Aligning EHR Data for Pediatric Leukemia With Standard Protocol Therapy

Area of Science:

• Pediatric Oncology
• Health Informatics
• Clinical Trial Management

Background:

• Children with acute lymphoblastic leukemia (ALL) are treated using risk-based protocols from the Children's Oncology Group (COG).
• Understanding the alignment between real-world clinical practice and COG protocol milestones is crucial but not well-established.
• Aggregate deidentified electronic health record (EHR) data present an opportunity to assess clinical practice in real-world settings.

Purpose of the Study:

• To evaluate the alignment of real-world clinical practice with established Children's Oncology Group (COG) protocol milestones for pediatric acute lymphoblastic leukemia (ALL).
• To assess the utility of aggregate deidentified electronic health record (EHR) data in understanding clinical practice variations and adherence to treatment guidelines.

Main Methods:

• A cohort of pediatric ALL patients was identified using Cerner Health Facts deidentified aggregate EHR data.
• Automated methods were developed to classify standard-risk precursor B-cell ALL patients, with manual review for procedural milestones.
• Milestone procedures, specifically lumbar punctures (LP), were analyzed relative to therapy initiation and aligned with COG standard induction therapy protocols.

Main Results:

• The study identified 7,728 pediatric ALL patients across 188,187 encounters.
• Lumbar puncture (LP) and bone marrow biopsy records were sufficiently present to evaluate guideline adherence.
• Analysis of 14 health systems showed that while not all systems had data for all recommended LP procedures, most demonstrated adherence to recommended timing for LPs.

Conclusions:

• Variations in data sources can lead to inconsistencies in aggregate EHR data, necessitating careful data interpretation and visualization.
• Despite data limitations, health systems showed strong alignment with recommended lumbar puncture (LP) milestones for pediatric ALL patients.
• Large-scale aggregate EHR data are valuable for assessing the concordance between recommended and actual clinical milestones in pediatric ALL treatment, informing future guideline development.