The New Year holiday offers a time to reflect upon past accomplishments and plan for new resolutions. This year, transition is also in the air, as I take on the wonderful honor of serving as ASA president in 2018 while making a career change at the same time. This first President’s Corner gives me an opportunity to reflect on both the job I am leaving and the one I am about to assume, while giving Amstat News readers an inside look into what makes this president tick.

Lisa LaVange

The US Food and Drug Administration (FDA) employs more mathematical statisticians in the GS 1529 series than any other government agency—more than 350 by recent count. Statistical reviewers across the seven FDA centers occupy this civil service position to carry out the important work of regulating food, drugs, biologics, medical devices, veterinary medicines, toxicological research, and tobacco products for the American people.

As director of the Office of Biostatistics in the Center for Drug Evaluation and Research (CDER) since 2011, I have had the good fortune to lead an extremely talented and dedicated staff who develop and apply statistical methodology for drug regulation. Our staff members—predominantly holding doctoral degrees in statistics or biostatistics—advise pharmaceutical companies on statistical aspects of drug development; review clinical trial protocols; assess the benefit and risk of new drugs, biologics, generic drugs, and biosimilar products based on data submitted to the FDA for approval; and monitor the risk of drugs after they are marketed and their use expands. Our mission is to make sure safe and effective drugs are available to improve the health of people in the US.

A day in the life of a CDER statistician involves assessing the integrity of clinical data submitted in support of a marketing application and assessing the validity of the sponsor’s analyses intended to provide evidence that the drug is safe and effective. But our work does not stop with these basic components of a statistical review. We conduct research in clinical trial design and analysis to improve the efficiency of drug development; we investigate other data sources—popularly termed real-world data—for evidence of safety signals that need further study; and we conduct meta-analysis to investigate trends across trials or drugs in the same class.

One of our most important jobs is to develop statistical policy and issue guidance so pharmaceutical companies know what to expect in our reviews. The development of statistical guidance documents is a frequent commitment agreed to under the Prescription Drug User Fee Acts (PDUFAs), first passed into law in 1992 and renewed every five years. With these acts, fees associated with marketing applications are used to provide additional resources for hiring review staff and improving processes to enable faster turnaround times for regulatory reviews and marketing decisions. The user-fee programs have shortened review times for new drug applications from 2.5–3 years to under a year, and their success has led to similar programs for biosimilar products and generic drugs.

About a year into my job at FDA, I wrote about the challenges statistical reviewers faced in CDER and made a case for the indispensable role statistical thinking plays in meeting those challenges. This past year, I gave a JSM talk about my six-year FDA career highlighting the accomplishments of the talented statisticians I have had the privilege to lead. As a brief sampling of those highlights, CDER’s Office of Biostatistics did the following:

• Grew from 165 statisticians and support staff to 216 (as of November 30, 2017). This growth is analogous to hiring an entire academic department—no small feat given the flat-line growth of doctoral-level candidates in our field.
• Issued statistical guidance on noninferiority trials, multiple endpoints, and statistical assessment of analytical similarity for biosimilar products; completed draft guidance on adaptive designs and meta-analysis. These were in addition to collaborating on numerous disease-specific guidances for both new and generic drugs.
• Sponsored collaborative workshops and research efforts in several areas of unmet medical need, most notably development of anti-bacterial products, products to treat a variety of rare diseases with no available therapies, and oncology/hematology products. Research topics spanned new methods for trial design, new biomarkers to enrich trial enrollment, and new surrogate endpoints suitable for accelerated approval of ground-breaking therapies. Trial designs that make use of prior information through Bayesian modeling were a breakthrough for CDER during this period as a way of substituting information when patients are scarce.
• Encouraged pharmaceutical sponsors to join forces in master protocols and platform trials designed to study more than one therapy in more than one disease in a perpetual fashion. The FDA supported efforts of patient advocacy groups to serve as a catalyst in these multi-organizational endeavors in breast cancer, lung cancer, and drug-resistant bacterial infections, as examples.
• Contributed to finding a solution to the opioid addiction problem through the evaluation of abuse-deterrent formulations and monitoring of drug abuse rates to determine the impact of marketing these formulations. Our leadership in developing and evaluating methodologies and data sources for post-market surveillance has been instrumental in FDA’s continuing efforts to combat this immense public health problem.

This is not an advertisement for employment as an FDA statistician, but clearly, the impact of the job—the sheer number of topics and interesting problems our staff addresses—speaks for itself. What an immensely rewarding place to work!

As I leave one of the best leadership jobs in statistics, you might ask, “What were you thinking?” My decision was based on several factors. First, I felt a strong urge to attend to the training of statisticians, having struggled in their hiring for so long. My tenure at FDA has given me new insights into the importance of statisticians’ being able to navigate data sources of all types, sizes, and levels of complexity to search for signals and solve complex problems. Our leadership in this area—our ability as data scientists to quantify uncertainty and make meaningful interpretations of the results—has never been needed more.

Second, I felt the need to resume my work in leadership training. Having launched a doctoral-level course in statistical leadership in 2011 and continuing with numerous short courses and lectures on the topic, the time is right to return to this important task in earnest.

Third, I wanted more dedicated time to serve as ASA president than continuing my leadership role at FDA would practically allow.

And fourth, my family calls. After six and a quarter years of commuting from Chapel Hill, North Carolina, to Washington, DC, it was time to go home.

On January 1, I rejoined the faculty in biostatistics in the Gillings School of Global Public Health at The University of North Carolina at Chapel Hill. Here, I serve as associate chair of the department and director of the Collaborative Studies Coordinating Center and am coordinating development of the data science curriculum and reinstatement of leadership training. I am excited to return to a workplace and colleagues I know well and enjoy working with. The department is fully supportive of my ASA presidency, and my commute has been reduced to a 10-minute walk.

My new responsibilities as a biostatistics faculty member at UNC align well with the various initiatives and activities underway at the ASA. In future issues of Amstat News, I look forward to sharing details of my 2018 ASA presidential initiatives in leadership and professional development with you. In the meantime, Happy New Year!