System-of-Systems Approach for Improving Quality of Kidney Transplant Decision-Making Support for Transplant Surgeons

An AI-powered decision-making tool is being developed to aid transplant surgeons in evaluating an existing deep-learning model. The goal is to improve the quality of decision-making support by incorporating individual surgeon practices. The kidney acceptance decision-making process is complex, requiring a transdisciplinary approach. Identifying kidneys that are hard to place can further complicate the evaluation process, which is why an AI-enabled decision-making tool is necessary. However, AI machine learning for healthcare decision-making is not yet widely researched, adopted, or accepted. The system-of-systems (SoS) approach aims to capture transplant surgeons’ unique decision-making practices while assessing the level of trust experienced during the process. This tool also offers a way to explore adapting large language model recommendations into healthcare decision-making. The tool objective is to maximize the balance between the following key performance parameters (KPPs) of this model: performance, acceptability, usability, affordability, and robustness. Potential decision-making architectures are generated, the system interfaces are measured, and architectures are assessed using genetic algorithms coupled with the fuzzy inference system (FIS). The results suggest the most desirable or near-optimal SoS meta-architecture that can be used to improve kidney transplant decision-making support for transplant surgeons, with the goal of creating a successful human-AI system interaction model.