Fully independent external validation of the Transcatheter Aortic Valve Replacement 30-day (TAVR-30) hospital readmission model

Abstract

Background: Early and late readmissions after Transcatheter Aortic Valve Replacement (TAVR) are common and associated with worse outcome. A risk prediction model (TAVR-30) was recently developed using readily available clinical variables to identify patients at risk for hospital readmission within 30 days after TAVR. We performed an independent external validation of the TAVR-30 model.

Methods: The Swedish TAVR-registry, linked together with other mandatory national registries was used to identify all TAVR procedures, variables from the original model, hospitalizations and deaths between the years 2008 to 2021.

Results: A total of 8459 patients underwent TAVR, 7693 patients had complete data and were included in the analysis. Out of these, 928 patients experienced a readmission within 30 days. Using the estimates from the original model, a concordance (c)-index of 0.51, a calibration slope of 0.07 and intercept of -0.62 were obtained respectively, overall implying poor model performance.

Conclusions: This independent external validation indicates poor performance of the TAVR-30 model in a Swedish setting. Further research is needed to develop more reliable tools for predicting the risk of early hospital readmission after TAVR, as well as, for providing a deeper understanding of how to develop risk models that performs well in patients with multiple underlying comorbidities.

Keywords: Prediction model; Readmission; TAVR.