Objective acuity measures for infusion therapy staffing: A predictive model approach—Phase I.

Abstract

e13553 Background: Acuity is a measure of the complexity of patient care and an estimate of nursing workload. The use of acuity tools is shown to improve care standards, nurse satisfaction and productivity. Existing acuity tools have barriers to adoption, including increased nursing workload, estimate variations among different nurses and clinical care settings, and poor understanding of validity of acuity measures. We aim to build a predictive model for prospectively determining acuity levels and staffing requirements in infusion therapy to address these gaps. This project will be implemented with a phased approach. Here we describe Phase I, to establish a standard measure and documentation process for acuity, then collect and analyze a large data set to quantify the relationship of acuity scores to selected patient characteristics. Methods: A US Oncology Network nursing committee developed evidence-based guidelines for acuity measurement and documentation in August 2023. Nine sites from four independent, community oncology practices were selected to perform acuity rating and documentation using the developed guidelines. Multiple training sessions on the guidelines were conducted with nurses from selected sites. Nurses documented acuity scores (level 1~6) and start time for patients’ infusion visits in the EHR 8/30/2023 – 11/30/2023. Acuity documentation rates were calculated and provided to each site weekly to gain feedback and engagement. Acuity scores were extracted from the electronic health record using natural language processing. Structured data on potential predictors including patient demographics, performance status, lab, and treatment-related information were also collected. Results: A total of 14,610 infusion room visits (5,139 distinct patients) with documented acuity scores were reported. The acuity documentation rate from each site was consistent by week and ranged 60 – 96%. Table reports prevalence of select features by acuity score. Conclusions: An operational workflow was successfully implemented to collect and analyze acuity scores and relevant features. Features of interest include chemotherapy Cycle 1 Day 1, five or more drugs and highly emetogenic treatment. In Phase II, the collected acuity scores and clinical features will be used to build a predictive model to objectively estimate acuity. Phase III will focus on the deployment of the predictive model to oncology practices. [Table: see text]