Using medication utilization information to develop an asthma severity classification model.

Area of Science:

• Health Informatics
• Pharmacoeconomics
• Respiratory Medicine

Background:

• Claims data are a common source for asthma research but lack detailed severity information.
• This limitation can negatively impact the accuracy and validity of study findings.
• A novel asthma severity classification model using medication data is needed.

Purpose of the Study:

• To develop and validate an asthma severity classification model.
• To utilize medication utilization patterns within Taiwan National Health Insurance claims data.
• To enhance the quality of asthma research based on claims data.

Main Methods:

• Utilized Taiwan National Health Insurance Research Database (2006-2012).
• Developed and validated a model based on medication utilization in the first year post-diagnosis.
• Assessed asthma exacerbations in subsequent years and compared model performance using AUC and Kaplan-Meier survival analysis.

Main Results:

• The developed model classified patients into severity groups based on medication use, with distinct patterns for oral corticosteroids, short-acting medications, and inhaled corticosteroids.
• Survival analysis indicated a two-group classification model yielded moderate performance (AUC = 0.671).
• Model performance was consistent in both development and validation datasets.

Conclusions:

• Accurate assessment of asthma severity is crucial for effective research.
• The developed medication utilization-based model addresses a gap in using claims data for asthma studies.
• This model offers a valuable tool to improve the validity and quality of claims-data-based asthma research.