Essays on blood supply management

• Main Author: Senga, Juan Ramon Liboro
• Other Authors: Geoffrey Chua
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Business::Operations management::Inventory control
• Online Access: https://hdl.handle.net/10356/153659
• Institution: Nanyang Technological University

There is a global shortage of blood with countries signalling the need to improve collection in order to address blood requirements. This dissertation focuses on three essays related to blood supply and blood donor management with the aim of addressing this global shortage. We do this by providing insights and tools into better blood inventory management, donor recruitment and retention, and disaster-related policy for blood banks. One of the defining features of blood donation is a fixed rest time between donations for donors. This effectively restricts the ability of willing donors to donate more often than they want to. From a blood bank's perspective, this also reduces the number of times they can ask a donor to donate. We call this process “regeneration” with blood donors being an example of regenerative capacity. In the first essay, regenerative capacity is analyzed, advancing blood inventory management and generalizing capacitated inventory systems. The essay considers both backlogging and lost-sales with an initial assumption that no donors drop-off from the donor pool. We prove the optimality of a threshold policy in both cases. We further characterize a rationing and donor-build up policy in the lost-sales setting if few donors are available because of a resource-conserving and capacity building mind-set from blood banks. This is then extended to a more realistic scenario where donors are allowed to drop-off. We then provide a numerical study that looks at the effect of reducing regeneration on the cost incurred. We find that because blood is considered a valuable resource, reducing regeneration can significantly reduce cost. The blood shortage problem becomes even more apparent when there are disasters. In this two-part second essay, we examine blood supply interventions and the dynamics of blood supply during disasters. For the first part of the second essay, we construct an analytical model which looks at the equivalence of three types of interventions in a steady-state system, namely interventions related to: (1) recruitment and retention, (2) donation frequency, and (3) reducing regeneration time. We also present a methodology for comparing the efficiency of each of the three interventions in reaching a desired target donation rate. Next, we use the steady-state results to calculate the more realistic non-steady state scenario through an iterative calculation procedure. We find that the equivalences between the interventions continue to hold even in the non-steady state. A proposed dynamic donation rate strategy that controls the donation rates is shown to be able to overcome donation volatility. In part two of the second essay, we gain further insights from the analytical model by constructing a stochastic inventory model with regenerative capacity, followed by a discrete-event simulation study. The results show that more mature blood banks have more strategies available at their disposal when it comes to trying to maintain fill-rates. We also observe that the dynamic donation rate strategy results in lower fill-rates during disasters when the fill-rate is low despite being able to meet donation targets, and that multiple waves with short lead times exacerbate this effect. Lastly, we introduce three ethically complex interventions: (1) demand deferral, (2) rationing of blood, and (3) tapping non-traditional donor pools and show that these can be implemented from both the analytical and simulation models. In the third essay, we address the trade-off between recruitment and retention of blood donors. As blood banks try to increase donation by using marketing campaigns to target groups of donors, budget constraints leave them with a choice of either recruiting new donors or retaining old ones. We use a multi-period network flow model and an equivalent mixed integer linear program to solve the problem given four sample countries. We find that if the objective function is to maximize number of donations, blood banks would focus on retaining those who have already donated rather than getting new donors. However, this leaves the blood bank vulnerable to having an older donor pool. We then introduce a hybrid objective function that balances maximizing new donors with maximizing the number of donations. A numerical study confirms that the hybrid objective increases new donors significantly while reducing the number of donations by a relatively smaller amount. The final analysis compares the strategies of short-term and long-term planning horizons, such that long-term planning horizons focus on recruitment of new donors at the start and retention of more regular donors towards the end.