A new study published in the International Journal of Population Data Science (IJPDS), highlights the accuracy of neonatal and perinatal health data routinely collected by Canada's Discharge Abstract Database (DAD). 

DAD is an administrative database that collects comprehensive data on hospitalisations across all Canadian provinces and territories, except Quebec. Due to its linkage with other administrative databases, DAD data can be used for population-based epidemiological research. However, the accuracy of this data, particularly related to preterm births, has not been comprehensively validated.

Researchers from the University of Manitoba’s Department of Pediatrics and Child Health compared DAD data with the Canadian Neonatal Network (CNN) database, which is well known for its high levels of accuracy. CNN collects data on very preterm infants from level 3 Neonatal Intensive Care Units (NICUs) across Canada. The study focused on neonates born before 33 weeks of gestation in Winnipeg, Manitoba, between 2010 and 2022. It compared 55 different parameters between the DAD and CNN databases, including maternal characteristics, diagnoses, neonatal diagnoses, outcomes, and procedures. 

The results showed variable accuracy across the parameters. Some neonatal variables like gestational age and birth weight were highly accurate. However, by contrast, some neonatal procedures like mechanical ventilation, laparotomy and gastrostomy had a lower accuracy. Other neonatal variables such as meningitis, necrotizing enterocolitis, bronchopulmonary dysplasia and sepsis, had variable accuracy.

Similarly, maternal variables, such as Caesarean delivery and multiple delivery, showed very good agreement, with others such as maternal pregnancy complications, e.g. gestational diabetes, substance abuse in pregnancy showing lower accuracy and agreement.

A major finding for the study was that the accuracy of DAD did not improve significantly after transitioning to electronic medical records. This suggests potential issues with data documentation by healthcare providers rather than the data entry system itself.

Overall, the study highlights that while administrative data like the DAD is useful for research, its accuracy can be inconsistent depending on the specific variables used. For preterm birth-related research, researchers should be cautious about assuming the accuracy of all the parameters in DAD and either consider the results of our research, or validate the key variables themselves independently.

This study calls for identifying the potential causes for this lack of accuracy in DAD such as inadequate or inaccurate documentation by physicians and or the errors occurring during the data abstraction process. Once the causes are identified, appropriate strategies should be adopted to improve the quality of DAD data. Collaborations with national organisations may be necessary to address these issues and improve the data quality in DAD.

 

Click here to view the full article

Dr Deepak Louis, Assistant professor, Max Rady College of Medicine Pediatrics and Child Health, University of Manitoba, Canada

Louis, D., Eshemokhai, P., Ruth, C., Cheung, K., Lix, L., Flaten, L., Shah, P. and Garland, A. (2024) “Validation of Preterm Birth Related Perinatal and Neonatal Data in the Canadian Discharge Abstract Database to Facilitate Long-term Outcomes Research of Individuals Born Preterm”, International Journal of Population Data Science, 9(1). doi: 10.23889/ijpds.v9i1.2380.