Linking Medicaid and Child Welfare Data to Identify High Service Users in Child Welfare
Data sharing across Medicaid and child welfare is rare, despite the fact that both systems play important roles in providing services for youth in foster care. By linking child welfare and Medicaid data, a new, more comprehensive perspective on service use is gained that can improve service delivery and outcomes for youth in care. Casey Family Programs and Mathematica Policy Research, Inc. partnered to take this on.
The challenge
Two sites participated in a study to link child welfare and Medicaid data to identify high service users. The study was based on a sample of children in foster care during a five year period in each state.
The undertaking
Tennessee
Descriptive and latent class analysis sample: Sample includes children who enter out-of-home care between July 1, 2011, and December 31, 2015. (N=21,672)
Predictive analysis sample: Sample includes children who enter out-of-home care between July 1, 2012, and January 31, 2014. (N=12,056)
Tri-County Area in Florida
Descriptive and latent class analysis sample: Sample includes children who enter out-of-home care between September 1, 2013, and December 31, 2015. (N=6,695)
Predictive analysis sample: Sample includes children who enter out-of-home care between July 1, 2012, and January 1, 2014. (N=8,290)
Pasco county, Pinellas county and Hillsborough county
Superutilization
Superutilization of services was measured across the data sources on the following dimensions:
-
Measured by:
- number of custody episodes
- number of placement moves
- number of child welfare services
- number of inpatient, outpatient, and emergency Medicaid services
- non-Medicaid mental health and substance abuse services per year (for Florida only)
-
Measured by:
- average share of time spent in group home or residential treatment care
-
Measured by:
- total length of stay in out-of-home custody
-
Measured by:
- cost of child welfare placements
- cost of child welfare services
When comparing children identified as experiencing superutilization to those who did not, we find those experiencing superutilization had the following attributes:
-
- Higher proportions of adolescents and infants
- Higher proportions investigated for maltreatment while in foster care
- Higher proportions removed from home due to a behavior problem
- Lower proportions removed from home for reasons of parental drug abuse
- Higher proportions with high needs and less developed life skills as measured on the Child and Adolescent Needs Assessment and Ansell-Casey Life Skills assessment, respectively
- Higher proportions with prior investigations and out-of-home placements
- Lower proportions of exiting state custody
- Higher proportions of exiting through emancipation, among those who exited state custody, and lower proportions exiting to guardianship and relative care
-
- Higher proportions of adolescents and lower proportions between 1 and 5 years old
- Higher proportions of males
- Higher proportions adopted and lower proportions reunified among those who exited state custody
- Higher proportions with prior investigations and out-of-home placements
Tennessee and Florida had many similar and a few distinct types of superutilization. In both locations, children were high service users due to:
In addition, both Tennessee and Florida had conceptually distinct types of superutilization.
Digging deeper
Both Tennessee and Florida were interested in a deeper understanding of placement instability.
Children in foster care who experience high placement instability are at greater risk for:
- emotional and behavioral disorders
- attachment disorders
- lower academic achievement
- lower likelihood of relational permanence
Relational permanence refers to many types of important long-term relationships that help a child or young person feel loved and connected – relationships with brothers and sisters, family friends and extended family, former foster family members and other caring adults.
Available prior child welfare, Medicaid, and non-Medicaid substance use and mental health data were used to predict high placement instability.
While each site had its own set of unique predictors, there were some similarities among the most important predictors identified for each site. Children had higher risk of placement instability if they:
- were older at the time of entry
- had more prior child welfare investigations
- previously spent more time in foster care
- had Medicaid-covered outpatient services for physical or behavioral health
- had Medicaid-covered emergency services for physical health incidents
Children had higher risk of placement instability over the course of 12 months after entering foster care if they:
-
- Were older at the time of entry into this custody episode
- Had more prior lifetime child welfare investigations
- Had more Medicaid-covered outpatient services for physical health in the year prior to this custody episode
- Had more Medicaid-covered outpatient services for behavioral health in the year prior to this custody episode
- Had more Medicaid-covered emergency services for physical health in the year prior to this custody episode
- Were removed for child behavioral reasons and neglect
- Spent more time in foster care prior to this custody episode over the course of their lifetime
- Received more in-home CW services in year prior to this custody episode
-
- Were older at the time of entry into this custody episode into this custody episode
- Spent more time in foster care prior to this custody episode over the course of their lifetime
- Had more prior lifetime child welfare investigations
- Had more Medicaid-covered outpatient services for physical health in the year prior to this custody episode
- Had more Medicaid-covered outpatient services for behavioral health in the year prior to this custody episode
- Had more Medicaid-covered emergency services for physical health in the year prior to this custody episode
- Had more non-Medicaid-covered substance abuse services in the year prior to this custody episode
- Had more Medicaid-covered inpatient services for behavioral health in the year prior to this custody episode
- Had more lifetime removals to out-of-home care
- Had more non-Medicaid covered mental health services in the year prior to this custody episode
What can we do with this information?
Child welfare agencies can partner with other systems to identify youth experiencing superutilization services, tailor interventions to better address their needs, and prevent placement instability for youth in care.
The following programs are listed in the California Evidence-Based Clearinghouse as having at least promising evidence for addressing adolescent behavioral health and placement stabilization:
- Aggression Replacement Training
- Attachment Biobehavioral Catch-up (ABC)
- Cognitive Behavioral Therapy
- Dialectical Behavior Therapy (DBT)
- Ecologically Based Family Therapy
- Family Group Decision Making
- Functional Family Therapy
- Generation PMTO (formerly Parent Management Training – Oregon Model)
- Keep (Keeping Foster and Kin Parents Supported and Trained)
- Keep Safe
- Life Space Crisis Intervention
- Multisystemic Therapy
- Multidimensional Family Therapy
- Neighbor to Family Sibling Foster Care Model
- PAX Good Behavior Game
- Teaching-Family Model
- Trauma-Focused Cognitive Behavioral Therapy
- Treatment Foster Care Oregon for Adolescents and Preschoolers
- Wraparound
In summary, this study provides much-needed insight, through linked administrative data and advanced statistical methods, into the superutilization of services among children in foster care. The results of this study can be used to improve the delivery of services to improve outcomes for children and families. Stay tuned for stories about how Florida and Tennessee are using these results.
Access an issue brief on child welfare and Medicaid.
More information about the study, including the executive summary and full report.
For more information contact the research team at Casey Family Programs Researchteam@casey.org or Elizabeth Weigensberg, the study’s project director at Mathematica Policy Research, at EWeigensberg@mathematica-mpr.com.
Research done in partnership with:
