Seeing Uncertainty: New Frontiers of Statistical Communication
When statistical sciences were in their infancy, the communicative benefits of statistics were widely touted. Thousands of different data points could be neatly summarized into a single figure, creating efficiencies in transmitting knowledge and clarity of understanding. As statistics has grown more complex, the ability of the statistical sciences to engage an easy transfer of knowledge has grown complex as well. A recent Cytel webinar panel noted that the FDA briefing on coronavirus vaccines struggled to capture Bayesian findings accurately, and within a number of biotechs and pharmaceuticals the inability to fully communicate how a certain design can de-risk a trial, means that clinical trials launch with greater uncertainties.
Even when these uncertainties are easy to quantify they are difficult to use for decision-making. Can a 0.91% increase in statistical power justify a 0.89 month extension of a clinical trial, should seem like a straightforward decision, but is not. When the numbers are not simple though, and much is at stake, the need for the clear communication of statistics for the understanding of tradeoffs becomes even more critical.
Clearer communication of uncertainties has often been thought to be the role of a statistician, but new technology can also help facilitate the transmission of statistical findings. A webinar by Cytel CSO Yannis Jemiai raises the question of how the entire design and selection of clinical trials will transform in the next few years. Clearer communication of statistical tradeoffs is a key part of the Reimagined Clinical Trial.
According to Yannis, one under-examined area of statistical communication is how to use technology to create precise visualizations that are intuitive for both statisticians and non-statisticians to use. These visualizations will not only capture probabilities, but also tradeoffs between parameters like speed and savings, speed and clinical trial power, and other relevant factors. They will also display more intuitively how statistical characteristics of design align with business goals like expected net present value.
Such visualization tools are available in Cytel’s new Solara software. You can learn more about such visualization tools and their role in strategic decision-making by clicking on the webinar link below.