Statistical Engagement in Regulatory Interactions

INTRODUCTION

It is indisputable that the discipline of statistics is quite important in many interactions between pharmaceutical companies and regulatory agencies regarding drug development, drug approval, and drug promotion (Lewi 2005). This importance may be due to the statistical methods applied to the raw data and/or the interpretation of the analyzed results (Marquardt 1987). This is true whether the application of statistics concerns randomized trials or observational databases for real-world evidence and cost-effectiveness evaluation. This points to a prominent role of the statistician within these activities. This role is often recognized within the pharmaceutical company and among regulatory agencies, as can be inferred from the various guidance documents and reflection papers issued by the US Food and Drug Administration (FDA), the European Medicines Agency (EMA), and the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH).

The design of studies is typically a joint activity between the statistician and the clinician, but the analytical methods are almost exclusively the provenance of the statistician. The interpretation of the results is also strongly influenced by the methods used in the analysis and the conclusions of the statistician. This prominent role of the statistician does not always extend to a strategic role within the company or to a broader representative role in regulatory interaction. Ideally, the statistician should think strategically about the totality of the data and thereby have a wider role than a purely statistical one internally and in regulatory interaction (Rockhold 2009). This is less about the definition of roles or overcoming barriers or the need to sell the statistical profession (Grieve 2002). Those are past battles that have been largely won. This is more about behaviors and knowing when to insert oneself into the game (Emir et al. 2013; Grieve 2005; Unwin 2007). Specific areas of statistical activity are examined in this chapter with the focus on highlighting behaviors to enhance the strategic and representative roles of the statistician in regulatory and other external interactions.

INTERNAL BEHAVIORS

The statisticians ability to be influential directly in regulatory interactions must first be established and recognized within the pharmaceutical company itself. Thus, the behaviors begin at home. Even the prominent internal role of the statistician in the design, analysis, and interpretation of study results is often not enough to consider the statistician as more than a supportive role in direct interaction with regulators. Access to the process as a strategic player often happens as a result of behaviors and successes on past projects. Statisticians should look to make presentations within the pharmaceutical company to non-statistics groups (e.g., regulatory, clinical, marketing, or outcomes research) relating statistical methods and principles to regulatory topics. For example, the concept of “estimands” is a statistical issue with a strong regulatory impact. Statisticians should take the initiative on explaining this concept and its application to clinical trial analysis and interpretation. Thus, when the inevitable regulatory issue of estimands arises in the design and/or analysis of clinical trials, the statistician should be perceived as the critical person to represent the company in the interaction with regulators.

Statisticians should also be well-versed in applicable regulations and review and comment on draft regulatory guidance. For example, in June 2018 the US FDA issued a guidance for industry titled, “Medical product communications that are consistent with the FDA-required labeling” (FDA 2018). This guidance recognized that sponsors have additional information about approved uses of their products that is consistent with the label, and communication of this information to patients and prescribers is helpful. The guidance deals with how FDA evaluates consistency of a product communication with the products labeling and gives general recommendations for conveying information in a truthful and nonmisleading way. While not overtly statistical in nature, the guidance does rely on the concept of “false and misleading.” To avoid being false and misleading there should be scientifically appropriate and statistically sound support for the promotional communication. The guidance points out that findings or the conclusions that can be drawn from supportive data should not be overstated. The communication requires proper contextual language including “limitations of the strength of evidence.” The statistician is the most appropriate person to make sure these criteria are satisfied, but regulatory and marketing groups within a pharmaceutical company may not recognize this. However, a statistician who takes the initiative to summarize the guidance and what communications may or may not satisfy the “false and misleading” criterion from a statistical perspective will enhance his/her strategic value.

Some external activities can also lead to a bigger strategic role in regulatory interaction. Publications on statistical methodology, applied to clinical trials or any other field in general, is one such activity. Making presentations at scientific meetings and serving on joint regulatory and industry working groups are others. However, because these activities may not be fully known or appreciated by senior management and do not include direct interaction between the project teams and the statistician, the influence of external activities may be limited. The direct, internal strategic activities are more effective in establishing the strategic role of the statistician in regulatory interaction. As discussed in the next section, with the extensive use of external Data-Monitoring Committees on important clinical trials, the role of the independent statistician in such committees offers another way to increase the strategic role of the statistician within the pharmaceutical company.