2022 Optimization Days
HEC Montréal, Québec, Canada, 16 — 18 May 2022
MA6 - Scheduling in healthcare
May 16, 2022 10:30 AM – 12:10 PM
Location: Raymond Chabot Grant Thornton (yellow)
Chaired by Dina Ben Tayeb
10:30 AM - 10:50 AM
Simulation of patient flow considering bed management in a university hospital
This project aims to analyze the flow of patients between the emergency services, the surgical block, and the hospitalization units of a university hospital through discrete event simulation. The main objective of the work is to generate an overview of the use of beds at the partner university hospital about the surgical block and the flow of the emergency department to explain how these resources interact with each other and how we can take them into account to improve the use and prediction of bed requirements. The simulation model generated is intended to serve as a basis for a reliable forecast for the next year on bed utilization, planning the surgical block schedule to meet seasonal demand, and studying length of stay and waiting times in the emergency units, among others. To meet the objectives, some of the actions carried out have been: the collection, processing, and analysis of sensitive data belonging to actual patients at the hospital; or the development of the model following a rigorous established methodology (the health system studied has been modeled using queuing theory), among others.
10:50 AM - 11:10 AM
Optimization of staff schedule and assignment of tasks: a case study in a Quebec hospital setting
Although nurse-rostering problem is a classical problem, real-life applications still raise challenges. In this research, overlapping and multiple lengths shifts type, hierarchical personnel divisions, work patterns and numerous contract types which vary in minimum and maximum working hours in a given period are particularly targeted. We have implemented the models using Julia and user friendly prototypes are to be transferred to the partners. In order to ensure a robust and flexible schedule, we investigate how an extra virtual shift can be added to secure a replacement in case of absenteeism, which is a recurring event in many working environments. Additionally, further analysis is conducted to carefully select appropriate weights for soft constraints in the objective function, which allows a more generalized and adaptable model. This model has even been applied to two case studies, each with their own particularities.
11:10 AM - 11:30 AM
Integration of length of stay in Master Surgery Scheduling
We seek to generate an optimal master surgery schedule for a university hospital while considering patient length of stay in the wards and the ICU. The goal is to avoid surgery cancellations due to a lack of downstream resources and to limit the need to move patients between the wards. Our approach of the length of stay is done at a granularity level that involves the diagnostics of patients, which are grouped by clustering algorithm under each specialty. To do so, we were provided with a year of data to exploit. The Master Surgery Scheduling is thus done on a long term horizon (one month) with a deterministic, linear programming model which seeks to both maximize bed utilization and room utilization while giving priority to patients based on their wait time. We are then able to provide an efficient MSS and also recommendations regarding elective patient scheduling. Emergency patients are treated through the allocation of slack capacity.
11:30 AM - 11:50 AM
Location-Allocation of Caregivers with Fuzzy Districting in the Home Healthcare Industry
We consider a caregiver location-allocation problem in the home healthcare sector for a given territory. The territory consists of several basic units, and patients are located in these basic units. Each basic unit has demands in terms of the number of patients, and caregivers are associated with basic units. The purpose of this study is to simultaneously investigate the location-allocation of caregivers and districting of the given territory with fuzzy boundaries. Each district is made up of multiple basic units, and in the case of fuzzy district boundaries, a basic unit may belong to more than one district if the basic unit is served by more than one caregiver from different districts. We further investigate the benefit of grouping caregivers into healthcare units. Caregivers are grouped according to the basic unit to which they are attached. We also extend the problem by considering the stochastic demands of basic units. The work satisfaction level of each caregiver is considered to be a function of the distance or travel time between the caregiver and the basic unit to which the caregiver is assigned. The objective of the problem is to maximize the sum of the work satisfaction levels of all caregivers.
11:50 AM - 12:10 PM
On integrating patients appointment grids and technologist schedules in a radiology center
In this work, we study the simultaneous optimization of appointment grid and technologist schedules in a radiology center. We develop a mixed integer programming model that provides an optimal allocation of personal resources to maximize the machine utilization and the number of treated patients. We evaluate the optimization model using a real case of the CHUM radiology department. We study different scenarios by testing several technologist rules and planning construction methods.