Event (Seminar) – Delta coverage: The analytics journey to implement a novel nurse deployment solution (Apr 21)

The Centre for Healthcare Engineering (CHE), Centre for Analytics and Artificial Intelligence Engineering (CARTE) and Operations Research (OR) Seminar Series welcome:

Jonathan E. Helm
Kelley School of Business, Indiana University

Date: Friday, April 21, 11:00 am – noon
Venue: GB220, Galbraith Building, 35 St George St, Toronto

Announcement with speaker Bio and map at: CARTE Events (also calendar tool); MIE Events

No registration required

Abstract: In January 2021, we embarked on a journey in partnership with Indiana University (IU) Health System, the largest health system in Indiana with 16 hospitals, to jointly develop a suite of advanced data and decision analytics to support a new internal travel nursing program at IU Health. This program was designed to leverage a flexible pool of resource nurses that can be moved between the 16 system hospitals. In contrast to traditional travel nursing programs designed for longer-term nursing shortages (e.g., 12 week contract during flu season), our program focuses on short term (days instead of weeks) deployments to respond to geographic and temporal fluctuations in hospital occupancies. The key challenge with a dynamic travel program is that nurses must be put “on-call” to travel 1-2 weeks in advance and notified of being “called-in” 24-24 hours in advance. This requires accurate forecasts of nurse demand across the 16 hospitals and the ability to make extremely complex decisions with trillions of possibility regarding which nurses to send to which hospitals on each day of a 30-day planning period. To support this program, we developed a machine learning-based occupancy forecasting model that accounts for different levels of patient acuity. Using distributional information from the forecast, we generate workload scenarios for the network that are fed into a two-stage stochastic program where the decision variables mimic the timing and type of decisions being made in practice. The decision support tool was implemented in October 2020 as a Microsoft Power BI application. We logged the performance of the recommendations from October 2021 to March of 2022 as a proof of value, with the tool running each day in real time. Analysis indicates system-wide improvements in all metrics: with reductions of 5% understaffing, 3% misallocation of resource nurses, and 1% overstaffing. The annualized savings estimated at over $400K. This talk focuses on the challenges of developing technical research within the real-world constraints of implementation and integration into hospital data warehouse and staffing systems as well as the challenges of an academia-industry collaboration for implementing technical work in a massive healthcare organization such as IU Health.