ORACLE

Important Notice

Research Prototype — Not for Clinical Use

This tool was developed using real-world data from more than 600 SynCardia 70cc Total Artificial Heart implantations. It predicts 12-month device outcome (device success — alive on device or transplanted — versus on-device death) using 13 pre-implant clinical variables.

The model demonstrated consistent but modest discrimination, with an optimism-corrected c-statistic of 0.65 and identical performance in temporal validation (c-statistic 0.65), suggesting stable predictive performance without substantial evidence of overfitting. However, the model has not undergone external validation in an independent patient cohort and has not been evaluated for clinical decision-making.

This tool is not a medical device and must not be used to guide the care of any individual patient, including treatment selection, candidacy assessment, risk counselling, or clinical management decisions.

The tool is intended solely for educational and research purposes, to explore how pre-implant characteristics may influence outcomes at a cohort level. Predictions reflect historical patterns observed within the training dataset and may not generalize to contemporary practice, other institutions, different patient populations, or other mechanical circulatory support devices.

Clinical decisions should always be based on comprehensive patient assessment and the judgment of appropriately qualified healthcare professionals.