COMPASS: A novel risk-adjustment model for catheter ablation in pediatric and congenital heart disease patients

Background: Robust risk-adjustment algorithms are often necessary if data from clinical registries is to be used to compare rates of important clinical outcomes between participating centers. Although such algorithms have been successfully developed for surgical and catheter-based cardiac interventions in children, outcomes of pediatric and congenital catheter ablation have not been modeled with respect to case mix. Methods: A working group was appointed by the Pediatric and Congenital Electrophysiology Society to develop a risk-adjustment algorithm for use in conjunction with a modernized, multicenter registry database. Expert consensus was used to develop relevant outcome measures, an inclusive list of possible predictors, and estimates of associated incremental risk. Historical data from the Pediatric Radiofrequency Ablation Registry was reanalyzed using multivariate regression to create statistical models of ablation outcomes. Results: Acute ablation failure and serious adverse event rates were modeled as outcomes. Statistical modeling was performed on 4486 cases performed in 19 centers. For ablation failure rate, a simple model including general category of arrhythmia mechanism and presence of structural congenital heart disease accounted for ~71% of outcome variance. The model was useful for identification of between-center variability in the historical data set. Although expert consensus predicted the need for a more complex model, predicted univariate effects were similar to those generated by statistical modeling. Serious adverse events were too infrequent to permit statistical association with any predictive variable, but could be compared with the mean rate observed among all centers. Conclusion: A substantial component of the intercenter variability of acute ablation outcomes in a historical database of pediatric and congenital ablation patients may be accounted for by a simple statistical model, exposing variations in outcome specific to centers. This will be a useful initial model for use a modern registry for pediatric catheter ablation outcomes.