In honor of the Department of Statistics' 50th Anniversary, we are highlighting our outstanding faculty and alumni of the past half-century.
Growing up, Carnegie Mellon Statistics Department alumnus Scott Berry (Ph.D. 94) engaged in dinner table conversation in which his father, renowned statistician Donald Berry, asked questions and raised issues rooted in critical statistical thinking.
“Invariably, when I got to college I found statistics interesting, and was good at it,” Scott said.
Today, he is president of Berry Consultants, which he co-founded with his father in 2000. To devote himself full-time to the new venture, Scott resigned from a tenure track assistant professorship at Texas A&M University. His father, who is a professor in the Dept. of Biostatistics at the University of Texas MD Anderson Cancer Center, works part-time at Berry Consultants.
Professor Jay Kadane was Scott’s advisor at Carnegie Mellon, and his father’s co-advisor at Yale. “I’m pleased to report that never once, in working with Scott, did I compare the two,” Jay said.
Headquartered in Austin, Texas, Berry Consultants is a statistical consulting company specializing in Bayesian statistics, an approach that is radically changing the way research is done throughout the medical industry in both device and drug development. “The world of biostatistics and clinical trials is very traditional with many established standard approaches, not because they are great, but because they are, well, standard. In the beginning, we were entering this very traditional world and doing very Bayesian things. “We had a great deal of success as more and more clients wanted to do innovative things,” Scott said.
For the first five years, the company consisted of himself and his father as the sole statisticians. As the demand for services grew, Scott hired fellow Carnegie Mellon Statistics alumnus Jason Connor (Ph.D. 06), who is today Director and Senior Statistical Scientist with expertise in Bayesian statistics and designing adaptive clinical trials. “When we hired Jason we were doing very Bayesian analyses which requires a deep understanding of, and the computational expertise of, the Bayesian approach,” Scott said. “There were very few people with that combination of expertise, and almost all were from Carnegie Mellon. The huge part of what we do is the Bayesian approach and the computational tools needed for that. “The education we received at Carnegie Mellon was crucial for us to be able to enter a different world with complete expertise of what we were doing. “It was more than just the technical skills. At Carnegie Mellon, we were always questioning things -- what is the root of the problem, and why do we do this. Jay Kadane was great, and taught us to do that. “Bringing the scientific approach to what we do also came from Carnegie Mellon. We have to be able to understand the science; the problem; be able to create solutions; be able to calculate the answers; and perhaps as important, to communicate them to the world,” Scott said.
In time, Scott brought three other Carnegie Mellon Statistics Department graduates aboard: Ashish Sanil (Ph.D. 98), Kert Viele (Ph.D. 96), and Mark Fitzgerald (Ph.D. 95). Mark, who was hired in 2013, is a Statistical Scientist with broad experience with statistical and mathematical modeling. Kert, who joined Berry Consultants in 2010, is a Director and Senior Statistical Scientist, and a software architect for FACTS (Fixed and Adaptive Clinical Trial Simulator). Ashish is a Senior Statistical Scientist and software architect responsible for the development and design of the core statistical components of the FACTS software.
“Berry Consultants specializes in the Bayesian approach to clinical trials and medical statistics, and Carnegie Mellon’s training in Bayesian statistics provided the essential skills for working in this area. “More recently, I have been working on the development of a software platform for simulating clinical trials. “I would not have been able to do this without the strong statistical computing emphasis in Carnegie Mellon’s program,” Ashish said.
Since its founding, Berry Consultants has designed more than 500 unique adaptive trials for medical device, biotech, and pharmaceutical companies, and has become an opinion leader in the science of clinical trials. Some of these trials have been groundbreaking trial designs, setting new standards for innovation and flexibility in trial design. These include the analysis supporting the first fully Bayesian approval by the Center for Drug Evaluation and Research (CDER) of the United States FDA (Pravastatin-Aspirin combination) and the statistical design for Time Magazine’s #2 Medical Breakthrough of 2007 (Veridex’s GeneSearch BLN Assay).
Berry Consultants markets software that allows pharmaceutical and drug companies to fully simulate the trial to optimize the design, frequently leading to more flexible and stronger designs. The software package, FACTS, is the most powerful clinical trial simulation tool available and, in many ways, has created a new industry of clinical trial simulation. Scott said the expertise in statistical computing that he, Ashish, Mark, Kert, and Jason learned at Carnegie Mellon provided the tools so critical to the creation and development of the software. The software is licensed to 25 percent of the top 20 pharmaceutical companies in the U.S., and more than 25 academic institutions throughout the world.
In 2013, Scott was named a Fellow of the American Statistical Association. His citation reads, “For outstanding contributions to innovative design and analysis of clinical trials, including a highly regarded textbook and an award-winning ASA short course; for excellence in the development and dissemination of Bayesian methods with applications in medicine and sports, for mentoring others in the statistical science community and for service to the ASA.”
As for future plans, he said, “We love what we do, and continue to go forward to changing the world in this way.”
The Berry statistical lineage is also continuing, as Scott’s children, Nick and Lindsay, are currently enrolled in graduate school for statistics. “It was the same kind of dinner table conversation with them. “They found it interesting, and so there was a natural gravitation to statistics like it was with me,” he said.