Predicting short-term outcomes after transcatheter aortic valve replacement for aortic stenosis

Area of Science:

• Cardiology
• Medical Informatics
• Health Services Research

Background:

• Transcatheter aortic valve replacement (TAVR) for aortic stenosis has expanded, but accurately predicting patient outcomes is still a challenge.
• Existing risk stratification methods may not fully capture the complexity of post-TAVR outcomes.
• There is a need for improved prediction models for both clinical and patient-centered outcomes.

Purpose of the Study:

• To develop and evaluate prediction models for 30-day clinical and patient-centered outcomes following TAVR.
• To assess the predictive performance of models using integrated data sources.
• To identify key predictors for short-term outcomes after TAVR in a large, diverse population.

Main Methods:

• A cohort of 1,565 adult patients undergoing TAVR from 2013-2019 was identified.
• Gradient boosting machines were used to develop prediction models incorporating data from the TVT Registry and electronic health records.
• Model performance was evaluated using area under the curve (AUC) for discrimination and calibration plots.

Main Results:

• The risk of adverse 30-day outcomes varied, from 1.3% for heart failure hospitalizations to 15.3% for all-cause emergency department visits.
• Model discrimination was moderate for death (AUC 0.60) and quality of life (AUC 0.62), but better for heart failure-related ED visits (AUC 0.76).
• The STS risk score only predicted death and all-cause hospitalization; older age predicted HF-related ED visits; race/ethnicity was not significant.

Conclusions:

• Predicting short-term clinical and patient-centered outcomes after TAVR remains difficult despite comprehensive data.
• Current models show moderate predictive power, highlighting the need for novel predictors.
• Further research is essential to enhance personalized decision-making and monitoring strategies for TAVR patients.