When you think of nurses, you think of people who take care of individuals who are ailing. But, when the nurse scheduling procedures at the University of Michigan were ailing they did not call a doctor. To improve the arduous scheduling decision process, the schedulers at the University of Michigan called on a interactive computer scheduler which used mathematical programming.

To revive the nurse scheduling process, the computer scheduler incorporated a set of policy decisions that effected both the nurses and the hospital. The process posed the scheduling decision as a large multiple-choice programming problem, which addressed the concerns of the nurses as well as the hospital.

The program allowed nurses to make their preferences about, days off, shifts, rotations, and work stretches etc. known. Nurses were given "Individual Aversion" forms to complete. Individual data was stored and updated as needed. The completed schedules adhered to nurses wishes as regularly as possible. However, when conflicts arose, the program decided in favor of the hospital.

The hospital had its own set of concerns to be considered: (a) having the minimum number of nurses needed for each shift on each day, (b) minimizing overtime, (c) preserving quality care and morale, and (d) decreasing turnover and absenteeism. In the past, the scheduling methods were not as understanding of the needs of both parties.

There were two methods previously used by the nurse schedulers. The first method was based on a trial and error approach. This approach was time consuming and frustrating; however, it allowed the nurses some flexibility. The second method, was based on a cyclic approach which utilized a rotating schedule. This approach was not flexible; however, it provided stability for the hospital.

The new method eliminated the drawbacks felt by the old methods. It decreased the amount of time need to make a six week schedule by 18-24 hours. It was less frustrating for the nurses and the hospital because their feelings and needs were taken into consideration before the schedules were generated. In addition, it allowed room for flexibility because the changing desires of the nurses were added to the system before upcoming scheduling periods.

The new nurse scheduling system was well received by the nursing administrators and the nurses at the University of Michigan. Members of the nursing staff agreed that new system had improved the quality of the schedules generated.