness, attention and documentation of suspected abuse cases (15). In Bleeker et al. study, a 9-question checklist was used in ED for collecting information from children suspected to be child abuse cases, which using this tool the number of detected cases increased (21). Hosseinkhani and colleagues determined the status of child abuse in the Iranian population and evaluated the va- lidity and reliability of a new questionnaire. They concluded that, their questionnaire is a new tool with acceptable valid- ity and reliability and can be applied in child abuse studies in Iran (22). Since currently there is no accepted standard for screening children at risk of child abuse in ED, researchers are trying to design and develop new decision rules or vali- date the existing tools. Therefore, the present study was de- signed with the same aim. It seems that carrying out simi- lar studies in other parts of the country with various cultural and economic statuses can provide more acceptable results for reaching a decision regarding the accuracy of this tool. 5. Conclusion: It seems that Escape is a suitable screening instrument for detection of at risk cases of child abuse presenting to ED. Based on the results of the present study, the accuracy of this screening tool is 99.2%, which is in the excellent range. 6. Appendix 6.1. Acknowledgements All authors would like to thank all the ED staff of Shahid Be- heshti and Amir Kola Hospitals, Babol, Mazandaran, Iran. 6.2. Author contribution All authors passed four criteria for authorship contribution based on recommendations of the International Committee of Medical Journal Editors. 6.3. Conflict of interest None 6.4. Funding None References 1. http://www.who.int/topics/child_abuse/en/. 2007. 2. Hobbs CJ, Hanks HG, Wynne JM. Child abuse and ne- glect: a clinician’s handbook: Elsevier Health Sciences; 1999. 3. Kemp A. Abuse in the family: An introduction: Wadsworth Publishing Company; 1998. 4. Ludwig S, Kornberg AE. Child abuse: a medical reference: Churchill Livingstone New York; 1992. 5. McMillan JA, Feigin RD, DeAngelis C, Jones MD. Oski’s pediatrics: principles & practice: Lippincott Williams & Wilkins; 2006. 6. Sayyari AA, Bagheri Yazdi SA, Jalili B, Khoshabi K, Shah- mohammadi D, Imanzadeh F, et al. Physical child- abuse in Tehran, Iran. Journal of rehabilitation (UWRS). 2002;6(6-7):7-13. 7. Farhoudian A, Izadian ES, Goodarzi RR, Sharifi V, Mo- hammadi M-R, Nejatisafa A-A, et al. Iran’s Contribution to Child and Adolescent Mental HealthResearch (1973– 2002): A Scientometric Analysis. Iranian Journal of Psy- chiatry. 2006;1(3):93-7. 8. Louwers EC, Affourtit MJ, Moll HA, de Koning HJ, Ko- rfage IJ. Screening for child abuse at emergency depart- ments: a systematic review. Archives of disease in child- hood. 2010;95(3):214-8. 9. Woodman J, Pitt M, Wentz R, Taylor B, Hodes D, Gilbert R. Performance of screening tests for child physical abuse in accident and emergency departments. 2008. 10. Raffle AE, Gray JAM. Screening: evidence and practice: Oxford University Press; 2007. 11. Louwers EC, Korfage IJ, Affourtit MJ, Ruige M, van den Elzen AP, de Koning HJ, et al. Accuracy of a screening in- strument to identify potential child abuse in emergency departments. Child abuse & neglect. 2014;38(7):1275-81. 12. Butchart A, Harvey AP, Mian M, Furniss T. Preventing child maltreatment: a guide to taking action and gener- ating evidence. 2006. 13. Hampton RL, Newberger EH. Child abuse incidence and reporting by hospitals: significance of severity, class, and race. American Journal of Public Health. 1985;75(1):56- 60. 14. Hussey JM, Chang JJ, Kotch JB. Child maltreatment in the United States: prevalence, risk factors, and adolescent This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com 5 Emergency. 2017; 5 (1): e8 health consequences. Pediatrics. 2006;118(3):933-42. 15. Benger JR, Pearce AV. Quality improvement report: Sim- ple intervention to improve detection of child abuse in emergency departments. BMJ: British Medical Journal. 2002:780-2. 16. Chang DC, Knight V, Ziegfeld S, Haider A, Warfield D, Paidas C. The tip of the iceberg for child abuse: the critical roles of the pediatric trauma service and its registry. Journal of Trauma and Acute Care Surgery. 2004;57(6):1189-98. 17. Holter JC, Friedman SB. Child abuse: early case finding in the emergency department. Pediatrics. 1968;42(1):128- 38. 18. Kempe CH, Silverman FN, Steele BF, Droegemueller W, Silver HK. The battered-child syndrome. Jama. 1984;251(24):3288-94. 19. Pless IB, Sibald A, Smith MA, Russell M. A reappraisal of the frequency of child abuse seen in pediatric emergency rooms. Child abuse & neglect. 1987;11(2):193-200. 20. Louwers EC, Korfage IJ, Affourtit MJ, Scheewe DJ, Van De Merwe MH, Vooijs-Moulaert A-FS, et al. Effects of sys- tematic screening and detection of child abuse in emer- gency departments. Pediatrics. 2012;130(3):457-64. 21. Bleeker G, Vet NJ, Haumann TJ, van Wijk IJ, Gemke RJ. [Increase in the number of reported cases of child abuse following adoption of a structured approach in the VU Medical Centre, Amsterdam, in the period 2001-2004]. Nederlands tijdschrift voor geneeskunde. 2005;149(29):1620-4. 22. Hosseinkhani Z, Nedjat S, Majdzadeh R, Mahram M, Aflatooni A. Design of the child abuse Questionnaire in Iran. Journal of School of Public Health and Institute of Public Health Research. 2014;11(3):29-38. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com Introduction Methods Results: Discussion: Conclusion: Appendix References