HCM
Health Care Management I

Invited Session

Time Slot: Tuesday Afternoon

Room: 002

Chair: Giuliana Carello

The value of exploiting information in Logic-Based Benders Decomposition for chronic outpatients scheduling by Answer Set Programming

Time: 14:10

Maddalena Nonato (DE, Università di Ferrara, Italy), Cappanera Paola, Gavanelli Marco , Roma Marco

Chronic patients living at home often need to take tests or receive treatments (health services in general) at hospital premises in a periodic way, as prescribed by well-assessed clinical pathways. In addition, due to increased life expectancy, the number of elderly patients having several Non-transmissible Chronic Diseases (NCDs) at the same time (co-morbidities) is steadily growing; these patients must follow several clinical pathways at the same time, thus making health service synchronization of paramount importance to ensure system efficiency and patient’s satisfaction. A further source of complexity is due to timing conflicts potentially arising between health services: for example, a treatment with a certain drug may invalidate the results of a medical examination if the two health services are done too close in time. Therefore, an appointment schedule should obey several time constraints, such as predetermined service frequencies, beside complying with patient needs. Moreover, resource availability, regarding both personnel and devices at hospital care units, is limited, thus reducing the options for booking. This criticality has been further exacerbated by the current pandemic situation in which, at hospitals, resources have been devoted mostly to patients with high intensity care needs. Inevitably, this has had negative effects on the management of chronic patients: on the one hand, many screening or monitoring activities have been delayed, and in many cases cancelled; on the other hand, many patients have preferred to avoid attending hospitals for fear of becoming infected. For these reasons, it is estimated that the service demand for chronic patients will grow significantly in the near future. This suggests managing the scheduling of the clinical pathways of enrolled patients in a centralized manner. From the patient point of view, synchronizing different services reduces the number of accesses to the hospital or even avoids potential duplications that may arise when the same health service is present in different protocols.In this work, we provide a declarative approach to this challenging problem, encoding it in Answer Set Programming (ASP). Computational results show that when all of the constraints are given in one shot, the resulting monolithic approach soon fails to provide a feasible solution as the dimension of the problem grows. In order to improve the scalability of the monolithic approach, we propose a decomposition of the problem inspired by Logic-Based Benders Decomposition (LBBD), a technique that extends Benders decomposition principles to other formalisms, such as Constraint Programming. According to LBBD, a master problem and a set of independent sub-problems are iteratively solved until all subproblems admit a feasible solution. Master problem and subproblems efficiently communicate and exchange information to reduce the search space. The propagation of information from one subproblem to others is of key importance in the design of an efficient system. The computational experience will thus purse two targets: first, we will show the impact on efficiency of using decomposition instead of the monolithic approach; second, we will compare several ways of propagating information. The decomposition algorithms as well as the monolithic approach will be tested on realistic instances.

An approach to healthcare services decision-making using fuzzy ELECTRE III. An outpatient chemotherapy service application.

Time: 14:30

Christine Di Martinelly (IESEG School of Management, Univ. Lille, CNRS, UMR 9221 – LEM – Lille Economie Management, F-59000 Lille, France ), Duenas Alejandra, Glaize Annabelle, Fagnot Isabelle

Each day, healthcare institutions face the challenges of maintaining high-quality care and patient satisfaction while dealing with budget restrictions and a constantly increasing patient load. Oncology hospitals share the same struggle. An outpatient chemotherapy service is a complex care system that includes a multi-step process, complex procedures and interdisciplinary caregivers. Bringing changes and improvements need to be carefully considered and accepted by all stakeholders. In such a context, we propose a decision-making approach based on business process improvement (BPI) and multi-criteria decision making to help an oncology department in their reorganization. BPI is used as a tool to structure the decision-aid process, and to identify the stakeholders, the improvement solutions and the criteria to assess them. As there is some ambiguity in the stakeholders’ consensus on the improvement solutions, a fuzzy-ELECTRE III method is used. An online questionnaire is given to ten stakeholders to identify the relative importance of the improvement solutions over the selected decision criteria. Their preferences are assessed and the concordance, discordance and credibility matrices are built. We have decided to add the weights of stakeholders’ characteristics to the decision matrix to include the differences of perspectives and experiences. A sensitivity analysis is finally performed to study the impact of a change in the method’s parameters. Results provide the priorities and the least preferred options for the outpatient services.

A framework for at-home blood collection considering donor preferences

Time: 14:50

Martina Doneda (Politecnico di Milano / IMATI CNR), Yalçındağ Semih, Lanzarone Ettore

In all health care systems, Blood Donation Supply Chains (BDSCs) play a fundamental role in several sectors, including hematology, trauma management and surgery. According to recent statistics (2019), the production of blood units for transfusion still falls short of satisfying the ever-increasing demand. This is due to a combination of factors, the main one being that blood can only be produced from human donors. Meeting demand has always been the main objective of the system, and several decision tools have been developed during the years to help reach it. However, during the COVID-19 pandemic, new and unexpected issues emerged. As blood drives were cancelled because of lockdowns and donors feared going to donations centers, shortages of units have increased dramatically. This is due to the lack of a direct channel between donors and blood banks. According to the Italian National Blood Center, in April 2020 the production of red blood cells was 36.4% less than the one of the previous year. At the same time, healthcare services are now moving towards a more decentralized paradigm. Considering these factors, in this work we consider an integrated framework to collect blood directly at the donors’ homes. To do so, we link the Blood Donation Appointment Scheduling (BDAS) and the Multi-Trip Vehicle Routing Problem with Time Windows (MTVRP-TW) problems in an organic way. Our approach consists of three planning phases: in the first one, we create possible time slots for donor appointments, with the aim of balancing the production of the different blood types, to ensure better manageable stock levels. Then, we turn these slots into real appointments, according to the received booking requests. Here, we compromise between the preferences expressed by calling donors and the requirements of the center, including scheduling and routing constraints. Lastly, a fleet of bloodmobiles is dispatched to perform the collection. The second and third phases are linked in such a way that the outcome of the former serves the purpose of helping the computation of the latter, providing an improved starting point.The proposed integrated framework is then tested and
validated against data obtained from the most prominent Italian blood provider. We also prove the cost-effectiveness of the implementation of the proposed solution.

Patient Appointment Scheduling for Chemotherapy treatments

Time: 15:10

Giuliana Carello (Politecnico di Milano), Passacantando Mauro, Tanfani Elena

The number of patients affected by cancer is expected to increase in the next years and, thus, the need for chemotherapy treatments will increase accordingly. We assume that chemotherapy treatments are delivered in an outpatient setting. Further we consider a cancer centre shared by different specialties where patients affected by different pathologies are treated. Managing the activities of shared cancer centres involves different decisions hierarchically linked (Lamè et al. 2016). First, at a strategic level, the capacity planning of the main resources involved must be determined. Then, at a tactical planning level, the available resources (such as beds, seats, nurses, clinicians, rooms for visit) must be allocated to the clinical specialties. At this stage the cyclic timetable determines the days when the different cancer groups are treated, referred to as Master Chemotherapy Planning (Carello et al. 2022). The clinician rostering over a mid-term planning horizon that covers the cyclic schedule must be determined, as well. Finally, at an operational level, the assignment and sequencing of patients needing chemotherapy must be determined for each day of a given planning horizon (Garaix et al., 2020). In this work, we focus on the operational level decision. We assume that the resources have been assigned to the pathologies, and that in each day the assignment of rooms to pathologies and clinicians is given. The set of days of the week in which patients can be treated, depending on their pathology, is therefore known as the solution of by ad hoc optimization models (Carello et al., 2022). The obtained weekly pattern repeats over a month. In this work, we consider the problem of assigning patients to days in the considered planning horizon. Further, we solve the appointment scheduling problem for each day. The problem can be seen as a multi-appointment scheduling problem (Marynissen and Demeulemeester, 2020) where we have to decide the starting time of each activity involved in the day hospital (in particular, oncologist visit and infusion delivery) for each patient. We consider different objectives using a patient centered approach aimed at maximizing the quality of the patient experience. We formulated the problem as a multi-objective optimization model and we tackle the problem by sequentially solving three ILP models, in a lexicographic multi-objective fashion. The models and solution procedure are tested on real data from an Italian hospital.

[1] Marynissen J., Demeulemeester E. (2020). Literature review on multi-appointment scheduling problems in hospital settings. European Journal of Operational Research, 272 (2), 407-419.
[2] Carello, G., Landa, P., Tànfani, E., Testi, A. (2022). Master chemotherapy planning and clinicians rostering in a hospital outpatient cancer centre. Central European Journal of Operations Research, 30, 159-187.
[3] Lamè G., Jouini O., Stat Le Cardinal J. (2016). Outpatient chemotherapy planning: A literature review with insight from a case study. IIE Transactions on Healthcare System Engineering, 6(3), 127-139.
[4] Garaix T., Rostami T., Xie X. (2020). Daily outpatient chemotherapy appointment scheduling with random deferrals. Flexible Services and Manufacturing Journal, 32(1), 129-153.