Analysis of Clinical Transplant Data: A Personal Comment

  • Gerhard Opelz
Conference paper
Part of the Lecture Notes in Medical Informatics book series (LNMED, volume 34)


A variety of methods exist for the computation of transplant success rates and their statistical evaluation. For someone not formally trained in statistics (like myself), expressions such as actuarial, regression, multivariate, etc., are impressive and bewildering. Like most transplant physicians, surgeons, or immunologists, I am inclined to respectfully accept p-values delivered by the computer. In my efforts to overcome my statistical deficiencies, one wisdom I adopted without difficulty was that “multivariate” analyses must be superior to “univariate” analyses. After all, the more comparisons and stratifications were carried out the better the result, or so I believed. It was no little surprise, therefore, when M. R. Mickey concluded, after having performed an extensive study based on the UCLA transplant registry data, that multivariate and univariate analyses essentially gave the same answers (1). Had just “some statistician” come to that conclusion, I would not have been bothered too much. That someone who I knew was a renowned member of the international biostatistical community had made the statement, however, did impress me. Could it be that what I had suspected all along but never dared to say was true? That all those complicated computations were necessary only if the data were “no good”? Or to be more precise, that if the patient numbers were sufficiently large, at least with renal transplant data, it did not matter what type of analysis one used? How did that compare with the popular notion that anything other than multivariate was primitive, unsophisticated, perhaps worthless? The revelation that “multivariate” is not unfailingly superior to “univariate” encouraged me to go ahead with writing this manuscript which, I can assure the reader, I was reluctant to do. My apology for contributing the manuscript to a journal of medical informatics is that even naive thoughts sometimes stimulate the experts, if to nothing else perhaps to educate those of us on the “user side”.


Graft Survival Graft Survival Rate Patient Survival Rate Graft Outcome Transplant Data 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Mickey MR: Multivariable analysis of one-year graft survival. In Clinical Kidney Transplants 1985 (P.I. Terasaki, Editor), UCLA Tissue Typing Laboratory 1985, pp. 27-44.Google Scholar
  2. 2.
    Mickey MR, Opelz G, Terasaki PI: Prospective estimates of success of kidney transplants. Transplant Proc 11:1914–1915, 1979.PubMedGoogle Scholar
  3. 3.
    Hennige M, Köhler CO, Opelz G: Multivariate prediction model of kidney transplant success rates. Transplantation 42:491–493, 1986.PubMedCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1987

Authors and Affiliations

  • Gerhard Opelz
    • 1
  1. 1.Department of Transplantation Immunology, Institute of ImmunologyUniversity of HeidelbergHeidelbergFR Germany

Personalised recommendations