Evaluation of the results of organ transplantation in France: is there a center effect?

The evaluation of the results of organ allografts for each type of organ and for each center is one of the prioritary missions of the Etablissement français des Greffes (EfG). The objectives, the methodology and the results of this evaluation have been defined and discussed with all the organ transplantation teams, after a preliminary work of the Conseil Médical et Scientifique of the Etablissement. This paper describes the evaluation experiment conducted by the EfG between 1995 and 1998. The main objectives of this first phase of the evaluation project are the identification of centers with outside of the norm results and the study of the relationship between the number of transplants performed by each center and the quality of their results. The chosen quality indicators are the excess in the patients mortality rate, computed one year after the first transplantation, for vital organs, and in the one year kidney graft loss rate for kidney transplantation. The excess of mortality is defined as the difference between observed and expected mortality rates. The expected mortality rate is estimated, for each program, by a statistical model based on a set of patient specific risk factors. All the vital organ transplantation teams who have performed more than 10 transplants between 1991 and 1995, and all the kidney transplantation teams who have performed more than 15 transplants between 1991 and 1996 have been included in the study. The main results of this evaluation experiment are the following ones: even if the intercenter variability of the results was statistically significant, it remained of low magnitude, particularly for heart, liver and kidney transplantation. None of the evaluated centers presented results outside of the norm. The results were positively correlated to the number of transplants for liver and lung transplantation. This was not the case for heart and kidney transplantation; this relationship is difficult to analyze for heart-lung transplantation, due to the small number of centers included. Thanks to the data base constituted since 1959 by all the organ transplantation teams, and in spite of the partially retrospective nature of this study, which explains its limits, this evaluation experiment, opens a perspective of extension to other domains of public health. In the future, however, this kind of evaluation should be prospective; a project aimed to developing the evaluation of the results of organ transplant actions in real time was def