Revisiting the center effect.

Based on a model analyzing multiple effects, center variations in one-year outcomes from the OPTN/ UNOS Transplant Registry were used to estimate scores for 246 centers reporting 40+ renal transplants during 1995-2005. The adjusted one-year graft survival (derived from scores and 5 fundamental covariates) varied from 72% to 99%, demonstrating that the center effect is still the most influential factor in early renal graft survival. The 5 fundamental variables (recipient/ recipient's pre-transplant waiting time, original disease, age and degree of sensitization (PRA), proportions of expanded versus standard criteria donor grafts, donor age, cold ischemia time, transplant CMV and flow cytometry crossmatch statuses, and type of induction therapy. The associations indicated that centers with higher scores tended to transplant "better" kidneys into donor sex and race and transplant year) combined to reduce center variation by -10%. After categorizing centers into 5 groups based on increasing scores, we analyzed the extent to which groups differed over 15 secondary factors related to primary, deceased-donor transplant survival. All variables were statistically associated with center groups, but not all of the significant differences were clinically interesting. The notable associations included "more favorable" recipients. However, one-year graft survival rates still increased stepwise with increasing center success at nearly all levels of secondary variables. Thus, differences in center-specific success rates were not the direct result of center practices with regard to known kidney transplant factors.