In the complex world of drug development, the interaction between the various stakeholders is essential for the success of a project. A fundamental part of communication is a basic understanding of the problems, terminology and way of thinking of your team members and partners.
Traditionally, statisticians have a particular way of thinking, which is very different from the medical way. Distributions, probabilities and hypothesis testing – which are basic concepts in statistics – have no counterpart in the medical world, whereas adverse events, treatment decisions and comorbidities sound strange to a statistician. Consequently, statisticians think in mean values and standard deviations, whereas clinicians think in “individuals” – they know the exact blood pressure of “patient x” but only have a vague idea on the average age of their patients.
The planning phase of a project is crucial for its success, not only because decisions made in this phase will be valid throughout the life cycle, but also because of the strict requirements from regulatory authorities for granting market access (which must be taken into account). Today, clinical trial methodology offers a wide spectrum of tools not only to minimize the sample size of a single study, but also to reduce the time used for the whole development process. All possible study designs should be discussed in a very early stage of planning. The worst case for a study protocol, and a nightmare for statisticians, is being asked to perform a sample size calculation and complete the sections for statistical methodology two weeks before a planned submission of a protocol to the ethics committee.
Statistical input into a study protocol is not limited to sample size calculation, and the selection of appropriate statistical methods. It also includes input into the selection of the “best” endpoint (if not defined by regulatory guidelines), strategies for early stopping, seamless study designs, handling of multiplicity problems and ensuring that all regulatory requirements from a methodology point of view are fulfilled.
To take full advantage of the statistical input not only in the planning stage – but rather in all stages of drug development – a basic understanding of statistical methods is necessary. For example, in the case of multiple endpoints, a discussion on the hierarchy of endpoints – and associated hypotheses – is much more efficient if all partners understand why an adjustment of the error of probability is necessary and what the consequences for final interpretation of the trial results will be. Another example is the definition of stopping rules and understanding the interpretation of simulations a statistician might present.
Each partner in the drug development process should undertake some effort to understand the role each partner has in order to allow for a successful interaction.
Basic statistical principles are relatively easy concepts to understand and can be easily taught by statisticians to other members of the clinical team.
For more information, visit www.crosnt.com/cros-academy