Improving HIV Surveillance Data by Using the ATra Black Box System to Assist Regional Deduplication Activities

Area of Science:

• Public Health
• Epidemiology
• Health Informatics

Background:

• High-quality HIV surveillance data is crucial for effective public health programs.
• The Data to Care initiative highlights the need for accurate HIV case records.
• Existing methods for resolving duplicate HIV case records across jurisdictions can be time-consuming.

Purpose of the Study:

• To identify and quantify duplicate and exact duplicate HIV case records across multiple Enhanced HIV/AIDS Reporting System (eHARS) databases.
• To compare an automated approach to traditional manual methods for resolving interstate duplicate HIV case records.
• To assess the efficiency and accuracy improvements offered by the automated system.

Main Methods:

• Utilized the ATra Black Box System for matching case records.
• Employed 6 eHARS variables (name, DOB, sex, SSN, race/ethnicity) and 4 system-calculated values for record linkage.
• Processed 799,326 uploaded records from 9 eHARS databases across 8 jurisdictions.

Main Results:

• Successfully matched 290,482 cases, identifying 55,460 exact case pairs in approximately 11 hours.
• Significant overlaps identified between specific jurisdictions (e.g., NYC and NYS at 51%).
• Jurisdictions reported an estimated 135 labor hours saved compared to manual methods.

Conclusions:

• The automated approach successfully identified previously undiscovered exact matches.
• This method reduces the time and effort required for resolving duplicate HIV case records.
• Improved accuracy and completeness of HIV surveillance data support public health policies; standardization of postprocessing is recommended.