An Evidence-Based Approach in Assigning Patients to Nurses

Purpose: This article presents a data-integrated simulation based method to evaluate patient-to-nurse assignments ahead of a shift. Unlike patient classification systems, which just estimate the workload, this simulation tool quantifies the workload of nurses by simulating individual nurse’s workflow for the entire shift. Design: To enable building the simulation model, the nurse workflow data was collected from a north Texas hospital for six months. The observational workflow data for six months was continuously collected via locating devices that were embedded with the nurse badges. Methods: After collecting and processing the data, a data-mining model called classification and regression trees (CART) was fitted to the data. Based on the historical data on nurse workflow, the CART model unearthed patterns in nurse workflow for different patient-to-nurse assignments. A simulation code was written to repeatedly sample from CART to predict the workflow of nurses for the entire shift ahead of time. Findings/ Conclusion: Generating 1000 scenarios of nurse workflow for the entire shift took only fewer than three minutes. Hence, it is possible to implement the simulation on an actual care unit to quantify workload and evaluate balance in workload among nurses. Introduction of this simulation tool enable charge nurses and hospital management to adapt evidence-based practice in assigning patients to nurses by using performance measures from the simulation.