Policy numbers

By Alex Piazza
apiazza@umich.edu

Andrew Moore’s two computer screens are blanketed with numbers and plots.

The University of Michigan data analyst is looking for patterns on how youth move through the foster care system.

“How long did they stay in foster care?”

“Did they have any run-ins with the law?”

With about 400,000 U.S. children in foster care, state agencies are looking for answers to these pressing questions.

Set aside public opinion, economic conditions and political discord for a moment. More and more state agencies are turning to big data to illuminate social problems and propel solutions for children and families.

Thus, the importance of Moore and his work at U-M’s Child and Adolescent Data Lab. School of Social Work faculty Joseph Ryan and Brian Perron launched the lab last year with one goal in mind: harness the power of data to improve outcomes for vulnerable youth and their families.

It’s too early to declare victory, but multiple state agencies across Michigan already have partnered with the lab to help them answer policy questions using big data.

“We are not interested in collecting data simply for the sake of writing articles and filling a library,” Ryan said. “From my perspective, that just seems like a waste of time. We are using real data to help inform policy and essentially improve the welfare of our state.”

A unique approach

Big data is nothing new to state government.

Many state agencies rely on data when making important policy decisions.

But how often do they share their data?

Not very often, according to Ryan, which poses some serious problems when public policy is concerned.

“When agencies use data, they often rely strictly on their own numbers,” he said.

Ryan recognized a similar problem in Illinois, so he partnered with a number of state agencies there to integrate massive datasets into a cohesive database. That database impacted how policy decisions were made in Illinois because agencies could use real-time data to develop effective interventions to address child welfare issues.

When Ryan joined U-M in 2011, he began reaching out to state agencies to determine how big data could help them tackle important issues.

Four years later, Ryan and Perron launched the Child and Adolescent Data Lab. And seven months in, state agencies are quickly recognizing its importance.

That includes the Michigan Department of Health and Human Services (MDHHS), which already has partnered with the lab to help improve the way in which the state’s child welfare system monitors performance.

“Big data plays an ever-important role in shaping policy decisions,” said Debora Buchanan, director of the MDHHS Division of Continuous Quality Improvement. “Through our partnership with the University of Michigan, we are able to tap into their resources and use big data to determine the most effective ways to move our state forward in a positive direction. This partnership has the potential to impact a variety of policy decisions across the state of Michigan in the coming years.”

Here’s how it works. State agencies identify a problem that could potentially be addressed through big data like:

• Rates of foster youth with mental health and substance use disorders
• Classroom attendance and achievement of youth within the juvenile justice system
• Maltreatment in substitute care settings
• Initial offending patterns and recidivism in juvenile court
• Obstacles Michigan foster youth face as they transition to adulthood

The state then contacts the lab and supplies it with massive datasets, which could include anything from child abuse records to Medicaid claims to education reports.

Agencies often require a quick turnaround, so the team sifts through these datasets in search of an answer.

For instance, the governor’s office was looking at ways to address prison reform in Michigan.

“Of the counties that have implemented various reform efforts, which ones worked?”

Ryan and Perron analyzed data from Livingston County, and then supplied the governor’s office with a research brief that outlined the county’s recent efforts to invest more resources into its community-based placements.

“In a traditional academic sense, faculty collect data and write at their own pace,” Ryan said. “The consequences are on us if we don’t get our papers out the door. But in terms of our partnership with the state, we understand they have questions that need to be answered quickly. That’s why our lab operates under a quick-response approach.”

Tangible results

Erin Rhodes understands the foster care system.

The University of Michigan junior was homeless and he lived in a facility with foster youth. It was then that Rhodes vowed to help improve the system.

Rhodes eventually overcame homelessness, enrolled in a community college and then transferred to U-M to study economics. He enrolled in the university’s Undergraduate Research Opportunity Program (UROP), and when searching for projects, he came across the Child and Adolescent Data Lab.

“This project really hits close to home because I understand the people they are studying,” he said. “My own personal experience definitely provides added motivation for me to find solutions to help improve the lives of foster youth.”

Rhodes is one of three U-M students who work in the lab, and he focuses on identifying behavioral patterns among foster youth.

Perron works closely with Rhodes and other students as they navigate complicated datasets, but the data itself won’t fix anything.

“Our work alone is not going to solve the problems facing our state, but it will certainly enrich our understanding of those issues,” Perron said. “That’s what makes this partnership so productive. We at the university have an expertise in data management, but the state agencies have expertise in the field so they can take our insights from the data and put it to good use.”

And as the state continues to pose additional questions, Perron and Ryan plan to recruit more students to assist with everyday tasks. And in terms of students’ overall educational experience, it’s certainly a value added.

“Learning something in the classroom isn’t always enough,” said Barbara Hiltz, U-M clinical assistant professor of social work. “That’s why we’re hoping to provide our students with more of a of hands-on experience. Often in academia, we do work, but we don’t always see the results. This is different. We are forming long-term partnerships with state agencies and we are helping to influence public policy. That’s pretty meaningful.”