Assembling ALL and AML Cohorts to Analyze Center-Level Variations in Induction Mortality and Evaluate Treatment Strategies for Relapsed AML

Background

The Pediatric Health Information System (PHIS) is a database of daily billing records from various units from 46 free-standing pediatric hospitals in the US. Utilizing PHIS, I will use methods similar to those already developed by the Aplenc lab to expand existing cohorts of pediatric patients with acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) to allow for important clinical questions in pediatric leukemia to be definitively addressed.

Project Goal

First, I plan to explore differences in mortality during induction, which is the first month of therapy for newly diagnosed ALL and AML patients. There is variability among death rates during induction across hospitals, suggesting that modifiable institutional characteristics can influence the mortality rate during this time period. Using the newly expanded cohort, I will analyze center-level variation in induction mortality with the goal of conclusively determining the role hospitals play in induction mortality, which could impact development of standardized patient care to potentially reduce mortality during this time period.

Second, I plan to evaluate treatment strategies for relapsed AML. Treatment outcomes for pediatric patients with AML have improved over the past few decades, with a five-year overall survival rate currently exceeding 60%. Still, more than one-third of patients experience relapse, associated with a significantly poorer prognosis upon relapse. There are limited data comparing chemotherapy options, and thus, there is no established standard of care for these patients. I will use the expanded AML cohort to identify a subpopulation of AML patients who experienced relapse and assess the comparative effectiveness of various treatments for relapsed AML.