Predicting the duration of surgeries is an important task because of the many dependencies between surgery processes and the hospital processes within other departments. Thus, accurate predictions allow for better coordinating patient processes throughout the hospital. Prior data-driven research provides evidence for accurate predictions of surgery durations enhancing the efficiency of surgery schedules. However, the current prediction models require large sets of features, which make their adoption more intricate. Moreover, prediction models focus on the surgery department and neglect potential effects on other departments. We use a unique dataset of about 17,000 surgeries to study how particular features and machine learning algorithms affect the prediction accuracy of major surgery steps. The prediction models that we study require few features and are easy to apply. The empirical findings can be useful for the design of surgery scheduling systems.
Riekert, Martin; Premm, Marc; Klein, Achim; Kirilov, Lyubomir; Kenngott, Hannes; Apitz, Martin; Wagner, Martin; and Ternes, Lena, (2017). "PREDICTING THE DURATION OF SURGERIES TO IMPROVE PROCESS EFFICIENCY IN HOSPITALS". In Proceedings of the 25th European Conference on Information Systems (ECIS), Guimarães, Portugal, June 5-10, 2017 (pp. 2842-2851). ISBN 978-0-9915567-0-0 Research-in-Progress Papers.