Real-time automated billing for tobacco treatment: developing and validating a scalable machine learning approach

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare

Background:

• Manual medical billing for tobacco cessation counseling is administratively burdensome and prone to errors.
• Inaccurate coding for services like CPT codes 99406/99407 can lead to revenue loss and hinder patient care.
• There is a need for automated solutions to improve efficiency and accuracy in medical billing.

Purpose of the Study:

• To develop CigStopper, a real-time, automated medical billing prototype.
• To identify eligible tobacco cessation care codes (CPT codes 99406/99407) using machine learning.
• To reduce administrative workload and improve billing accuracy in healthcare.

Main Methods:

• A synthetic corpus of clinical notes was generated using ChatGPT prompt engineering.
• The dataset was annotated by clinicians for CPT code eligibility.
• Multiple machine learning models, including decision trees and random forests, were trained to predict billing code eligibility.

Main Results:

• Decision tree and random forest models demonstrated superior performance.
• The models achieved a mean PRC AUC of 0.857 and an F1 score of 0.835.
• Generalizability testing confirmed the effectiveness of tree-based models on deidentified clinical notes.

Conclusions:

• CigStopper shows potential for automating billing processes, improving efficiency in tobacco cessation care.
• Machine learning models trained on synthetic data provide a foundation for clinical implementation.
• Automating billing tasks can simplify healthcare administration and enhance financial sustainability.