This article originally appeared in the School of Social Work 2024 Fall Impact Report.
When it comes to spending taxpayer money on work that makes a difference, data talks.
Since 1997, University of North Carolina at Chapel Hill School of Social Work Research Professor Dean Duncan has been gathering and analyzing data for North Carolina’s Division of Social Services. He has linked big data sets to identify trends, locate outliers that indicate where the division might want to devote more resources, and to document the efficacy of an intervention.
“We help the state with analyses they don’t have time to do,” Duncan said. “It’s a long-term partnership.”
Duncan joined the School’s faculty in 1997 to lead a project funded by NCDSS, studying the impact of Work First — North Carolina’s version of the federal Temporary Assistance for Needy Families program. The state implemented TANF under a federal waiver. Duncan and his team collected and analyzed data to indicate how well Work First was operating, especially in boosting employment.
By building longitudinal files based on extracts from NCDSS files and linking them to employment data, Duncan tracked how long people stayed on Work First, whether they found work and how much they earned, as well as family size, education and race. The following year, he folded in data from the Food and Nutrition Services (food stamp) program.
Duncan and his team developed a website to share data, updated monthly, with social services departments in other counties, enabling them to observe and improve the effectiveness of their programs.
In the early 2000s, Duncan’s team partnered with Lynn Usher, now distinguished professor emeritus at the School of Social Work, to add information on county child welfare programs to the website. They began by posting an experiences report Usher and his team developed and added reports of abuse and neglect, foster care entries, and federal and state performance indicators over time. North Carolina is one of the few states with a public-facing website to make such information available.
“We use our data for all sorts of analysis,” Duncan said. “We track disproportionality in reports of maltreatment and entries into foster care as well as what happens to kids aging out of foster care.”
In 2012, Duncan began working with NCDSS and Duke’s Center for Child & Family Health on Project Broadcast and received a $3 million grant from the U.S. Children’s Bureau to move to a trauma-informed child welfare system. As part of that effort, Duncan and his team began tracking the behavioral health care diagnoses of children in foster care and the number of prescribed psychotropic medications.
Heather McAllister, section chief for Child and Family Services at the N.C. Department of Health and Human Services, said North Carolina has an ongoing interest in foster care children and trauma-informed care. She collaborated with Duncan on Project Broadcast.
By putting Duncan’s data analysis to use in Project Broadcast, McAllister’s division implemented a trauma screening tool and piloted it in several North Carolina counties. Project Broadcast collected information about the types of trauma children experienced prior to foster care placement and their ongoing needs for intervention.
National research conducted more than a decade ago showed that children in foster care received psychotropic medications at a higher rate than other populations. The division wanted to improve monitoring of psychotropic medications prescribed for children in foster care to ensure that the drugs weren’t overused to manage behavior and that there was adequate follow-up treatment. The data Duncan’s team collected on trauma enabled McAllister’s section to determine what services needed to be put in place.
“Research shows if you accurately identify trauma in children, and if you apply appropriate interventions,” McAllister said, “you see more accurate diagnoses and potential reduction in psychotropic medications and more referrals to necessary therapies.”
Duncan’s data, some of which comes from Medicaid claims, revealed red flags that McAllister’s section investigated more closely, outliers such as disparities in diagnoses and prescriptions across ages, populations or regions.
“(Duncan’s research) gives us a better picture of what’s happening statewide, and we can develop better interventions,” McAllister said.
Though Project Broadcast has ended, McAllister can take the process Duncan used and re-implement his data analysis work. That has revealed insights into how some children needing intervention were overlooked.
Because Duncan has been working with child welfare and cross-systems data for many years, McAllister said, “he understands and has unique insight into what is important to us, like reduction in placement changes. He can connect the dots.”
Duncan said that linking data sets is critical to his research success.
“Data resides in different systems, designed by different people and in different structures,” he said. “There’s a richness when you link data. We can follow people from one system to another. You get a better picture of what’s going on by being able to look across multiple systems.”
Duncan’s research identifies issues upstream before they become significant problems. Understanding the nature of a problem leads to more effective interventions. And one of the best ways to secure funding for new initiatives, McAllister said, is to have data that shows the impact.
“(Duncan’s) work helps inform us about future programming,” she said. “He brought us tools we need to identify and show impact. Having that information has propelled us forward.”
