What does administrative data tell us about youth behavioral health emergency service usage?

Why is this issue important?

Thousands of Rhode Island children visit an emergency department for a behavioral health crisis each year. The state has long recognized that behavioral health—not just physical health—is critical for children’s well-being. However, recent research suggests that treating behavioral health using emergency services is less effective than non-emergency services. Identifying gaps in service provision and understanding patterns of service use among at-risk children with behavioral health needs is vital to providing effective services that allow children to thrive.

What did we do?

We conducted a descriptive analysis of a cohort of children who were hospitalized or who visited the emergency department (ED) in 2017 with a primary behavioral health diagnosis. We follow this cohort for three years, from 2016 to 2018.

By combining Medicaid claims, early childhood demographics, and data from the Department of Children, Youth, and Families on placements of children, we examined service use patterns to highlight potential points for intervention.

What have we learned (so far)?

The main cohort of 3,235 individuals had 2,899 inpatient visits and 4,064 ED visits in 2017 alone. The most striking finding of the analysis showed that 35% of these individuals account for 86% of the group’s total Medicaid spending (see Figure 1 for an illustration of this concentration of spending). Further, a small fraction (15%) of claimants visited emergency departments four or more times over three years, with a nontrivial number visiting 20-80 times.

These “high-utilizer” children tend to have more co-occurring and complex behavioral health disorders, which are likely more difficult for families to manage. They also tend to visit the emergency department more often and have frequent placements with the Department of Children, Youth, and Families.

What happens next?

Rhode Island’s Executive Office of Health and Human Services (EOHHS) is taking the findings of this analysis into consideration as they determine potential points of intervention to provide more effective services to these at-risk children. Potential actions include considering both a behavioral health mobile crisis center and exploring ways to strengthen provider ties across sectors.

The project also had an unexpected impact by revealing a data quality problem which has since been improved: Medicaid claim IDs are now correctly matched to data in DCYF’s system for nearly ten times as many children as before the project started.