Session 7: Health Care OM

Parallel Sessions

Session 7: Health Care OM – Chair: Erwin Hans

René Bekker (Vrije Universiteit Amsterdam) – Access times in appointment-driven systems and level-dependent queues
Motivated by health care applications, we study access times in appointment-driven systems. The access time is the number of days between a request for an appointment and the day that the appointment can take place. To meet target access times, we allow for overbookings, as they often occur in practice. We argue that such a system can naturally be modeled as an MAP/G/1 queue; the level corresponds to the access time and the phase to the dynamics of the number of free slots at the first available day. To allow for overbookings, we analyze a level-dependent version of the MAP/G/1 queue, leading to intuitively appealing results. The model sheds light on appropriate occupancy targets for e.g. outpatient departments.

Gréanne Leeftink (University of Twente) – Scheduling multi-disciplinary cancer clinics
Many hospitals start multidisciplinary clinics to assure timely care for their cancer patients. Scheduling appointments in a multi-disciplinary clinic is complex, since coordination between disciplines is required. The design of a blueprint schedule for a multi-disciplinary clinic with open access requirements requires an integrated optimization approach, in which all appointment schedules are jointly optimized. This research is motivated by a Dutch hospital, which uses a multi-disciplinary cancer clinic to communicate the diagnosis and to explain the treatment plan to their patients. Furthermore, also regular patients are seen by the clinicians. All involved clinicians therefore require a blueprint schedule, in which multiple patient types can be scheduled. We design these blueprint schedules by optimizing the patient waiting time, clinician idle time, and clinician overtime. As scheduling decisions at multiple time intervals are involved, and patient routing is stochastic, we model this system as a stochastic integer program. The stochastic integer program is solved with a sample average approximation approach. Numerical experiments evaluate the performance of the sample average approximation approach. We test the suitability of the approach for the hospital’s problem at hand, compare our results with the current hospital schedules, and present the associated savings. Using this approach, robust blueprint schedules can be found for a multi-disciplinary clinic of the Dutch hospital.

Nico Dellaert (Eindhoven University of Technology) – Optimizing chemotherapy appointment-templates between minimum flowtime and minimum makespan
We study the problem of scheduling outpatient chemotherapy infusion appointments at oncology clinics. The patients are prepared during a fixed initial period of their infusion appointments. During the remainder of the appointments, the patients are mainly monitored by the nurses and if needed taken care of. For preparing a patient and setting up the infusion device, one nurse must be fully assigned to the patient. Nurses who are not on a break and not busy with preparing patients, simultaneously monitor up to a certain number of patients who are already receiving infusion. The infusion duration can significantly differ from patient to patient. The objective is to have appointments start as early as possible and the ending appointment(s) to be completed as early as possible within the work-shift. We solve this problem using integer programming. By adjusting two parameters in the objective function, the solution can be tuned between flowtime minimization and makespan minimization while prioritizing selected appointments for starting as close as possible to their ready-times. Our numerical results show that the model can be solved to optimality with short computation times for large real-world size instances using commercial solver software. This method can be used both for generating appointment-templates in online scheduling and for offline scheduling when all patients of the intended day are already determined.