# Ensemble Prediction of Time to Event Outcomes with Competing Risks: A Case Study of Surgical Complications in Crohn's Disease

## TL;DR

This paper introduces a new ensemble machine learning algorithm that predicts the risk of major abdominal surgery within 5 years for Crohn's disease patients, effectively handling competing risks and censored data.

## Contribution

It presents a novel pseudo-observation based ensemble approach that extends existing machine learning methods to right-censored event data in medical prognosis.

## Key findings

- Effective prediction of surgical risk within 5 years.
- Extension of machine learning methods to censored data.
- New estimators for model evaluation and optimization.

## Abstract

We develop a novel algorithm to predict the occurrence of major abdominal surgery within 5 years following Crohn's disease diagnosis using a panel of 29 baseline covariates from the Swedish population registers. We model pseudo-observations based on the Aalen-Johansen estimator of the cause-specific cumulative incidence with an ensemble of modern machine learning approaches. Pseudo-observation pre-processing easily extends all existing or new machine learning procedures to right-censored event history data. We propose pseudo-observation based estimators for the area under the time varying ROC curve, for optimizing the ensemble, and the predictiveness curve, for evaluating and summarizing predictive performance.