Quality Assessment in Paediatric Cardiology: Experiences from Leveraging a Clinical Data Warehouse.

BACKGROUND

The generation of quality metrics for paediatric heart centre programmes frequently relies on registry data, with all the known benefits and disadvantages. This retrospective monocentric study introduces an algorithm capable of processing unedited clinical data to identify mortality risk factors following paediatric cardiac surgery.

METHODS

Patients who had undergone cardiac surgery in the department during the period from 2011 to 2020 were included when aged < 18 years. Congenital heart disease (CHD) was categorised into four diagnosis groups through hierarchical integration of the index surgery and CHD diagnosis. We evaluated preoperative, demographic, periprocedural, and postsurgical risk factors.

RESULTS

A total of 1700 patients with 2157 hospitalization encounters were included. The risk factors for elevated mortality with the highest degree of significance were extracorporeal membrane oxygenation (hazard ratio 13.97, p < 0.001), weight < 2500 g, patients in the univentricular heart group I, and the creatinine ratio.

CONCLUSIONS

Beyond confirming established predictors such as ECMO and low body weight < 2500 g, the present analysis highlights the creatinine ratio as a strong laboratory-based predictor of mortality. The applied framework serves as a foundational step towards enabling the real-time utilisation of raw datasets across multiple centres, thereby supporting privacy-preserving and efficient quality metric assessment as well as enhanced risk stratification.