Establishing a high-risk neuroblastoma cohort using the Pediatric Health Information System Database

Area of Science:

• Pediatric Oncology
• Health Informatics
• Clinical Epidemiology

Background:

• The absence of specific International Classification of Diseases, 9th Revision (ICD-9) codes for neuroblastoma hinders patient identification in administrative databases.
• Accurate patient identification is crucial for epidemiological studies and comparative effectiveness research in pediatric cancers.

Purpose of the Study:

• To develop and validate a reliable algorithm for identifying a high-risk neuroblastoma cohort within the Pediatric Health Information System (PHIS) database.
• To create a valuable data resource for future research on high-risk neuroblastoma.

Main Methods:

• A three-step algorithm was designed, incorporating ICD-9 codes, specific exclusion criteria, and manual review of chemotherapy billing data.
• The algorithm was applied to the PHIS database to assemble a cohort of high-risk neuroblastoma patients (n = 952).
• Validation was performed at a single institution, assessing sensitivity and positive predictive value (PPV).

Main Results:

• The developed algorithm demonstrated high performance, achieving a sensitivity of 89.1% and a positive predictive value (PPV) of 96.1%.
• A cohort of 952 high-risk neuroblastoma patients was successfully identified and assembled.
• The methodology provides a robust approach for identifying similar patient cohorts in administrative datasets.

Conclusions:

• The algorithm effectively overcomes the limitation of missing ICD-9 codes for neuroblastoma identification.
• The identified high-risk neuroblastoma cohort serves as a valuable resource for future clinical epidemiology and comparative effectiveness studies.
• This approach enhances the utility of administrative databases for research in rare pediatric cancers.