Being a Statistician: An Art, a Science or Just Job?

Posted by Cytel

Oct 8, 2015 4:25:00 PM

A recent American Statistical Association conference featured a town hall meeting to discuss the role of the statistician within the pharmaceutical industry, and in particular within the subfield of precision medicine. The panel of distinguished academics, regulators and members of industry confronted the concern that the era of precision medicine heralds an emerging need for statisticians to play a different role than the ones normally associated with technique and craft. Statisticians are now charged with discovering insightful applications to new statistical methodologies, and ensuring that clinical development reflects as many innovative statistical methods as necessary to achieve a successful study. According to the panel, statisticians may even find themselves in a position where they are required to anticipate scientific and technological advances for a study’s benefit.

This led the panel to ask whether the role of an industry statistician should be understood as diverging from the role of statistics within the pharmaceutical industry. Panelists were asked to comment on these differences, as they relate to current enterprises in precision medicine and missing data.

One view that particularly affected members of Cytel attending the session was raised by Tom Fleming of the University of Washington. Professor Fleming noted that the often cited triad of technique, art and communication seemed to apply now more than ever to statisticians in the pharmaceutical industry. Whereas solid statistical technique is necessary to solve statistical problems with the rigor and analysis necessary for clinical trials, what he called ‘statistical art’ might require the use of these same techniques in ways that are innovative or unconventional. As an example from our own experience, the idea that Bayesian methodology should be the chosen framework for pediatric trials where the therapy already has an adult dose, not only requires mastery of Bayesian techniques but an awareness of the scope of Bayesian statistics. (Bayesian pediatric trials.)

Communicating such innovations to non-statisticians remains a complex issue within the industry. Living in the age of infographics certainly helps with communication to some degree, but we also see more software packages and analytical tools responding to the need for easy communication with enhanced visuals when generating charts, reports or PIPs. Zoran Antonijevic once argued that within the context of adaptive clinical trials, the art of communication might sometimes require courage from more introverted statisticians. This is likely true for precision medicine as well.

Where art and communication seem to overlap is that precision medicine raises new statistical problems that by definition require innovative statistical solutions. We know, for example, that choice of subpopulation, selection of surrogate endpoints, and design of enrichment designs raise complications that have never been seen before now. Explaining this to the non-statistician can be quite challenging.

Topics: Statistical Programming, statistical communication

The Cytel blog keeps you up to speed with the latest developments in biostatistics and clinical biometrics.  Sign up for updates direct to your inbox. You can unsubscribe at any time.


Posts by Topic

see all

Recent Posts