University of Twente Student Theses
Pediatric Respiratory Exacerbation Detection : Innovating Asthma Care with Artificial Intelligence (PREDICTA) : An AI Model to predict pediatric asthma exacerbations and personalized risk factors
Ruuls, Tamara (2024) Pediatric Respiratory Exacerbation Detection : Innovating Asthma Care with Artificial Intelligence (PREDICTA) : An AI Model to predict pediatric asthma exacerbations and personalized risk factors.
|
PDF
6MB |
| Abstract: | Pediatric asthma significantly impacts children's quality of life, with uncontrolled cases leading to unpredictable exacerbations driven by various demographic, clinical, and environmental factors. Given the complexity of these interactions, machine learning (ML) could play a vital role in identifying patterns for predicting asthma exacerbations. However, existing models lack generalizability and clinical utility. The PREDICTA study addresses these gaps by developing ML models, specifically LSTM and XGBoost, for predicting exacerbations within routine pediatric asthma care. Using 3.5 years of electronic health data from MST, the study defined severe exacerbations as requiring hospitalization and certain medications, while moderate cases involved the same medications without hospitalization. Despite limitations due to class imbalance (129 exacerbations in 2.3 million data points), the LSTM model showed promise for personalized, time-dependent predictions, whereas XGBoost faced challenges with individualized predictions. Interviews with pediatricians underscored the importance of transparency, risk factor identification, and decision-making support. The study proposes three applications: a personal risk dashboard for patients, a risk dashboard for clinicians, and an eHealth tool for at-risk patients. This work highlights the need for standardized exacerbation definitions and further model improvements, paving the way for more personalized, effective asthma management. |
| Item Type: | Essay (Master) |
| Faculty: | TNW: Science and Technology |
| Subject: | 44 medicine, 50 technical science in general, 54 computer science |
| Programme: | Technical Medicine MSc (60033) |
| Link to this item: | https://purl.utwente.nl/essays/104526 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page