The doctoral thesis of Lauri Neuvonen, DI, “Supporting decision making in complex multiobjective problems: Practical tools and experiences from the healthcare context.” will be publicly examined at the Aalto University School of Business on Friday, February 23, 2024.
In many countries, healthcare organizations face increasing pressure for providing more services while also suffering from lack of ressources, both problems exacerbated by aging populations. This development highlights the importance of resource efficiency. At the same time healthcare decisions often have to take into account multiple, potentially conflicting objectives, such as maximizing health benefits and minimizing resource use, and complex interactions between parts of the system. These overlapping requirements make them a challenging and, on the other hand, an interesting application area for multiobjective optimization tools. Multiobjective optimization is a methodology that helps in finding high performing decision recommendations in situations where several, potentially conflicting objectives are pursued. Recent developments in both computing power and algorithms have made such tools viable in supporting decision making related to healthcare problems of practical scope.
This Dissertation develops multiobjective optimization approaches and explores their use in three practical healthcare decision making problems: i) mitigating the impacts of the COVID-19 epidemic, ii) improving the efficiency of the Finnish colorectal cancer screening program, and iii) designing a hospital network for carrying out hip and knee replacement surgeries. These approaches help accommodate uncertainties affecting the performance of the found solutions. They also accommodate hidden or partial information about the decision-maker’s preferences. The overall focus in the approaches is on modeling the problem setting in high enough accuracy for the solutions to provide practical insights, while also being able to leverage multiobjective optimization techniques in finding the most promising solutions.
The contributions of this Dissertation are two-fold: First, it presents multiobjective optimization approaches, supported by other analytical techniques, that can be used to develop decision recommendations for real-life, complex healthcare decision making problems. These approaches help generate insights that would have been difficult to obtain without the use of model-based tools. A second, more general contribution of the Dissertation is the demonstration of the usability, challenges, and benefits of multiobjective optimization in supporting decision making in the field of healthcare.