Quantifying disparity: Race, socioeconomic status, and survival disparities in AML

Abstract Background: Acute Myeloid Leukemia (AML) outcomes are shaped by far more factors than disease biology perse. Other factors including race, ethnicity, socioeconomic status, and healthcare access weigh in. Many studies have suggested that disparities impact outcomes. However, the data are scattered, and the lack of a recent synthesis of quantitative effect sizes obscures the overall landscape, making it difficult to grasp the bigger picture. We conducted a systematic review and meta-analysis to test how comparable these studies really are, and exploring what they mean for interpreting results. Methods: Following PRISMA guidelines, we systematically searched PubMed, Web of Science, and Cochrane Library for studies published between 2015 and 2025. Search terms encompassed AML, myelodysplastic syndromes (MDS), chemotherapy, hematopoietic cell transplantation, race, ethnicity, socioeconomic status (SES), and disparities. Inclusion criteria were adult patient populations, original research reporting hazard ratios (HRs) or odds ratios (ORs) for Overall survival (OS), mortality, or treatment utilization, and comparisons by racial/ethnic group or SES. Out of 5,529 records, after excluding 361 duplicate records (722 entries), 79 studies met eligibility. Data extraction was performed in duplicate. Of these studies only 7 were found to be comparable depending based on their subgroup analysis. Meta-analyses were conducted with heterogeneity assessed via I² and Cochrane's Q; p < 0.05 was considered statistically significant. Results: Hispanic vs Non-Hispanic White (NHW) males (OS): 2 studies; pooled HR = 0.70 (95% CI: 0.53–0.93), I² = 0%, Cochrane's Q 0.94, p-value 0.33 suggesting consistent findings across studies and significantly improved OS in Hispanic males. Non-Hispanic Black (NHB) vs NHW males (mortality): 2 studies; pooled HR = 1.26 (95% CI: 1.08–1.47), I² = 72.9%, Cochrane's Q 2.13, p-value 0.14 indicating significantly higher mortality among NHB males but with substantial heterogeneity, suggesting differences in study design, populations, or confounder adjustment. Receipt of systemic chemotherapy (NHB vs NHW): 3 studies; pooled OR = 0.93 (95% CI: 0.74–1.17), p = 0.54, I² = 72.9%, Cochrane's Q 3.69, p-value 0.54 showing no statistically significant difference, but marked variability between studies. Conclusions: Only a small subset of the identified studies was directly comparable for pooled analysis, limiting the generalizability of findings. The low I² in the Hispanic vs NHW OS analysis suggests some degree of agreement between those studies, whereas the high I² in the NHB mortality and chemotherapy receipt analyses indicates that the majority of the variability in effect sizes is due to differences between studies rather than chance. This heterogeneity, combined with potential publication bias and varying adjustment for confounders, warrants cautious interpretation. While these pooled results suggest better OS for Hispanic males and higher mortality risk for NHB males with AML, we still lack the level of standardization, consistency, direct comparability needed to clarify the relationships between race/ethnicity, treatment utilization, and survival in AML.