Developing case-finding algorithms for second events of oropharyngeal cancer using administrative data: A...

Area of Science:

• Oncology
• Biostatistics
• Cancer Research

Background:

• Second event-free survival is key for evaluating cancer treatment effectiveness.
• Administrative data, like cancer registries, often lack explicit documentation of second events (recurrences or new primary cancers).
• Validated algorithms are needed to accurately identify second events in oropharyngeal cancer patients using existing administrative data.

Purpose of the Study:

• To develop and validate algorithms for identifying second events in oropharyngeal cancer patients using administrative data.
• To assess the performance of these algorithms in terms of sensitivity, specificity, and predictive values.

Main Methods:

• Classification and regression tree models were employed to develop the algorithms.
• Chart review data served as the reference standard for validation.
• Key performance metrics including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated.

Main Results:

• A high-sensitivity algorithm demonstrated 87.9% sensitivity and 84.5% specificity.
• A high-PPV algorithm achieved 99.1% specificity and 94.2% PPV.
• Both algorithms showed acceptable accuracy and NPV for identifying second events.

Conclusions:

• The developed algorithms demonstrate acceptable validity for identifying second events after primary treatment of oropharyngeal cancers.
• These validated algorithms can improve the assessment of treatment efficacy by leveraging administrative data.
• The study provides a method to overcome limitations in administrative data for cancer research.