An official website of the United States government
Determining emergency department (ED) nurse staffing decisions to balance quality of service and staffing costs can be extremely challenging, especially when there is a high level of uncertainty in patient demand. Increasing data availability and continuing advancements in predictive analytics provide an opportunity to mitigate demand uncertainty by using demand forecasts. In this work, we study a two-stage prediction-driven staffing framework where the prediction models are integrated with the base (made weeks in advance) and surge (made nearly real-time) nurse staffing decisions in the ED. We quantify the benefit of having the ability to use the more expensive surge staffing and identify the importance of balancing demand uncertainty versus system stochasticity. We also propose a near-optimal two-stage staffing policy that is straightforward to interpret and implement. Last, we develop a unified framework that combines parameter estimation, real-time demand forecasts, and nurse staffing in the ED. High-fidelity simulation experiments for the ED demonstrate that the proposed framework has the potential to reduce annual staffing costs by 10%–16% ($2 M–$3 M) while guaranteeing timely access to care. This paper was accepted by David Simchi-Levi, healthcare management. Funding: J. Dong was partially supported by the Division of Civil, Mechanical and Manufacturing Innovation [Grant CMMI-1944209]. Supplemental Material: The data files are available at https://doi.org/10.1287/mnsc.2021.02781 .
more »
« less
This content will become publicly available on December 1, 2026
Implementing a prediction driven framework for emergency department nurse staffing to optimize real time decisions
Abstract This study implemented and evaluated a prediction-driven nurse staffing framework in a large adult emergency department. The framework leveraged a two-stage prediction model that forecasted patient volume and guided staffing decisions. Using a pre-post study design, we compared patient throughput (measured by door-to-evaluation time, active treatment time, boarding time, length of stay, and left-without-being-seen rate) and cost outcomes (measured as hourly nurse staffing costs) before and after implementation. The model achieved an RMSE of 11.261 and MAPE of 13.414% at the base stage, and an RMSE of 9.973 and MAPE of 12.126% at the surge stage. The framework reduced hourly staffing costs by $162.04 without negatively affecting throughput. Reducing one nurse per hour from the recommended level increased wait times by two minutes, with an additional 2.3-min increase when staffing dropped below 20% of recommendations. These findings highlight the potential of prediction-driven staffing to reduce costs while maintaining patient throughput.
more »
« less
- Award ID(s):
- 1944209
- PAR ID:
- 10670862
- Publisher / Repository:
- npj Health Systems
- Date Published:
- Journal Name:
- npj Health Systems
- Volume:
- 2
- Issue:
- 1
- ISSN:
- 3005-1959
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract IntroductionIn order to be positioned to address the increasing strain of burnout and worsening nurse shortage, a better understanding of factors that contribute to nursing workload is required. This study aims to examine the difference between order‐based and clinically perceived nursing workloads and to quantify factors that contribute to a higher clinically perceived workload. DesignA retrospective cohort study was used on an observational dataset. MethodsWe combined patient flow, nurse staffing and assignment, and workload intensity data and used multivariate linear regression to analyze how various shift, patient, and nurse‐level factors, beyond order‐based workload, affect nurses' clinically perceived workload. ResultsAmong 53% of our samples, the clinically perceived workload is higher than the order‐based workload. Factors associated with a higher clinically perceived workload include weekend or night shifts, shifts with a higher census, patients within the first 24 h of admission, and male patients. ConclusionsThe order‐based workload measures tended to underestimate nurses' clinically perceived workload. We identified and quantified factors that contribute to a higher clinically perceived workload, discussed the potential mechanisms as to how these factors affect the clinically perceived workload, and proposed targeted interventions to better manage nursing workload. Clinical RelevanceBy identifying factors associated with a high clinically perceived workload, the nurse manager can provide appropriate interventions to lighten nursing workload, which may further reduce the risk of nurse burnout and shortage.more » « less
-
Abstract Rapid nurse decision-making is needed to detect patient deterioration and prevent mortality. Current approaches to support nurses’ decisions involve diagnostic data processing and providing a decision with little explanation. Our team aimed to demonstrate the utility of attention architecture to model sequential nurse–patient care actions. Experienced nurses and students completed patient care simulations. Nurse actions were systematically coded and analyzed using our model, consisting of an attention encoder to sequentially process and predict nurse behavior. Performance of our model was compared to recurrent neural networks and long-short term memory models based on accuracy, precision, recall, and F1 score. Behavioral data from 24 nurses (11 experienced nurses and 13 nursing students) were collected during patient care simulations. Nineteen unique types of actions were distilled down to 8 common actions. There were 33 episodes captured (i.e., 33 unique sequences of patient care actions), including a total of 1024 actions (i.e., an average of 31 ± 11 actions). Results showed that the attention model outperformed the other models on all metrics except for precision. Our team demonstrated that machine learning can model sequential nurse actions. These results could be leveraged to provide real-time guidance to support novice nurses’ decision-making in the simulated environment.more » « less
-
Nursing homes (NHs) are critical facilities for caring frail older adults with around-the-clock formal care and personal assistance. To ensure quality of care for NH residents, an adequate staffing level is of great importance. Current NH staffing practice is mainly based on experience and regulation. The objective of this paper is to investigate the viability of experience-based and regulation-based strategies, as well as alternative staffing strategies to meet the heterogeneous service demand of NH residents at reduced labor cost under various scenarios of census compositions. We propose a predictive analytics integrated computer simulation model to characterize the heterogeneous service demand of NH residents, and further evaluate and identify promising staffing strategies at the facility level. Specifically, we propose a predictive model based on latent survival analysis to characterize diverse length-of-stay (LOS) with multiple discharge dispositions among NH residents. Further, we develop a simulation model with the incorporation of predictive analytics and domain knowledge to characterize the heterogeneous service demand of NH residents on different types of caregivers over time. Based on the simulation model, we develop a graphical user interface for the simulator to evaluate different staffing strategies at the facility level and inform NH administrators about promising strategies. We use real NH data to validate the proposed model and demonstrate its effectiveness. The proposed predictive LOS model considering multiple discharge dispositions exhibits superior prediction performance and offers better staffing decisions at reduced costs than those without the consideration. With the improved modeling fidelity via integrating predictive analytics with computer simulation, the proposed model is flexible to evaluate various staffing strategies using total labor cost as a performance metric, and can identify promising staffing strategies to meet the service demand of NH residents. Promising staffing strategies with the suggested staff-to-resident (SR) ratio can significantly reduce the total labor cost of multiple types of caregivers, as compared to the benchmark strategies, such as the SR ratios based on industrial practice or minimum requirement of state regulation. Moreover, we construct multiple scenarios of different census compositions of NH residents to demonstrate the capability of the proposed model. Our proposed model can facilitate NH staffing decision making to meet the heterogeneous service demand of NH residents at reduced labor costs.more » « less
-
Dissolved oxygen (DO) is a vital parameter in aquaculture, directly influencing fish health, growth and overall productivity. However, existing DO prediction methods largely rely on complex physical models that demand numerous parameter measurements and high computational cost. This study introduces a hybrid framework that integrates transfer learning (TL) with physics‐guided neural networks (PGNNs) based on long short‐term memory models (PG‐LSTM‐TL), to accurately predict pond DO levels from only 1 month of sparse and irregularly sampled nighttime data. The TL strategy leverages knowledge from a data‐rich tidal estuary and adapts it to a data‐scarce aquaculture setting, whereas the physics‐guided component embeds two key domain constraints: The temperature‐dependent DO solubility governed by Henry's law and the monotonic nighttime DO depletion trend. Unlike conventional TL approaches that primarily rely on fine‐tuning or statistical feature alignment, we embed physics‐guided constraints directly into the model learning stage. This integration mitigates negative transfer arising from environmental heterogeneity while preserving physically plausible multi‐step DO trajectories. Results show that, for each daily cycle (9:00 p.m.–5:00 a.m.), the model uses just the first four hourly measurements to recursively forecast DO until dawn, achieving an overall mean absolute error (MAE) of 0.3593 mg/L and root mean square error (RMSE) of 0.4813 mg/L. This represents improvements of 35.5% (MAE) and 32.4% (RMSE) over conventional models. By combining physical insights with data efficiency, the proposed approach offers a practical and scalable solution for reliable DO prediction in small‐scale aquaculture systems worldwide.more » « less
