Child Abuse Clinical Decision Support to Improve Detection, Evaluation, and Reporting


Almost 1,600 children die annually due to maltreatment, making it one of the leading causes of morbidity and mortality in children. A significant proportion of children who are victims of physical abuse have previously been evaluated by physicians, but signs of abuse were not detected. Despite evidence-based recommendations for screening children under 2 years of age with suspected abuse, physicians fail to consistently do so. This case study’s objective was to design, develop, and implement a child abuse-clinical decision support system (CA-CDSS) into the electronic health record (EHR) within the emergency department (ED). 

Child abuse concept. Teddy bear covering eyes in an empty room, front view, copy space


The CA-CDSS was originally developed, validated, and implemented within EDs at University of Pittsburgh. A grant proposal was funded by the Patient-Centered Outcomes Research Institute (PCORI) to further disseminate and implement the CA-CDSS in two different health care systems (University of Wisconsin and Northwell Health) with 2 different EHRs. After meetings with key stakeholders, a prototype was developed for the CA-CDSS. The CA-CDSS consisted of the following components: a trigger system (including high risk chief complaint triggers for suspected abuse and a 5 question child abuse screen (CAS) – figure shown), ED status board alert, provider alerts within clinical documentation, and injury-specific order sets. Usability testing was conducted on the prototype and revisions were made. Ultimately, the revised CA-CDSS was implemented into the EHRs and data was collected from June 2019 – April 2020.  


During usability testing, the high risk of missing a potential case of abuse was a facilitator for nurses, advanced practice providers, and physicians to utilize the CA-CDSS. Facilitators for nurse completion of the CAS included having clear definitions of each question and ease of use of the form. For advanced practice providers and physicians, it was important to not trigger the alert too early (i.e. alert upon clicking name of patient) as patient evaluation was a key step in detection for abuse. This testing ensured optimized use and clinical workflow integration of the CA-CDSS. Upon implementation of the CA-CDSS at 4 EDs within Northwell Health, the CAS was completed by nurses during 72% of visits for children less than the age of 13. When abuse was suspected, providers were fully compliant with injury specific order sets, based on American Academy of Pediatrics, 96% (22/23) of the time. 


 This multi-component CA-CDSS was adopted at a high rate by nurses in the ED to aid in identification of children who should be further evaluated by ED providers for possible child maltreatment, which may have otherwise been missed in the setting of the ED. The CA-CDSS was successfully implemented within a short time frame (6 months) and advanced care providers and physicians were fully compliant with order sets in many of the cases. This implementation project was able to adapt an evidenced based model of evaluating for child abuse and disseminate into additional EHRs. The case study proved to be an excellent example of collaborative success between multiple academic institutes and within each institute the importance of key stakeholders (i.e. child abuse pediatricians, ED providers and nurses, information technologists, and implementation science experts) in implementing CDSS.