Improving TAVR Procedural Volume Disparities Among Active Cancer: Propensity Score and Machine Learning Case-Control Analysis of 101 Million+ Hospitalizations 2016-2018.

Document Type

Conference Proceeding

Publication Date


Publication Title

Journal of the Society for Cardiovascular Angiography & Interventions



We performed the largest multi-year nationwide study of mortality, cost, and procedural volume of transcatheter aortic valve replacement (TAVR) for patients with and without active cancer.


This case-control study of the above outcomes was conducted using the United States’ largest all-payer hospitalized dataset, the 2016-2018 National Inpatient Sample (NIS). Machine Learning-augmented Propensity Score adjusted multivariable regression (ML-PSr) was utilized, weighted by the NIS complex survey design, and adjusted for known confounders (including NIS-calculated mortality risk by DRG and the likelihood of undergoing TAVR versus medical management).


Among 101,521,656 hospitalizations, 444,748 (0.14%) underwent TAVR among whom 254,992 (0.08%) had active cancer. Amid the stepwise increase in TAVR volume annually across all patients, patients with versus without active cancer were significantly less likely to undergo TAVR in 2016 (0.07 versus 0.14%), 2017 (0.08 versus 0.15%), and 2018 (0.09 versus 0.17%) (all p<0.001), despite comparable mortality. Procedural disparity was noted among age and morality risk-matched patients in 2016 (4.07 versus 11.56%, p<0.001), but rates were comparable in 2017 and 2018. There were no significant procedural disparities among patients with versus without cancer by sex, race, income, or urban density, but there were for regions (the highest TAVR volume in patients with cancer were done in the Northeast [4.80%] and New England [4.38%], p=0.048). In multivariable regression across all years, active cancer did not significantly increase mortality following TAVR (OR 0.47, 95%CI 0.19-1.18; p=0.109), but actually reduced mortality in sub-group analysis among those with active cancer (OR 0.24, 95%CI 0.10-0.58; p=0.002). Sub-group analysis among those with versus without cancer found only marginally increased total costs ($2,117, 95%CI $642-4876, p<0.001).


This longitudinal propensity score analysis indicates that TAVR may have similar safety and cost for patients with and without active cancer, which may be reflected in the increasing rate of TAVR among the former.




3 Suppl

First Page



SCAI 2022 Scientific Sessions, May 19-22, 2022, Atlanta, GA.