Data accuracy in the Ontario birth Registry: a chart re-abstraction study

Area of Science:

• Health Informatics
• Reproductive Health
• Data Quality Assurance

Background:

• The Birth Registry of Ontario (BORN) was established in 2009 to enhance healthcare through critical data on pregnancy, birth, and early childhood.
• High-quality routinely-collected health data is crucial for patient care, health system planning, and research globally.

Purpose of the Study:

• To verify the accuracy of selected data elements recorded in the Ontario birth registry.
• To assess the quality of routinely-collected health data for improving healthcare provision.

Main Methods:

• Compared data re-abstracted from patient records against data in the Ontario birth registry.
• Audited 100 linked mother-newborn charts from each of 10 purposively selected hospitals.
• Utilized percent agreement, kappa statistics (categorical), and intra-class correlation coefficients (continuous) for 29 data elements.

Main Results:

• Overall agreement ranged from 56.9% to 99.8%, with 76% of elements exceeding 90% agreement.
• Almost perfect or substantial agreement was found for 12 categorical elements.
• Moderate-to-excellent agreement was observed for four continuous data elements.

Conclusions:

• The evaluated data elements in the birth registry demonstrated good agreement with patient chart data.
• Data elements with moderate kappa or low ICC require further investigation to ensure data integrity.