An empirical analysis of predictors for workload estimation in healthcare

Abstract

The limited availability of resources makes the resource allocation strategy a pivotal aspect for every clinical department. Allocation is usually done on the basis of a workload estimation, which is performed by human experts. Experts have to dedicate a significant amount of time to the workload estimation, and the usefulness of estimations depends on the expert’s ability to understand very different conditions and situations. Machine learning-based predictors can help in reduce the burden on human experts, and can provide some guarantees at least in terms of repeatability of the delivered performance. However, it is unclear how good their estimations would be, compared to those of experts. In this paper we address this question by exploiting 6 algorithms for estimating the workload of future activities of a real-world department. Results suggest that this is a promising avenue for future investigations aimed to optimising the use of resources of clinical departments.