A carregar...
Publicação
From Data to Decision: Data Visualization to Support Clinical Decision-Making in Acute Myocardial Infarction
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 2.49 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
ABSTRACT - Accurate mortality prediction in acute myocardial infarction (AMI) is essential for guiding clinical decision-making and improving patient outcomes. Machine learning (ML) models show strong predictive performance, but their use in practice is constrained by interpretability. This study explores how interactive data visualization can improve the usability and understanding of datasets that feed into ML-based AMI mortality predictions, making them easier for clinicians to interpret. An interactive web-based dashboard was developed in D3.js, allowing dynamic exploration of patient data and designed for future integration of model-based risk estimates. The study followed a design study methodology, with iterative input from cardiologists to identify needs, refine the interface, and assess its value for decision-making. The dataset used was drawn from previous work and included demographic, clinical, and laboratory variables relevant to AMI prognosis. Evaluation focused on two dimensions: design principles, examined through heuristic analysis by experts in data visualization, and clinical utility, assessed by cardiologists to determine its clinical applicability. Results showed high ratings for ease of use and facilitating conditions, with moderate perceived usefulness and intention to use. Lower scores for habit and social influence were expected, as the tool has not yet been integrated into routine practice. These findings indicate that the dashboard is usable and potentially valuable in clinical settings, providing a foundation for future integration of predictive models. The heuristic evaluation additionally highlighted opportunities for design refinement, informing directions for future iterations of the tool.