Uncertainty in the blood donation appointment scheduling: Key factors and research perspectives

We consider the management of a blood collection center, which includes the features of both a production system and a service provider. In particular, we analyze the scheduling of donors and the related appointment system, addressing the so-called Blood Donation Appointment Scheduling (BDAS) problem. From the production system viewpoint, the requirement is to balance the production of whole blood units between days in order to meet the requirement of a constant supply of blood to hospitals and transfusion centers; from the service provider perspective, appointments reduce waiting times and improve the service quality perceived by donors. Thus, the goals of the BDAS are to guarantee a quite constant production of whole blood units and to reduce physicians' overtimes while including appointments and free slots for donors without a reservation. A framework for the BDAS problem has been recently proposed, in which slots are first preallocated to the different blood types and then assigned to the donors when they call to make a reservation. However, this framework refers to a deterministic setting in which all input parameters are assumed to be known in advance. On the contrary, the BDAS problem is stochastic in nature and includes stochastic parameters that must be predicted from historical data. In this paper, we first analyze the possible uncertainty sources to determine the most critical ones. Then, we propose research directions to properly include them in the BDAS framework, considering both stochastic programming and robust optimization methodologies.