Background: Adolescents and young adults (AYA: 15-39 years) with acute lymphoblastic leukemia (ALL) have inferior survival when compared with children (1-14 years). An approach is lacking that includes both patients enrolled and not enrolled in clinical trials, and includes the contribution of health care delivery, treatment, and clinical prognosticators. Methods: We assembled a retrospective cohort of ALL patients diagnosed between 1-39 years (AYA: n = 93; child: n = 91) and treated at a single institution between 1990 and 2010, irrespective of clinical trial enrollment. We modeled relapse risk (i) during therapy and (ii) after completing therapy. Results: On-therapy relapse: AYA experienced an increased risk of on-therapy relapse versus children (HR, 10.5; P = 0.004). In multivariable analysis restricted to AYA, independent predictors of relapse included lack of clinical trial enrollment (HR, 2.6, P = 0.04) and nonwhite race/ethnicity (HR, 2.2; P = 0.05). Relapse after completing therapy: When compared with children, AYA experienced an increased risk of relapse after completing therapy (HR, 7.7; P < 0.001). In multivariable analysis restricted to AYA, longer therapy (months of maintenance: HR, 0.7; P < 0.001; months of consolidation: HR, 0.8; P = 0.03) protected against relapse. Conclusions: Among AYA, aspects of health care delivery (clinical trial enrollment, nonwhite race/ethnicity) are associated with relapse during therapy, and aspects of treatment (shorter duration of maintenance and consolidation) are associated with relapse after completing therapy. Impact: These findings highlight the importance of clinical trial enrollment and therapy duration (maintenance, consolidation) in ensuring durable remissions in AYA ALL. Future studies encompassing health care delivery, treatment, and biology are needed.