What do regulatory statisticians expect from me?

December 2013

Focus on the Study Protocol:

Statisticians working for the regulator (e.g. EMA, FDA) expect a study that is well-designed, not just from a clinical perspective but also from a statistical perspective. A study that is designed in a statistically sound manner is also far more likely to succeed!

The importance of comprehensive statistical planning is an often overlooked aspect of the study protocol, viewed as less essential to the study design. However, it is important to note that the International Conference on Harmonisation (ICH) E9 guidelines dictate that “the protocol and subsequent amendments should be approved by … the trial statistician.”

Every aspect of a study’s design is potentially critical for determining how statistical analyses will be performed and, in turn, can affect study outcome. Among the many components of a study protocol that are statistically relevant:

  • Endpoint definition
  • Randomization scheme and additional methods to avoid bias in study results such as blinding
  • Number of study sites
  • Sample size considerations
  • Treatment of incomplete data and outliers
  • Target population (age, gender and race covariate effects)
  • Analysis populations definitions

As all of these components need to be factored into statistical analyses, it is imperative that the trial statistician be involved from the very start of the process of designing a study.