O4349: Implementation of secure data sharing and digital health tools to improve identification and care delivery for children in foster and kinship care

Background: Identification of children in foster and kinship care and knowledge of their unique needs is critical to quality healthcare for this vulnerable population. Digital health tools can effectively identify and support decision-making in populations with special healthcare needs. While data sharing between healthcare and child welfare systems is a priority in Washington state, there are currently no automated tools in place for health care providers to identify children in foster care.

Methods: Over an 18-month period, secure data sharing was implemented at a large academic health system along with targeted high yield education. Insurance verification codes, diagnosis codes, and patient emergency contact information within the electronic health record (EHR) were utilized to alert providers that a child may be in foster/kinship care. This alert directed providers to a set of tools including diagnosis codes, documentation templates, orders, community resources, and links to evidenced-based guidelines.

Using a quasi-experimental pre-post design, baseline data were collected 3 years pre- and 6 months post-implementation, comparing a general Medicaid to a foster/kinship care cohort. Implementation outcomes included correct usage of foster care diagnosis codes, referral to foster care-specific services and use of EHR tools. Additionally, qualitative interviews with 25 providers across 13 clinics were conducted, transcribed, and analyzed using thematic analysis.

Results: Among patients 0-21 years with Medicaid insurance, those in foster care comprised 5.1% (n=840) of the pre-implementation group (N=16,489) and 4.7% (n=312) of the post-implementation group (N=6,649). Several different identifiers were used to indicate whether a child was in foster/kinship care and the majority of children were identified by their insurance verification codes rather than an indication within the electronic health record (Table 1). Usage of the Foster Care diagnosis codes increased at well child exams, with no meaningful change at problem-focused visits. The use of a specific foster care billing modifier at wellness exams and referrals to foster care-specific care management services increased (Figure 1).

Key qualitative themes identified in the provider interviews were: (1) knowledge that a patient is foster care involved is essential to providing quality care; (2) the digital tools helped with identification and care provision for children in foster care; and (3) subsequent opportunities to improve utilization were identified.

Conclusion: Implementation of secure data sharing in a large health system to identify children in foster care, coupled with digital tools and education, led to increased use of diagnostic coding, enhanced billing, and improved utilization of targeted services. Feedback will be used to adapt implementation strategies to further increase utilization. Findings encourage piloting bidirectional data sharing between health and child welfare systems.