10:30 AM - 11:00 AM
Surgical Scheduling to Smooth Demand for Resources
With growing demand for healthcare resources, pressure on efficient usage of available bed capacity is increasing. Peaks in bed demand are due to variability in admissions and lengths-of-stay. With a balanced schedule in elective surgeries, peak traffic is levelled across the week, hence, reducing overcrowding without turning away any patients or increasing bed capacity. This study presents a two phase approach for scheduling the surgeries which has been implemented in Hamilton Health Sciences in Ontario, Canada.
In the first phase, a Monte Carlo simulation type model is applied to estimate the patient demand for beds in the hospital during a typical week. The second phase involves running an integer programming model to minimize the required beds for elective inpatients admitted for surgery to the hospital, by changing the day of surgery blocks. This demonstrates the opportunities for smoothing the expected patient demand for beds by adjusting OR schedule. This decision is made at the tactical level. The demands for different surgeries have been estimated based on a two year historical data and the integer programming model has been solved using GAMS/CoinBonmin MINLP Solver. The optimal schedule reduces the demand for beds between 7% on Fridays and 69% on Mondays. The model can be extended to cover the demand for other resources such ICU beds, too.
11:00 AM - 11:30 AM
Is overtime always a relevant performance measure in healthcare?
Overtime is in the inpatient scheduling literature one of the most frequently used performance measurements. This is the case as: (1) it is in reality an important cost factor in hospitals and (2) as a performance measure it easily works together with many operations research methods.
In this research, we are trying to determine whether there are cases when overtime, and other operating room related performance measures, may not be entirely applicable to real inpatient scheduling problems.
This is an important question for both researchers and practitioners working on inpatient scheduling. For researchers it is important to understand whether overtime, given their setting, can be included into the objective of their method. Also for practitioners it is important as they try to identify methods from the literature that apply to their setting.
In order to tackle this question we use: (1) a very detailed simulation model of a real hospital operating room department and (2) the results of an analysis of a real setting.
Using simulation, we found that the patient scheduling methods we tested did not significantly affect overtime. Using data analysis, at this stage of our research, we think that the strongest determinant of overtime is the surgery duration estimation error whereas other relevant factors are emergency arrivals and surgery rescheduling.
Those results suggest that, in an inpatient surgery scheduling setting, caution should be applied whenever overtime is used as an optimization criteria.
11:30 AM - 12:00 PM
Medium term surgery scheduling with patient release dates
This work deals with the Advance Scheduling Problem, which is a key problem in operating theatre planning and in surgery wards managing. The problem consists in determining the set of patients to be operated on and their scheduling over a given planning horizon.
We consider an elective waiting list: each patient in the waiting list is characterized by a waiting time, a surgery duration and an urgency level. Based on his/her urgency level, a maximum waiting time before treatment is computed for each patient (deadline). We consider a medium term planning horizon up to three months. Therefore, we also consider new elective patients who join the waiting list during the planning horizon. New patients are characterized, as those already in the list, by an urgency level, a surgery duration and a deadline. Furthermore, a release date is given, i.e. the date in which the patient is registered in the waiting list. The set of operating blocks available in the planning horizon is given, each characterized by a surgery time capacity.
The aim is to provide the best possible quality of service from the patient point of view. This means that the surgery must be performed before deteriorating patient clinical condition and possibly keeping the waiting time as small as possible. Thus the objective function accounts for waiting time and tardiness of patients.
Different MILP formulations and MILP based heuristics are proposed. All the proposed methods are tested and compared on a set of real-life based instances. Their behavior, with respect to both computational time and quality of the obtained solutions, is evaluated. The impact of planning horizon length and number of blocks is analyzed.