Background: Literature suggests that inpatient hematology-oncology (heme-onc) clinical experiences may have a deleterious impact on trainee interest in heme-onc. However, the impact of inpatient heme-onc service structure on trainee education is not known. To delineate the factors affecting trainee education, we aimed to characterize the structure of inpatient heme-onc services at comprehensive cancer centers (CCCs) and the associated impact on trainees. Methods: We designed a survey with Likert-type and open-ended questions using established procedures. We surveyed heme-onc fellowship program directors (PDs) at all 50 U.S. CCCs with fellowship programs. We performed point biserial and Pearson correlations on Likert-type responses and content analysis on open-ended responses. Results: Eighteen out of 50 (36%) PDs responded and described 117 services. Services were either primary (35%), consult (44%), or both (21%). Fellows were present on 93% and residents on 54% of services. PDs rated the educational value of most inpatient services as either “very high” or “somewhat high” for both fellows (73% very high, 13% somewhat high) and residents (49% very high, 33% somewhat high). PDs perceived that fellows derived greater educational value from consult-only vs primary/mixed services [R(105) = 0.25, p= 0.01] and provided greater reciprocal benefit to consult-only vs primary/mixed services [R(100) = 0.28, p= 0.004]. Hospital type (county, university, or veteran’s) and the presence or absence of a formal curriculum on the service did not correlate with perceived educational benefit. Multiple open-ended comments suggested that heme-onc attendings and fellows should be consultative rather than primary service providers, programs should separate teaching and non-teaching services, and service/education balance optimization should be prioritized. Inadequate service staffing and lack of institutional support were identified as barriers to inpatient education. Conclusions: PDs felt that inpatient heme-onc services had high educational value for trainees, particularly for fellows. Both fellows and services may derive mutual benefit from fellow participation on inpatient consult services structured separately from non-teaching primary services. Hospital type and the presence or absence of a structured curriculum may not be critical for education on inpatient heme-onc services. Appropriate service/education balance, staffing, and institutional support are crucial to optimize inpatient education. Future studies should explore the trainee perspective around the educational value of inpatient clinical experiences. PDs and institutions can use the study findings to craft beneficial inpatient educational experiences for heme-onc trainees.