부산대학교 도서관

The article discusses the development of prediction models for childhood asthma based on population-based family history health records. The Predictive Asthma Risk Score (PARS) and the Asthma Predictive Index (API) are commonly used tools, but they have limitations in terms of sensitivity and specificity. The authors propose a new prediction model that incorporates the child's and parental history of comorbidities, using machine learning techniques. The study found that parental asthma, a child's allergic comorbidities, and the child's history of respiratory infections are the best predictors of pediatric asthma. The article emphasizes the importance of considering a holistic picture of a child's health status in improving the accuracy of asthma prediction algorithms. [Extracted from the article]