Editor’s Note

As the Editor of Statistics in Biopharmaceutical Research (SBR), I try to ensure that the journal continually provides important, useful, and timely information relevant to ongoing issues in clinical trials, regulatory science, biostatistical methodology and applications, and emerging topics in our professional community. Together with our excellent editorial team, I have organized and published a several special sections or issues, such as “Statistical Challenges in the Conduct and Management of Ongoing Clinical Trials during the COVID-19 Pandemic” (Hamasaki et al. 2020), “Roles of Hypothesis Testing, p-Values, and Decision-Making in Biopharmaceutical Research” (Hamasaki et al. 2021), and “Statistical Issues and Challenges in Clinical Trials for COVID-19 Treatments, Vaccines, Medical Devices and Diagnostics” (Cooner et al. 2022).

SBR typically publishes articles that discuss statistical methodology and information regarding the use of statistics in all phases of research, development, and practice in the pharmaceutical, biopharmaceutical, device, and diagnostics industries.

The article by Professor Scott R. Evans, the George Washington University Biostatistics Center and the Department of Biostatistics and Bioinformatics (Evans 2022), is unusual for SBR. It is a transcription of his invited talk from August 6, 2019 at the U.S. Food and Drug Administration Statistical Association (FDASA) Seminar. The recording is available at https://collaboration.fda.gov/pqxxxxogesrm. This article is free to access and available for anyone to read.

There are no perfect statistical models or methods for designing and analyzing clinical trials. Methods have advantages and disadvantages. As statisticians, we need comprehensive understanding of the pros and cons of alternative approaches to choose the approach that is best suited for practical problems at hand. Assumptions and limitations need clear articulation to ensure transparency and appropriate contextual interpretation based on the reliability and robustness of the evidence.

After collecting and analyzing data in clinical trials, sensitivity analysis are often conducted to evaluate whether incorrect conclusions may have been drawn due to departures between characteristics of the observed data and assumptions required for validity of the applied model. From this perspective, the fewer the number of assumptions, the better. Assumptions should be reasonably justified and, to the extent possible, verifiable through data, regardless of whether they are implicit or explicit.

During the last two years as the Editor of SBR, I have observed that some of the models which are labeled “innovative” or “novel,” require strong and nonconfirmable assumptions, when designing and analyzing clinical trials. In some cases, these assumptions and the resulting implications of nonconfirmable assumptions are not clearly articulated. Such assumptions may make results from the trials less robust and difficult to interpret. This increases the likelihood that people will draw incorrect conclusions, ultimately increasing the likelihood that patients will receive potentially ineffective or unsafe treatments.

Evans (2022) raises concerns about some of the trends that are happening in our profession noting how sacrifices to scientific rigor are branded as innovation. I believe that we need any innovative solutions to share certain characteristics, for example, proven robustness with interpretable results. This article provides us an important opportunity to think about the future direction of statistical innovation.

I express deep gratitude to the FDA for providing their permission to publish the transcription in SBR. I thank Dr. Gene Pennello for serving as the guest editor for this article.

Funding

The author(s) reported there is no funding associated with the work featured in this article.