A Visual Explorer for Retrospective Trajectories in Children Following Heart Surgery.

Children after cardiac surgery with cardiopulmonary bypass are susceptible to a variety of complications and poor outcomes, e.g., acute kidney injury or infection. Postoperative monitoring in children is a complex process, characterized by frequent interventions, age-varying reference ranges, different data sources, among other factors. Visual analysis of retrospective data could offer significant clinical value by improving understanding of trajectories and potential factors for complications. Therefore, we aim to create an interactive visualization tool to assist in exploration of retrospective postoperative trajectories of children following heart surgery. We present the prototype TRACE (Trajectory(-based) Retrospective Analysis for Clinical Exploration) that was developed in close collaboration with clinicians. Feedback was collected using a questionnaire with qualitative and quantitative items, designed to assess usability, utility in clinical, research and teaching settings, and effectiveness of the proposed visualizations. The tool achieved a mean system usability scale score of 82.5. Future work will focus on releasing TRACE to researchers with access to the ELISE dataset, a comprehensive, open pediatric ICU dataset, facilitating data understanding and analysis of trajectories in critically ill children following cardiopulmonary bypass.