Skip to main content

Big data, big help in child maltreatment

Rebecca Rebbe examines issues most of us prefer not to acknowledge. An assistant professor in UNC’s School of Social Work, Rebbe uses big data to gain insights into which children who appear before medical practitioners with symptoms that point to potential child abuse or neglect are referred to child protective services, and which ones aren’t. When and how do communities and systems intervene? Does income factor in? What about race or ethnicity?

Thousands and thousands of data points, each representing a child brought to an emergency room or admitted to a hospital with injuries consistent with maltreatment.

“I don’t get asked questions about my work much at dinner parties,” Rebbe said.

Currently, Rebbe is studying how child maltreatment and reporting practices may have changed with the onset of COVID, funded by a federal NICHD R21 exploratory and developmental research grant. The results were published in June 2023 in The Journal of Pediatrics under the headline “The Impact of COVID-19 on Infant Maltreatment Emergency Department and Inpatient Medical Encounters.” Her research shows that incidents of child maltreatment didn’t increase initially but did during the second year of COVID. The results suggest the stressors of the pandemic were not limited to the initial lockdown period. Rebbe is continuing to research and analyze reporting practices.

As an undergraduate at the University of Portland in Oregon, child abuse wasn’t on her radar. Neither was the field of social work until she enrolled in a service-learning class her freshman year and the social work professor who led it suggested she consider the field.

“I was intrigued by the idea of understanding people and their environment simultaneously,” she said, “and then with the very important piece of doing something about it.”

Social justice, being a problem solver, trying to tackle some of society’s challenges, “that’s what drew me in,” she said.

Rebbe did a practicum with Oregon’s child welfare agency, and after graduation in 2004, saw some of the subtleties and range of family situations and their impact on youth during a year as an AmeriCorps volunteer in a high school. That led her to complete an Ed.M. in risk and prevention at Harvard that mixed a focus on children with policy and practice. She followed up with an MSW from Boston College in 2007. She worked as an in-home crisis intervention clinician in Boston for a year, then returned to the Pacific Northwest, this time in Washington.

In Seattle, Rebbe worked first as an ongoing caseworker with the Washington state child welfare agency, then as a child protective services investigator. Promoted to supervisor, she oversaw a unit of legally free adolescents, teens whose parents’ parental rights had been terminated.

“They were kids with significant trauma histories,” she said.

Rebbe stayed at the Washington state agency for six years, longer than she’d initially anticipated. But the cases were complex, and she was interested in how the system worked and for whom. She came across a wide variety of situations that led to more questions. The questions built until she realized that she didn’t have the skills to answer them. A doctorate would enable her to address some of them.

“I was intrigued by some of the cases I was seeing: Were they outliers or typical?” she said. She wanted to learn what was going on in the larger system and the broader population level. “I realized I needed advanced research skills to find answers regarding how to make things better and understand how and when to intervene.”

When to intervene is often a gray area. A case in Dallas in April made national news when child protective services took a newborn from her parents after a home birth because the parents opted to treat the baby’s jaundice at home rather than hospitalize her. The doctor reported the parents to CPS, who incorrectly listed the mother’s name as someone who had a criminal record and wrote “alleged father” before the father’s name. CPS put the newborn in foster care. It took weeks before the parents, both Black, were able to regain custody.

Social workers, teachers and health care personnel, among others, are mandated to report incidences of suspected child maltreatment. But, like the Dallas case, it’s a judgment call on what constitutes harm. The obvious cases are the exceptions. Other situations require examination of myriad circumstances and pieces of information to figure out when and how to intercede.

CPS operates similar to a fire department, Rebbe said. Its personnel rely on the community to alert them that something needs their attention; then they respond.

“But we don’t know much about when those judgment calls are made, who makes them and how,” she said. “We don’t know much about which kids have been under consideration for being called in.”

For instance, if a parent with a substance use problem leaves a child unsupervised, is that maltreatment? That’s a gray area. What time of day is it? How old is the child? How long is too long? Rebbe’s research offers insight.

With big data sets, she can, for instance, see all of the children who presented at a hospital, their diagnoses, and their race and ethnicity. She can check another set to see which visit was paid for by public insurance, an indication of poverty, then match all that up with the data set on notifications to CPS. From that, she can glean who by various demographics is more likely to be reported to CPS, or even identify risk factors at birth.

“Population-based data enables us to have large scale understanding of what’s happening across the systems,” Rebbe said.

Such understanding garners some predictive knowledge and where to target resources. Her research might help early intervention programs identify families that need prevention programming or help other researchers develop interventions. “Policymakers also need to understand how our systems are working,” she said. “Then we can identify changes that need to be made to reach families that need help.”

Big data is a new way to answer questions, said Brett Drake, a professor of data science for the social good in practice in the Brown School of Social Work at Washington University in St. Louis. It’s agile and relevant and streamlines the research process. “It’s like having a new pair of eyes on the question,” he said.

An expert in child maltreatment, Drake is well familiar with Rebbe’s work. He noted Rebbe is part of a new wave of data scientists who were brought up with computers.

“She’s an instinctive scientist,” he said. “She thinks logically in a structured way and is creative in understanding what she sees. She’s becoming one of the dominant scientists in this field.”

One of her colleagues at UNC, Ankur Srivastava, an assistant professor in the School of Social Work and a core faculty member at Global Social Development Innovations, said Rebbe’s work with big data sets her apart from her peers.

“Very few people in social work are using big data sets the way she is,” he said. “She’s aware of the political climate, and she thinks about the impact her work can have in that context, how her research can be interpreted and talked through. She uses the data to ask good questions.”

The issues Rebbe tackles are complex; there are no quick and easy answers. Her work isn’t discouraging, she said, but society’s laxity is.

“As a society, we don’t treat or support kids as well as we should; nor do we support parents and recognize how challenging it is to raise kids.”

Injury deaths are the most common form of death for children in the United States. Society doesn’t want to talk about it, and politicians don’t want to deal with it, she said. Through her research, Rebbe aims to give more clarity to identifying maltreatment and when and how to intervene.

“Getting involved with families we don’t need to be involved with can be detrimental and have bad impacts as well,” she said. “Figuring out when and how to intervene is important. Getting it wrong in either direction can hurt.”

by Nancy E. Oates