Sarah Kirchner
Appointment scheduling in hospitals - sequencing and scheduling using time aggregation
VCLA hosted a talk by Sarah Kirchner on August 4, 2015.
DATE: | Tuesday, August 4, 2015 |
TIME: | 10:30 |
VENUE: | Menger room, Favoritenstrasse 11, 3rd floor |
ABSTRACT
In Germany as well as in many other countries, hospital services provided for admitted patients are settled using diagnosis-related groups (DRGs). That is, patients are grouped according to their diagnosis, received services and demographic characteristics into a DRG. The hospital receives a fixed reimbursement dependent on this DRG. This payment scheme provides incentives for hospitals to aim for a short hospitalization of patients. Almost all patients receive more than one medical service during their hospital stay and there may be dependencies between these services. To facilitate the requirements of these multiple dependent services, coordinated appointment calendars for the different resources of a hospital are needed. At the moment, it is common practice that medical staff at a resource sequentially assigns appointment times to incoming requests disregarding all other services the patient may need and often without considering the impact the decision has on the length of the patients stay. Additional to the admitted patients, many hospital units also provide ambulatory services to patients. Often they also need to consider walk-in patients that arrive at the hospital without prior notice. We assume that accurate information about the exact execution time for a request is only needed on the day of execution itself. Before that, it is sufficient to know the day (or week for requests in the far future) the request is scheduled to be processed. A time-indexed IP formulation for the problem is proposed in which timeslots are aggregated for the time after the next day. To cope with the requirements of walk-in patients additional heuristics are proposed. Based on real data from the university hospital in Aachen, test data sets are generated and the solution approach is tested.