AUTHORS
Aman Wadhwa, MD, MSPH, Yanjun Chen2, B. Grey Vandeberg, MS3, Lindsey Hageman, MPH4, Lucia Juarez, PhD5, Ariann Nassel, MA6, Elizabeth Baker, PhD7, Gabriela Oates, PhD8, Wendy Landier, PhD9, Smita Bhatia, MD, MPH10
AFFILIATIONS
1 University of Alabama at Birmingham, Birmingham, Alabama, United States
2 University of Alabama at Birmingham, Birmingham, Alabama, United States
3 University of Alabama at Birmingham, Birmingham, Alabama, United States
4 University of Alabama at Birmingham, Birmingham, Alabama, United States
5 University of Alabama at Birmingham, Birmingham, Alabama, United States
6 University of Alabama at Birmingham, Birmingham, Alabama, United States
7 University of Alabama at Birmingham, Birmingham, Alabama, United States
8 University of Alabama at Birmingham, Birmingham, Alabama, United States
9 University of Alabama at Birmingham, Birmingham, Alabama, United States
10 University of Alabama at Birmingham, Birmingham, Alabama, United States
ABSTRACT
Background: Prolonged mercaptopurine (6MP) exposure is required to maintain durable remissions in childhood acute lymphoblastic leukemia (ALL). Poor 6MP adherence resulting in low/varying 6MP exposure is associated with greater relapse risk (JAMA Oncol, 2016). In addition to age (adolescence), individual-level social determinants of health (SDoH: low annual household income/parental education) are associated with poor 6MP adherence in children with ALL (JCO, 2012; Blood, 2014). Despite a well-established association in non-oncology populations, the role of area-level SDoH in 6MP adherence remains unstudied.
Methods: We leveraged the Children’s Oncology Group (COG) study AALL03N1 where the primary aim had been to describe the prevalence/predictors of 6MP adherence measured using Medication Event Monitoring Systems [MEMS] TrackCaps. For the current study, we examined the association between area-level SDoH and 6MP adherence after accounting for the individual-level SDoH variables. Patients were enrolled ≥6 month after start of maintenance. Parents (patients if ≥18) reported patient sex and race/ethnicity, annual household income, parental education, and household structure. Residential ZIP-code at time of enrollment was linked to the Centers for Disease Control and Prevention Social Vulnerability Index (SVI) to determine association with the area-level SVI summary ranking (range, 0-1, greater values=higher vulnerability) and its four sub-themes: (1) socioeconomic status (SES), (2) household characteristics, (3) racial/ethnic minority status, and (4) housing type and transportation. Scores were divided into tertiles to identify residential areas of least vulnerability (T1), moderate vulnerability (T2), and most vulnerability (T3). The association between SVI and 6MP adherence was examined using generalized estimating equations adjusting for age at study, sex, and the following individual-level SDoH variables: race/ethnicity, income, education, and household structure (single parent/single child; single parent/multiple children; nuclear family); fitted mean adherence rates (least square means) from these models were reported.
Results: Median age of the 469 patients at study enrollment was 5.5y (range, 1.2-20y), 67% were male and 33.7% were non-Hispanic White. Low individual-level SES (annual household income ≤$50,000 and parental education <high school [HS]) were reported by 26% of the parents. Median SVI was 0.59 (0.04-0.97). After adjusting for individual-level SDoH variables, we identified the following area-level SDoH variables to be associated with poor 6MP adherence:
SVI summary score: Children living in the most vulnerable neighborhoods (T3) had lower 6MP adherence compared to those living in the least vulnerable (T1) neighborhoods (mean adherence: 88.3% vs 93.0%, P=0.02).
SVI SES sub-theme: Children living in most socioeconomically vulnerable neighborhoods (T3) had lower 6MP adherence when compared with those living in least vulnerable socioeconomic neighborhoods (T1) (mean adherence: 88.3% vs 92.1%, P=0.09); the difference did not reach statistical significance but trend aligned with summary score.
SVI household characteristics sub-theme: Residence in neighborhoods characterized by a high proportion of older or younger individuals, civilians with disability, single-parent households, poor English proficiency was significantly associated with lower 6MP adherence (mean adherence: T3=88.7% vs T1=92.7%, P=0.04).
SVI minority status sub-theme: Children living in neighborhoods with highest proportion of racial/ethnic minorities had lower 6MP adherence (mean adherence: T3=89.0% vs T1=93.7%, P=0.009).
SVI housing type and transportation sub-theme: Living in neighborhoods with a high proportion of multi-unit structures, mobile homes, crowded housing group quarters, and limited vehicle access did not influence 6MP adherence (mean adherence: T3=91.6% vs T1=92.2%, P=0.7).
Conclusions: Area-level SDoH were associated with poor 6MP adherence even after adjusting for individual-level SDoH. These findings suggest that patients need to be screened for adverse SDoH and vulnerable subpopulations provided anticipatory assistance to maintain high adherence levels with the overarching goal of minimizing the risk of relapse.
DISCLOSURES
Aman Wadhwa, MD, MSPH: No relevant conflicts of interest to declare.