This app calculates the
aTRIMpost
, a random forest based risk score for 30 day survival post TAVI.
The undnerlying random forest model
was trained on 22,283 TAVI interventions in the
German Aortic Valve Registry
of the years 2011-2016.
On test data (5864 interventions from 2017) classification performance reached an AUC of 0.74; 95%-CI [0.70; 0.78].
For details see Leha et al., 2023, European Heart Journal - Digital Health,
doi:10.1093/ehjdh/ztad021
.
The break down plot is based on synthetic data generated using
doi:10.18637/jss.v074.i11
to resemble the properties of the test data.