Deduplication of temporary names in an immunization registry using probabilistic linkage

Area of Science:

• Public Health
• Health Informatics
• Data Science

Background:

• Immunization Information Systems (IIS) often receive records with temporary child names.
• Vaccinations administered before a legal name is assigned can lead to duplicate records.

Purpose of the Study:

• To apply probabilistic record linkage for deduplicating temporary name records in an IIS.
• To assess the impact of deduplication on childhood vaccination coverage rates.

Main Methods:

• Cross-sectional study utilizing vaccination records from the Michigan Care Improvement Registry (MCIR).
• Probabilistic record linkage was used to match temporary name records to legal name records.
• Analyzed changes in vaccination coverage for the 4:3:1:3:3:1:4 series among 19-35 month olds.

Main Results:

• 58% (9803 of 16,806) of temporary name records were successfully linked to a legal name record.
• For children with temporary names, 4:3:1:3:3:1:4 series completion increased from 1.5% to 39.6% post-reconciliation.
• Statewide, 4:3:1:3:3:1:4 series coverage for 19-35 month olds increased from 66.2% to 68.2%.

Conclusions:

• Probabilistic linkage is an effective method for identifying and deduplicating temporary name records in IIS.
• Reconciling these records significantly improves the accuracy of vaccination coverage data.
• Accurate IIS data is crucial for monitoring and improving public health initiatives like childhood immunization programs.