Liangyuan Na, Jean Pauphilet, Ali Haddad-Sisakht, Louis Raison, Audrey Silver, Patricia Veronneau, Nicole Vogt, Dimitris Bertsimas (2024) Optimization Automates Emergency Department Nurse Scheduling at Hartford Hospital. INFORMS Journal on Applied Analytics.
To optimize nurse staffing in the emergency department (ED), Hartford Hospital has been collaborating with academics and consultants to schedule nurse shifts over each six-week staffing cycle. We develop and implement two-phase optimization models: a robust optimization model to find optimal staffing levels given the uncertainty in patient demands, followed by a pair of mixed-integer problems to generate individual schedules including work, trainee, and preceptor shifts for each nurse. Our approach leads to less costly (5%–8%) staffing with better coverage of patient care (8%–25%) and higher nurse satisfaction (5%). Moreover, nurses can work fewer shifts on weekends (17%), holidays (14%), and overtime (85%), as well as be assigned to more diverse positions (3.6) and more daily training opportunities (0.95). We implement our framework into an automated end-to-end scheduling optimization software, deployed for use at Hartford Hospital since March 2023. The software collects preferences from more than 200 ED nurses and enables managers to optimize schedules with guided dynamic adjustments. This transformative implementation streamlines a previously labor-expensive staffing process (currently taking more than 88 manual hours per cycle) and delivers schedules that are more suitable for patients and nurses together, with an annual projected cost saving of around $720,000.
沒有留言:
張貼留言