Abstract WP085: Development of an integrative platform for stroke care systems

Introduction: Establishing an efficient network among facilities addresses the unavoidable imbalance between medical needs and stroke facilities. To achieve this goal, assessing the current facility's capacity, defining its roles in the stroke care system, and monitoring network performance are of higher priority. A targeted platform is developed to facilitate these objectives. Methods: The data were extracted from national quality assessment and audit data, which collected qualifying and treatment information of stroke care hospitals in Korea. They reconstructed the hospital-level data format, including the hospital capacity and treatment volumes, such as recanalization therapy, patient volume, and transfers. Given the unfixed care system without certified hubs, K-means and core expansion network algorithms are utilized to identify the hubs. These processes delineated the roles of hubs in terms of performance and network dynamics. Subsequently, hospital groups with strong transfer connections were clustered heuristically into Hub-and-Spoke models, allowing for a quantitative analysis of their overall performance. The platform visualizes them as diagrams and interactive maps to present the network structures, practices, and degrees of connection. Results: In 2018, it is estimated that on average, 3400 ischemic stroke patients are admitted nationwide and undergo 300 transfers per month, with 292 and 272 intravenous thrombolysis and endovascular treatment performed, respectively. They comprised 158 hubs and 884 spokes on the hospital level, forming 88 stroke clusters in South Korea (Figure 1). They classified eight disparate Hub-and-Spoke models regarding cluster network structures, presenting the various numbers of hubs and centralizing patterns (Figure 2). This platform displayed the location of facilities within each cluster and degrees of interconnectivity throughout the nation. It visualized the performance and regional locations according to the user's interest and purpose. Conclusion: With the network-based platform, the stroke care system can be explored effectively for multipurpose use.