What We Can Learn From Failure: An EHR-Based Child Protection Alert System.

Area of Science:

• Pediatrics
• Public Health
• Health Informatics

Background:

• Child maltreatment poses significant risks to child well-being.
• Effective identification systems are crucial for timely intervention and protection.
• Existing clinical decision support tools for child protection require evaluation.

Purpose of the Study:

• To evaluate the diagnostic accuracy and efficacy of a novel Child Protection Alert System (CPAS).
• To assess the system's ability to identify children with confirmed or suspected maltreatment using triggering diagnoses.
• To identify barriers in the design and implementation of CPAS for protecting at-risk children.

Main Methods:

• Retrospective review of 666 electronic health records (EHRs) of patients evaluated by a child protection team (2009-2014).
• Analysis of EHRs for the presence of pop-up alerts, text-based visual alerts, and qualifying diagnoses indicative of child maltreatment.
• Assessment of the CPAS's diagnostic accuracy, including sensitivity and specificity for child maltreatment identification.

Main Results:

• Only 21.7% of children with confirmed/suspected maltreatment had a qualifying longitudinal diagnosis listed.
• The pop-up alert demonstrated low sensitivity (13.9%) but 100% specificity.
• The text-based visual alert was present in 44 out of 45 relevant cases, indicating good visibility once triggered.

Conclusions:

• The CPAS offers a potential method for supporting clinical decision-making in identifying children at risk of abuse.
• Significant barriers exist in the effective design and implementation of CPAS, particularly concerning alert sensitivity.
• Further research and system refinement are necessary to enhance CPAS efficacy in protecting vulnerable children.