Yodit Seifu, Victoria Chang, and Junjing Lin, ASA Biopharmaceutical Section

Margaret (Meg) Gamalo is senior director – biostatistics, global product development – inflammation and immunology at Pfizer Innovative Health. She combines expertise in biostatistics, regulatory, and adult and pediatric drug development. Previously, she was a research adviser of global statistical sciences at Eli Lilly, where she helped secure the approval of baricitinib as the first modern treatment for atopic dermatitis and position baricitinib as a first-in-indication treatment for alopecia areata. Gamalo also helped in the clinical trials for baricitinib as a treatment for COVID-19.

Prior to working at Lilly, Gamalo was a mathematical statistician for the Center for Drug Evaluation and Research at the US Food and Drug Administration supporting regulatory reviews of drugs in the infectious diseases and ophthalmology therapeutic areas. She is an expert in pediatric extrapolation and leads the Pediatric Innovation Task Force at the Biotechnology Innovation Organization. On this task force, she heads a cross-functional think tank on why extrapolation needs to be a default strategy in pediatric drug development.

Gamalo is also an active member of the European Forum for Good Clinical Practice – Children’s Medicine Working Party, working to establish decision criteria for the inclusion of adolescents in adult research.

In the statistics profession, Gamalo is a member of the executive committee of the Biopharmaceutical Section and has served in multiple administrative and scientific capacities within the section since 2014. She enjoys writing and mentors a group of statisticians in research activities on topics related to Bayesian methods, evidence synthesis, causal estimation for RWD/RWE, and policy-oriented work on pediatric drug development.

BIOP Report: Congratulations on your recent appointment as the editor-in-chief of the Journal of Biopharmaceutical Statistics (JBS). What is your vision on its mission, history, and current status?

MG: The Journal of Biopharmaceutical Statistics (JBS) has been around for about three decades. It started back in the early 1990s and, since that time, you can see how the industry has evolved over the years told through the collective perspective of statisticians. It is fascinating to study the progression of scientific research published in the journal reflecting both the scientific and evolving landscape of drug development. For example, the first two decades exhibited statistical issues ranging from proper use of one-sided vs. two-sided tests in hypothesis testing, appropriate analysis for stability and expiry, bioequivalence, noninferiority and superiority, subgroups and multicenter trials, interim analysis and adaptive designs, multiplicity, linear models analysis of endpoints, meta-analysis, among others for which the journal has provided the stage for consensus and application.

Since then, the journal has seen a transformation in scientific topics influenced by shifts in research and development and the growing number of stakeholders. For example, a big trend has been the move to a more patient-centric health care model, closely followed by the impact that technology has on all areas of the life sciences and the changing business models. Most recently, many of the published manuscripts revolve on topics such as ecient trial designs, including the use of external data, estimation and uncertainty for go-no-go decisions, finding optimal individualized treatment rules or biomarker-guided treatments, treatment heterogeneity, and considerations for regional and payer evaluation of multi-regional clinical trials, among others. Some proposed solutions apply an agile, iterative test-and-learn approach, rather than running long and expensive development processes to concoct the perfect solution. Other solutions often require modern technological innovation such as artificial intelligence (AI) and machine learning (ML), which provide significant opportunities to enhance drug discovery, clinical development, and commercialization.

When I wrote my editorial in December 2019 in time for the first issue under my editorship, I pondered the role of scientific publishing in an evolving industry. I realized that while our problems and insights today are much different from when the journal was first envisioned, not much has changed in terms of its vision and mission. The journal remains committed to the principle that better education of statisticians enables informed debate and decision-making about the valid application of new methodologies and aids in addressing misuse of statistical concepts (e.g., p-values) for scientific discoveries.

The journal will continue to strive to achieve excellence by publishing articles that present important new advances in an everchanging field of statistics within the pharmaceutical industry. JBS will always be a platform for education and dissemination of statistical research and innovation, a repository for statistical solutions and choices, and a forum for scientific opinion on issues impacting methodology and applications in the pharmaceutical industry.

BIOP Report: There are several statistical journals with a focus on statistics issues in drug development and many more with a broad focus on medical statistics and biostatistics. In your mind, what are the competing journals for JBS? What is your view on the relationship between JBS and these competing journals? What kind of journal would you like to see JBS become in the next few years?

MG: There are two ways of thinking about the existence of other journals (i.e., whether to look at them as competitors or whether their existence complements the journal and enlarges the collective influence of statistics in the biopharmaceutical industry by broadening readership and impact of scientific publication). I prefer the latter, embracing the perspective of abundance (i.e., there is an abundance of scientific thought and research material for everyone). This, I think, is the better long-term strategy for scientific publication and consistent with how science has evolved throughout history.

While Galileo Galilei may have stated, “In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual,” we all know science is also strengthened through constant validation—the reproducibility of results, the conduct of peer-reviewed, open literature research. Hence, the existence of other journals is also essential for quality scientific and statistical research.

Furthermore, science and statistics moves and grows with collaboration. There is more and more to know in the world, and one can only have so much in our heads. In fact, the share of stuff we know as individuals is declining in any field. Inevitably, we need to collaborate as most research problems require multiple kinds of expertise.

Collaboration in science and statistics is good for making bold advances (i.e., innovation). Interaction with people with different perspectives or approaches prevents us from getting tunnel vision. Hence, one objective in scientific publishing should be to collectively grow the biopharmaceutical statistics field together through a feedback loop of continuous scientific and statistical innovation.

Of note, innovation is key because it is multiplicative, meaning that the same input generates greater output far beyond the biopharmaceutical statistics field. The real problem, then, in scientific publication is sustaining statistical innovation, which I believe is anchored on drivers of innovation (i.e., problems, constraints, and opportunities).

While the biopharmaceutical industry may be a mature field with established and highly regulated scientific and statistical problems, our problems can be unique given the interface of different stakeholders (e.g., payor, regulator, patients, physicians, pharmaceutical companies). Capitalizing on this uniqueness to formulate innovative scientific and statistical solutions could have far-reaching implications on other fields.

With this paradigm in mind, the journal will strive to promote multidisciplinary collaborative research as a logical response to the expanse and pace of scientific revolution and transformation of the field. Not one discipline will be able to integrate all aspects of any problem or issue of interest alone with proper applicability and sufficient viability or competitiveness across a multi-stake holder industry. Increased focus on statistical research that has good cross-functional appeal apart from just pure statistical interest is important. The journal believes that effective science is achieved not just by one discipline, but by the collaborative and multidisciplinary effort within the entire biopharmaceutical field.

BIOP Report: What are the new measures you are taking to continue improving the quality and visibility of the journal and make it truly impactful?

MG: One of the initiatives I instituted when I assumed the role of editor-in-chief was to ensure diversity in the editorial board. We have increased the number of women associate editors (AEs), and currently there is a diverse pool of AEs representing pharmaceutical companies, academia, and regulatory agencies from various geographies. I realized that a journal relies on multiple and varied voices having a wide range of experiences. In fact, a diverse and inclusive editorial board brings the different perspectives a journal needs to ensure quality and unleash value-driving insights, methods, and practices. Another consequence of a diverse pool of AEs is having a network to a broader pool of peer reviewers who are much more engaged.

We screen papers for appropriateness to be published in the journal (i.e., papers that fit the aims and goal of the journal). We identify novel methods and applications and aspire to ensure relevance of topics while maintaining scientific integrity. We also plan special issues to address questions or bring up more discussion about new and hot topics and weave manuscripts coherently through different viewpoints or perspectives.

Finally, we rely on our reviewers to provide quality scientific and statistical reviews.

Diversity and inclusion must happen, even in scientific publishing. Science and statistics will not meet its potential until the research culture enables and supports contributors from all backgrounds and circumstances and contributions of all kinds based on the interests, skills, and resources available. Failure to achieve diversity and inclusion of all stakeholders in science and statistics will slow progress in discovery and translation of knowledge to solving humanity’s most pressing problems.

Another initiative we are accessing currently is to harness the power of the crowd by highlighting key innovations or discussions in social media. I think as a leader in scientific publication, we need to be more proactive in disseminating information than merely passive repositories of scientific thinking brought to use through a Google search. It is estimated that ~68% of Americans get their news from social media.

The ease of use of social media platforms for communicating and disseminating information also makes them attractive to scientists. Furthermore, I believe social media gives us the opportunity to engage directly with a wide range of audiences and helps us understand our readers.

Within the next five years, we have several other initiatives planned, as well (e.g., partnering with a statistics professional organization on proceedings or narrowing the proceedings on hot topics or breakthroughs). There will be a few more changes along the way as we try to think about what the role of scientific publishing in a personalized information age means. How can we adapt to that environment and perhaps influence it as well, particularly in biopharmaceutical statistics? I am open to suggestions, and anyone is free to reach out with ideas on how we can improve.

BIOP Report: JBS has successfully published several special issues in the past and attracted much attention. Do you plan to publish more special issues in the coming years, and what are the topics of these special issues?

MG: We recently launched a call for papers for special issues on real-world data/evidence (RWD/RWE) and in the implementation of estimands in clinical research. RWD/RWE is a growing area that still requires thought and consensus on how it can be applied, given the range of possibilities. In the future, there will be a blurring of how evidence of effectiveness of investigational new drugs will be established. Moreover, this field will continue to grow in the next decade, finding interconnections with clinical trials, clinical practice through electronic health records, and digital health.

Estimands, on the other hand, is a complex issue that is quite difficult to explore upfront and implement in the planning stage of a clinical trial. However, currently, it is changing the way we design trials, write the objectives, collect the data, conduct the trial, and perform analysis because the framework requires us to be more unambiguous about the questions we would like to answer. The complexity is also due to the presence of multiple scientific questions of interest about relevant treatment effects, interpretation of study results, and added value of drugs to different stakeholders (i.e., regulatory, prescribers, patients, and payers).

Our guest editors are busy connecting with key scientific leaders in this field, as well as disseminating the effort through different social media venues. What we believe to be important is to have a balanced and informative issue that will serve as a definitive reference for these topics to many scientists and statisticians in our field.

For these special issues, the reviews will be rolling (i.e., once the reviews are completed, the accepted manuscript will be posted on the webpage right away for access). Once a sufficient number of manuscripts is collected and reviewed, we will close the issue and publish it in print. I urge those who are interested to reach out to Junjing Lin (Takeda), Helen Qi (Bristol Myers Squibb), Yodit Seifu (Merck), or Bill Wang (Merck).

We will also launch another special issue on pediatrics very soon. Of note, the majority of the investigational drugs being studied in adults will [be studied in] pediatric patients, as well—either through a requirement or in pursuit of an incentive. The challenges of running trials in children are accelerating efforts in innovative trial design and analysis. We hope this special issue will provide a simple but comprehensive guide for statisticians/clinical research scientists to determine the extent of development in a pediatric trial in accordance with the principles of extrapolation and how these trials can be streamlined to be as lean as possible to ensure they provide the maximum information with the minimum number of pediatric patients exposed to research risk.

We are trying to have this special issue coincide with a global virtual workshop on extrapolation in pediatric drug development. This workshop and subsequent special is still in the planning stages, but will focus on statisticians’ and clinicians’ experiences in pediatric drug development.

BIOP Report: As far as we know, many statistical journals are facing challenges in finding highly qualified reviewers to complete reviews in a short timeframe—say one and a half months. What do you think JBS can do to address issues such as delayed manuscript review?

MG: Indeed, this is a big problem in scientific publishing. Our managing editor, Victoria Chang (BeiGene), has been excellent in reminding AEs when the reviews are needed. We have been able to manage asking the reviewers to actively turn in their reviews. Of course, there are some review slips here and there as this is all voluntary work. For special issues, we ask the guest editors to have their own system of ensuring expedient reviews by having a standby review committee.

As I have mentioned earlier, having a diverse set of associate editors has been very helpful. We still have plans to expand our editorial board with folks from the European Union and Asia. In fact, if you have an interest in serving as an AE and you have the passion to provide service to biopharmaceutical statistics and the society, please reach out to us.

We also encourage young scientists to take part in this endeavor actively. It does not require that one must be well experienced to serve as a reviewer. I think the major criteria to be a good reviewer are curiosity, critical thinking, and the ability to ask good questions.

I am aware that many of us will say we are eyeballs deep with work. On the other hand, I argue that we need to take care of ourselves, as well. One way of doing that is ensuring we retain our scientific and statistical thinking and keeping up to date on innovation and new statistical techniques. I always am reminded that we may need to keep learning if we want to be relevant and as we live longer. Perhaps Mahatma Gandhi was right in saying that we need to “learn as if you were to live forever.”

BIOP Report: Do you have any advice for the statisticians who would like to submit manuscripts to JBS?

MG: My general advice to all statisticians, and not just to those who would like to submit manuscripts, is be curious and tell a good story. Most breakthrough discoveries started with curiosity—the impulse to seek new information and experiences and explore novel possibilities. Curiosity is beneficial for all because it cultivates many levels, whether it is one’s organization or, more broadly, the society. In fact, curiosity helps society make better decisions. When we are curious, we think deeply and rationally about decisions and come up with more creative solutions.

Curiosity does not necessarily have to result in a monumental breakthrough, and certainly publishing in a scientific journal does not necessarily mean only novel solutions are entertained. Science moves by increments, not by leaps and bounds. Sometimes, the insight to the problem is enough.

What I also see as a problem is that there are times when statisticians hold back on their ideas, fearing that they may be too obvious. I think great ideas may sometimes seem obvious because the solution has all the parts of the question lining up and shedding light on a solution. Therefore, ‘obvious’ answers are not visible to most people, partly because most people are not thinking about the question. Ideas only come to those who recognize a problem and look for innovative solutions. I posit that even Einstein could possibly not find a solution if he had the wrong question.

Inside every scientific discovery, there is a good story. I think it is important to share that story, as I am sure there are a lot of insights that went through (e.g., how the problem came about, why it is an important problem to pursue, how the solution was discovered, why other solutions failed). These experiences are actually very informative and could help many researchers out there know or understand what works and what does not.

Brilliant statisticians may sometimes be dissuaded by writing not because they do not know how to write but by not trying. We are our own limit. We need to believe that something different can happen in order to break old patterns, and we can choose that new outlook at any time. Part of being an effective statistician is not only to develop or apply sophisticated numerical calculations, but also being an effective communicator in writing and in speaking.

BIOP Report: As we know, you had extensive working experience in government and the pharmaceutical industry. What is the impact of this unique experience on your perspectives on the role of statistics and statisticians in clinical trials research?

MG: I learned the principles of drug development at the FDA, and I learned how to apply them while understanding the challenges of drug development in industry. I realized many of the regulations in clinical trials are common sense and centered on ensuring the safety of patients. Having reviewed hundreds of investigational new drugs (INDs) and new drug applications (NDAs), what is the right thing to do is sometimes very easy to spot because it is rational. I can also see the difficulty with implementing mitigations from the industry side for some of the concerns raised by regulatory agencies.

I learned to assess what is ideal and what is applicable. In the case of the latter, there is no perfect solution most of the time, but the quality of the medicinal product and patient safety are paramount. Hence, when I think about my job responsibilities and the role of a statistician in the biopharmaceutical industry, it may just be encapsulated by the provisions of Title 21 of the Code of Federal Regulations, which is consistent with ensuring good clinical practice. Of note, good clinical practice recognizes that protecting data integrity is part and parcel of ensuring safety of patients.

More broadly, I think statisticians need to be involved as key decision-makers. When we can understand and interpret data correctly, our ability to identify crucial areas requiring attention in drug development are enhanced and our proposals for mitigating these key areas are likely to respond to the needs of our organization or the industry. In an age where data is essential for making big decisions, whether in business or government, statisticians need to be at the table so we can assist and encourage informed decision-making.

However, this also entails that we need to be able to communicate in the language that is understandable by nonstatisticians. Statisticians may need to be comfortable communicating about the problem not just in terms of numbers. We need to understand the whole problem and not just numerical ramifications (e.g., scientific, clinical, regulatory, payor, etc.) Having a holistic view is what we need, so we can provide more valuable and insightful feedback.

My experience on both sides of the industry (regulatory and pharma) has also shaped my collaboration with many statisticians in the industry. I have been more discerning on what topics are most impactful, and so I think statisticians need to influence scientific thinking and progress in the biopharmaceutical industry. Particularly, I learned how to think big, start small, and learn fast—our role in the industry is to have a broad vision while being mindful of how we act on it.

For more than five years, I have been using a great deal of Bayesian methodology. However, I realized that most sample sizes are driven by the number of exposures needed to have sufficient data to establish safety. Hence the value of Bayesian methodology in terms of efficiency in late phases of development may not be apparent to encourage a strong push for change. However, in areas of unmet need and in pediatrics, the use of Bayesian methodology is clear because of feasibility and because of ethical principles of not having duplicative information to warrant translating conclusions from one population to another. That situation gave me a better perspective to focus on what innovative advances can bring meaningful change in policy. In fact, most of these innovative methods have been expanded to applications of propensity scoring methods to augment clinical trials, particularly in pediatrics, orphan diseases, and unmet medical need indications. So even in scientific research and policy, the words of Justice Ruth Bader-Ginsburg reverberate: “Real change, enduring change, happens one step at a time.”

BIOP Report: The COVID-19 pandemic is having a significant and long-lasting impact on how clinical trials are conducted. Meanwhile, government, universities, and many pharmaceutical companies are working together to find vaccines and new treatments for this disease. Do you have a plan to use the journal as a venue to promote the discussion on challenging issues arising in clinical trial designs and analyses?

MG: Statistics in Biopharmaceutical Research (SBR) is already having a series of special issues on the impact of COVID-19 on clinical trials and in COVID-19–related research. Some of my friends and colleagues are already working hard on that area and I am amazed at the speed of coordination and implementation. Two of the manuscripts I have been involved in writing will be published in that endeavor.

Consistent with the spirit of collaboration I mentioned previously, I decided not to go with another special issue on COVID-19 in JBS because it would then be in competition with the SBR effort. Instead, any COVID-19–related research identified as helpful to the scientific community or related research activities will be given priority and expedited review. This allows us to publish any research and findings with greater speed and agility.

The main issue with developing drugs for COVID-19 is speed of innovation. However, many of the tools for acceleration have been discussed extensively in literature (e.g., adaptive design, data sharing, etc.) What is lacking, from a statistical perspective, is on knowledge of appropriate endpoints in relation to patient population. Hence, COVID-19 disease progression models are needed to learn about how to conduct COVID-19 treatment clinical trials. Ensuring clinical trials have a common set of data that can help inform other development is also important. However, there is little data available to assess how this can impact speed of development. The best would be when data is already out there, what can we learn from it so we can be better prepared should there be another catastrophic event of similar nature in the future?

I do encourage statisticians to contribute to the scientific efforts for developing treatments for COVID-19. I think it is a worthwhile endeavor and reminds us how interconnected we are. If we do not collaborate, we will be in this situation for a long time.

I believe this is the best time for science and statistics. In fact, as I have mentioned previously, with uncertainty comes great innovation. Problems and constraints are backdrops for opportunity. As a cheery reminder, Sir Isaac Newton produced an unbelievable number of exceptional results, including seminal experiments on the law of universal gravitation, while quarantined during the London plague of 1665–1666, though I believe he must have been curious and persistent even before that.