The scientific work of Peter Bauer

This June, Peter Bauer – chief editor of Biometrical Journal from 2000–2003 and member of the editorial board since 1996 – celebrated his 65th birthday. Peter Bauer became known as one of the originators of the “adaptive designs”, a field of research that recently gained considerable popularity, not only among academia but likewise among regulatory bodies and the pharmaceutical industry. It is characteristic for Peter Bauer, that while “adaptive designs” make their way into main stream statistics, he is already on the search for new challenges and now works, among others, on statistical problems relevant to bioinformatics. His sense for dynamic research areas can be easily seen from his whole professional career where he covered a wide range of research fields including informatics, binomial and poisson processes, model robustness, optimal designs, multiple testing, sequential analysis, adaptive designs and bioinformatics. Peter Bauer started his scientific career at the medical faculty of the University of Vienna, was appointed 1985 as head of the Department of Medical Documentation and Statistics in Cologne and returned to the University of Vienna in 1994 to become the head of the Department of Medical Statistics at the medical faculty. Besides his co-operation in many clinical research projects he built up an active research group. He currently decided to accept the lead of the Core Unit for Medical Statistics and Informatics, the superordinate unit of his and several other departments since the faculty became an independent university in 2004. In view of the proceeding structural changes at the new Medical University taking this lead is just one more example for his persistent willingness to explore new challenges. As a physicist by education, Peter Bauer was part of the pioneering epoch of medical computer science in Vienna. In the late sixties he developed a programming language to model chemical reactions and worked on computer based diagnostic programs for liverdiseases. Confronted with the need for statistical expertise at the medical faculty, he turned to statistics and became a British Council Research Fellow at the Department of Statistics in Edinburgh where David Finney was department head at this time. He returned to Vienna, fascinated by the British school of medical statistics, and got involved in several challenging consulting projects that were the starting point of methodological research papers. Motivated by practical experience with high dimensional compartment models he developed a method to specify the power for the discrimination between different regression models and investigated the related statistical design concept nowadays known as T-optimality. The method was published in two research papers of the “Biometrische Zeitschrift” (Bauer, 1975a, b) one year before the journal was renamed to “Biometrical Journal”. Another example for his practically oriented methodological research is the analysis of a cross-sectional study on the caries protective effects of fluorid. To model the hazard for the occurrence of caries Peter Bauer forestalled the development of frailty models (Bauer et al., 1980, 1982). In a fruitful cooperation with Peter Hackl he contributed to the theory of quality control and multiple testing. Peter Bauer published his first paper on multiple comparisons in 1985 in the Biometrical Journal (Bauer and Hackl, 1985). The multiplicity issue in statistical inference was the subject of many subsequent works with applications in dose finding (e.g. Budde and Bauer, 1989; Bauer and Budde, 1994; Bauer et al., 1998; Bauer, Brannath and Posch, 2000), model selection (Bauer, P tscher and Hackl, 1988), equivalence testing (e.g. Bauer and Bauer, 1994)