The 2016 ASA Biopharmaceutical Section Regulatory-Industry Statistics (RIST) Workshop has a theme of Statistical Innovation: Better Decisions Through Better Methods, and the casual observer may believe that such a topic may be solely limited to innovative ways to perform analysis. However, many non-statisticians (and even some statisticians) forget about the important role that statisticians play early on in the design of a clinical trial. But while statisticians may concern themselves with the more statistical aspects of study design, such as identifying a sample size to appropriately power the trial, managing type I error through multiple endpoints or interim analyses, or limiting bias through the use of carefully constructed randomization schemes, they may never consider the importance of data collection and data quality and its impact on the trial. This is unfortunate! The quality of the data impacts the validity and integrity of the final analysis, and our ability to communicate meaningful results. In fact, the analysis is entirely dependent on the data! We’ve all heard the adage “garbage in, garbage out”, and truer words were never spoken. The most elegant and cutting-edge analysis or sophisticated design won’t rescue a study if the data do not reflect our expectations. Often data issues are identified after the trial is completed, necessitating numerous sensitivity analyses, and risking potential regulatory scrutiny. Ideally, data quality should be assessed during the trial in order to correct the ongoing study.
As statisticians, we need to take an active role in preserving data quality, and our unique skills can aid the development team in identifying problematic data. Beyond protecting the final analysis, our role in assessing data quality protects the well-being of study participants, making us agents of good clinical practice. Consider participating in the pre-Workshop short course An Overview of Methods to Assess Data Integrity in Clinical Trials.