Black and Latinx Primary Caregiver Considerations for Developing and Implementing a Machine Learning-Based Model for...

Area of Science:

• Public Health
• Health Informatics
• Social Science

Background:

• Child abuse and neglect is a significant epidemic, with potential links to racial and social class stereotypes.
• Electronic health records (EHRs) offer opportunities to address child abuse and neglect.
• Involving marginalized communities is crucial for developing unbiased machine learning (ML) models using EHR data.

Purpose of the Study:

• To gather viewpoints from Black and Latinx primary caregivers on child abuse and neglect.
• To inform the design of ML-based models for detecting child abuse and neglect in emergency departments (EDs).
• To address racial bias and guide future interventions.

Main Methods:

• A qualitative study was conducted.
• In-depth interviews were performed with 20 Black and Latinx primary caregivers.
• Participants' children received care at a pediatric tertiary-care ED.

Main Results:

• Caregivers' perspectives on defining child abuse and neglect.
• Caregivers' experiences with healthcare providers and medical documentation.
• Caregivers' perceptions of child protective services.

Conclusions:

• Primary caregiver insights are essential for developing effective ML models for child abuse and neglect detection in EDs.
• Accurate definitions of child abuse and neglect from a caregiver's viewpoint are critical.
• Miscommunication and potential harm from ML models necessitate careful consideration and further research.