Ieva Sriubaite¹ (with Anthony Harris² and Andrew M. Jones³)

1 Monash University/CINCH

2 Monash University

3 University of York

This paper employs methods of supervised machine learning to construct a risk adjustment tool for a set of outcomes that describes the economic consequences of the road-traffic injury. We focus on the prediction of healthcare costs and benefits from medical care in terms of both productivity as well personal well-being (the quality of life). Using the Victorian State Trauma Registry, we select all patients who experienced a major trauma in a road-traffic related accident in Victoria. To tackle statistically challenging empirical distributions we set up an ensemble machine learning algorithm - the Super Learner algorithm that is based on several parametric and non-parametric algorithms including regularized regressions, decision trees and random forests. Our findings demonstrate that the Super Learner is effective and performs best in predicting all outcomes considered in this paper.