Mid-term nurse rostering considering cross-training effects

Hospitals experience challenging times in which both the economic pressure and the challenges of uncertain demand for care increase. One of the most prominent problems in health care operations is the nurse scheduling problem, where nurse rosters are created to cover demand. Cross-training, i.e. educating nurses to work in units other than their dedicated one, offers an opportunity to react to the issues mentioned above within the field of nurse scheduling. We contribute to the nurse cross-training literature in three ways: First, we propose a framework to define and visualize cross-training policies. Second, we introduce a new cross-training policy where each unit trains one dedicated nurse for each other unit. Third, we are the first who develop a mid-term model creating and applying cross-training policies in nurse rostering. Within this new mid-term model, we make use of parameters that allow to control the trade-off between flexibility of nurses and the continuity of care. In two computational studies with 6400 instances we compare our newly developed cross-training policy with three existing policies from the literature, demonstrate the superiority regarding demand coverage and overtime per number of cross-trainings, and compare the effects of cross-training intensity, i.e. the number of cross-trained nurses, with cross-training breadth, i.e. the number of departments a nurse is cross-trained for.