MIE Research Seminar
A Dynamic Decision Framework for Multi-Dose Vaccine Admin. in Low-Resource Settings: Balancing Cost, Coverage & Wastage
Content
Speaker: Gizem Nemutlu, Assistant Professor of Data Analytics at the Brandeis School of Business and Economics, Research Affiliate at the Center for Health Technology and Analytics at Massachusetts General Hospital
Abstract
In many low-resource settings, healthcare workers face a recurring tradeoff: open a multi-dose vaccine vial late in the day and risk wasting unused doses, or turn children away to preserve inventory for future sessions. This tension is especially consequential when demand is uncertain, replenishment is infrequent, and multiple vial sizes are available. We study how immunization programs can design the inventory mix across vial sizes and manage vial-opening decisions during clinic sessions.
We develop a finite-horizon Markov decision process that determines which vial size to open and when, if ever, to terminate service as a function of remaining clinic time, days to replenishment, and inventory. Using data for routine pediatric vaccines, we show that strategically combining small and large vials can reduce open-vial wastage by more than 60% and increase expected coverage by up to 20% relative to current practice, without increasing total cost. The optimal policy exhibits intuitive threshold structures that motivate simple time-based switching rules, which retain nearly all of the gains from the dynamic policy. Pareto-frontier analyses further identify cost-effective vial mixes and stock levels that balance coverage, wastage, and cost.
Our results show that even a modest allocation to smaller vials, combined with small adjustments in total stock, can substantially improve vaccine access and dose utilization. The framework provides practical guidance for both clinic-level vial-opening decisions and higher-level inventory and buffer-stock planning.
Bio
Gizem Nemutlu is an Assistant Professor of Data Analytics at the Brandeis School of Business and Economics and a Research Affiliate at the Center for Health Technology and Analytics at Massachusetts General Hospital. Her research focuses on healthcare operations, global health, and the design of data-driven policies to improve health outcomes, including healthcare delivery and resource allocation.
Her work uses decision modeling, simulation, and empirical analysis to study a range of healthcare problems, including vaccine delivery, cancer screening and treatment, and access to healthcare. A central theme of her research is understanding how operational decisions can improve the use of limited healthcare resources, particularly in settings where capacity, supply, and access constraints are significant.
Her current work includes research on vaccine delivery and multi-dose vial management, examining how operational decisions around vial selection and service delivery can reduce vaccine wastage while maintaining access to immunization. She is also an Early-Career Network Researcher with the NCI-funded Translational Liver Cancer Consortium, where she contributes to research on improving the early detection and risk stratification of liver cancer.
Dr. Nemutlu is a Council Member of the INFORMS Health Applications Society and teaches Business Analytics and Operations Management at Brandeis University.
Host: Prof. Muge Capan
We hope you can join us for this exciting seminar! Refreshments and opportunities for questions and discussion will be available post-seminar.