Predicting Asthma Using Clinical Indexes

Area of Science:

• Pediatric Pulmonology
• Clinical Prediction Modeling
• Respiratory Medicine

Background:

• Asthma is a heterogeneous condition, not a single disease, with varying phenotypes in preschoolers.
• Not all preschool children with recurrent wheezing develop asthma, necessitating predictive tools.
• Current diagnostic methods lack accurate genetic or biochemical markers for early asthma prediction.

Purpose of the Study:

• To review and summarize frequently used asthma predictive models for preschool children.
• To discuss the advantages and disadvantages of these predictive models.
• To evaluate the performance of models in predicting childhood asthma development.

Main Methods:

• Systematic review of seven popular asthma predictive models.
• Models analyzed include original API, Isle of Wight, PIAMA, modified API, ucAPI, APT Leicestersher, and ademAPI.
• Evaluation focused on clinical parameters and predictive accuracy (LR values).

Main Results:

• The original Asthma Predictive Index (API) demonstrated a positive Likelihood Ratio (LR) of approximately 7.4.
• The API is simple, requiring only four clinical parameters and eosinophil count from blood samples.
• The API's negative LR is suboptimal, limiting its utility in ruling out asthma development.

Conclusions:

• The original API is a practical and effective tool for predicting asthma in preschool children, usable in diverse healthcare settings.
• While useful for identifying at-risk children, the API cannot definitively exclude the future development of asthma.
• Further research into more accurate biomarkers and refined predictive models is warranted for early asthma diagnosis.