Monitoring Perioperative Services Using 3D Multi-Objective Performance Frontiers.

Affiliation

Elhajj AJ(1), Rizzo DM(2), An GC(3), Pandit JJ(4), Tsai MH(5)(6)(7).
Author information:
(1)College of Engineering and Mathematical Sciences, University of Vermont, Burlington, VT, USA.
(2)Department of Civil & Environmental Engineering, University of Vermont, Burlington, VT, USA.
(3)Department of Surgery, University of Vermont Larner College of Medicine, Burlington, VT, USA.
(4)Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK.
(5)Department of Anesthesiology, University of Vermont Larner College of Medicine, 111 Colchester Avenue, Burlington, VT, USA. [Email]
(6)Department of Orthopaedics and Rehabilitation
(by courtesy), University of Vermont Larner College of Medicine, 111 Colchester Avenue, Burlington, VT, USA. [Email]
(7)Department of Surgery
(by courtesy), University of Vermont Larner College of Medicine, 111 Colchester Avenue, Burlington, VT, USA. [Email]

Abstract

The Acute Care Surgery model has been widely adopted by hospitals across the United States, with Acute Care Surgery services managing Emergency General Surgery patients that were previously being treated by General Surgery. In this analysis, we evaluate the impact of an Acute Care Surgery service model on General Surgery at the University of Vermont Medical Center using three metrics: under-utilized time, spillover time, and a financial ratio of work Relative Value Units over clinical Full Time Equivalents. These metrics are evaluated and used to identify three-dimensional Pareto optimality of General Surgery prior to and after the October 2015 tactical allocation to the Acute Care Surgery model. Our analysis was further substantiated using a Markov Chain Monte Carlo model for Bayesian Inference. We applied multi-objective Pareto and Bayesian breakpoint analysis to three operating room metrics to assess the impact of new operating room management decisions. In the two-dimensional space of Fig. 2, panel a), the post-tactical allocation front lies closer to the origin representing more optimal solutions for productivity and under-utilized time. The post-tactical allocation front is also closer to the origin for productivity and spillover time as shown in the two-dimensional space of Fig. 2, panel b). The results of the three-dimensional multi-objective analysis of Fig. 3 illustrate that the GS post-tactical allocation Pareto-surface is contained within a much smaller volume of space than the GS pre-tactical allocation Pareto-surface. The post-tactical allocation Pareto-surface is slightly lower along the z-axis, representing lower productivity than the pre-tactical allocation surface. This methodology might contribute to the external benchmarking and monitoring of perioperative services by visualizing the operational implications following tactical decisions in operating room management.