On controversial statistical issues in clinical research

In clinical development of a test treatment under investigation, clinical trials are often conducted for evaluation of safety and efficacy of the test treatment. To provide an accurate and reliable assessment, adequate and well-controlled clinical trials using valid study designs are necessarily conducted for obtaining substantial evidence of safety and efficacy of the test treatment under investigation. In practice, however, some debatable issues are commonly encountered regardless compliance with good statistics practice and good clinical practice. These issues include, but are not limited to: 1) appropriateness of statistical hypotheses for clinical investigation; 2) correctness of power analysis assumptions; 3) integrity of randomization and blinding; 4) post hoc endpoint selection; 5) impact of protocol amendments on the characteristics of the trial population; 6) multiplicity in clinical trials; 7) missing data imputation; 8) adaptive design methods; and 9) independence of a data monitoring committee. In this article, these issues are briefly described. The impact of these issues on the evaluation of the safety and efficacy of the test treatment under investigation are discussed with examples whenever applicable. Some recommendations regarding possible resolutions of these issues are also provided.

Duke Scholars

Author Shein-Chung Chow Biostatistics & Bioinformatics, Division of Biostatistics