Automated Patient Linking for Electronic Health Record and Child Welfare Databases

Area of Science:

• Health Informatics
• Child Welfare Systems
• Data Integration

Background:

• Over 400,000 children are in U.S. protective custody.
• Fragmented data systems hinder critical health information transfer to caregivers.
• This data gap impacts the health outcomes of children in foster care.

Purpose of the Study:

• To develop and assess automated algorithms for linking child welfare data with electronic health records.
• To improve the accessibility of comprehensive health histories for children in protective custody.

Main Methods:

• Created ten custom algorithms for data linkage.
• Employed deterministic matching (name, DOB, gender) and non-deterministic matching (handling variations).
• Evaluated the performance of both matching strategies.

Main Results:

• Deterministic matching achieved a 91.3% linkage rate.
• Non-deterministic matching successfully linked 71.3% of remaining records.
• Demonstrated a scalable, automated approach to merge disparate datasets.

Conclusions:

• Integrated data sharing is crucial for enhancing health outcomes in protective custody.
• The developed algorithms offer a viable solution for merging child welfare and health data.
• This automated, scalable method can bridge critical information gaps for vulnerable children.