The association between hospital capacity and survival after allogeneic bone marrow transplantation (allo-BMT) was examined using the dataset accumulated by the Japan Society of Hematopoietic Cell Transplantations (JSHCT). The subjects were 3134 patients who received first allo-BMTs between 1991 and 1997 reported to the JSHCT. They were divided into three groups by cumulative hospital experience of allo-BMTs: low volume (capacity) (LV; < or = 25 cases), moderate volume (capacity) (MV; 26-75 cases) and high volume (capacity) (HV; > or = 76 cases). Using a proportional hazards model, the association of hospital experience with early survival at day 100 (D100S), and overall survival (OS) were examined. For leukemia patients, leukemia-free survival (LFS) was also analyzed. When HV was defined as the reference group, the hazard ratios (HRs) of OS for all subjects were 1.10 (95% confidence interval; 0.97-1.25) for MV and 1.25 for LV (1.08-1.44). The HRs with D100S were 1.20 (0.96-1.51) for MV and 1.40 (1.08-1.80) for LV. Larger values were observed for OS and D100S in cases of leukemia. Survival after BMT from sibling donors was clearly influenced by hospital experience, but this was not the case from unrelated donors. These findings suggest that size of the transplant team should be considered in order to improve the outcome of sibling BMT in general.