Identifying neonatal intensive care (NICU) admissions using administrative claims data

Area of Science:

• Health Informatics
• Pediatrics
• Health Services Research

Background:

• Accurate identification of neonatal intensive care unit (NICU) admissions is crucial for healthcare analysis.
• Administrative claims data offer a valuable resource for epidemiological studies but require precise definitions for specific patient cohorts.

Purpose of the Study:

• To develop and validate a reliable method for identifying NICU admissions using administrative claims data.
• To establish a flexible and repeatable definition for NICU admissions applicable across different datasets.

Main Methods:

• Retrospective cohort study utilizing de-identified claims data from 2016-2020.
• Development of a definition for NICU admissions based on International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM), Current Procedural Terminology (CPT), and revenue codes.
• Statistical analysis including Kappa statistics, positive predictive values (PPV), and 95% confidence intervals (CI) to assess code agreement and accuracy.

Main Results:

• Combined use of CPT and revenue codes identified the highest proportion of NICU hospitalizations (13.7%), aligning with preterm birth rates.
• Revenue codes alone identified more NICU hospitalizations (3.3%) than CPT codes alone (1.5%).
• When CPT codes were the standard, revenue codes demonstrated high accuracy, identifying 86% of NICU admissions (sensitivity) and 97% of non-NICU admissions (specificity).

Conclusions:

• A robust definition for identifying neonatal admissions was successfully developed using administrative data.
• The established definition of NICU codes is highly adaptable, repeatable, and flexible for application in diverse datasets.
• This method enhances the utility of administrative claims data for research on neonatal care and outcomes.