Validation of infant immunization billing codes in administrative data

Area of Science:

• Health Informatics
• Vaccinology
• Public Health

Background:

• Ontario's universal health care system utilizes administrative data for health studies.
• Vaccine-specific billing codes were implemented in 2011.
• Validating these datasets is crucial for accurate post-marketing vaccine surveillance.

Purpose of the Study:

• To validate Ontario's administrative health datasets for assessing infant immunization status.
• To evaluate the performance of Ontario Health Insurance Plan (OHIP) database codes for primary infant vaccines.

Main Methods:

• Compared OHIP database codes against Electronic Medical Record Administrative data Linked Database (EMRALD) as the reference standard.
• Assessed four primary infant immunizations: DTaP-IPV-Hib, pneumococcal, MMR, and meningococcal C vaccines.
• Calculated specificity, sensitivity, positive predictive value (PPV), and negative predictive value (NPV).

Main Results:

• OHIP claims demonstrated high specificity (81-92%) and PPV (97-99%).
• Sensitivity ranged from 70-83%, with NPVs between 13-46%.
• Using both specific and general codes improved sensitivity during the new code introduction phase.

Conclusions:

• Ontario's administrative data offers high specificity and PPV for vaccine billing claims but has limitations in sensitivity and NPV.
• These data have utility for post-marketing vaccine studies but are insufficient for measuring vaccine coverage and effectiveness.
• A comprehensive, linkable provincial immunization registry is recommended as a priority.