Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 499 - New Methods for Machine Learning
Type: Contributed
Date/Time: Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
Sponsor: Section on Statistical Learning and Data Science
Abstract #312694
Title: Random Forest-Classification of Area Socioeconomic Status (SES) and Mortality Risk in Pediatric Acute Lymphoblastic Leukemia (ALL) in the US
Author(s): Fatima Boukari* and Hacene Boukari and Md Hossain
Companies: Delaware State University and Delaware State University and Nemours Biomedical Research, A.I. DuPont Children's Hospital
Keywords: socioeconomic status; random forest; lymphoblastic leukemia; SEER registry

Socioeconomic status (SES), a composite measure of numerous interacting indicators known to impact cancer outcomes, scantly receives systematic investigation. This study aims to explore the patterns of co-occurrence of county-level SES indicators and to examine associations between these patterns and pediatric ALL mortality risk. The study includes 19875 ALL patients diagnosed at ages 0-19 years during 1973-2016 in SEER registry areas. County-level 25 SES attributes were collected from 2013-17 ACS survey. Application of unsupervised random forest identified four groups with distinct patterns of SES attributes encompassing poverty, education, income, residential stability, immigration, cultural and language barriers. Five factors accounting for 88% of variance in SES attributes were identified: a) poverty/income/educational disadvantage b) immigration-related features; c) housing instability; d) crowding; and e) absence of moving. After adjustment for known prognostic factors – histologic subgroups, race-ethnicity, sex and age at diagnosis in frailty model– pediatric ALL patients displayed substantially distinct risk of mortality across pattern-based classifications and factors.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program