From unlocking new insights to making clinical developments more relevant to people across the world, advanced statistical methods are transforming clinical trials.
We interviewed four of the women leading innovation here at Phastar, Hebing Wang (Principal Statistician, China), Samantha Hinsley (Statistics Manager, UK), Helen Marshall (Principal Statistician, UK), and Vicky Marriott (Head of Statistics and Data Science, UK).
Below, they discuss the developments they are most excited about, their own groundbreaking work and how we can encourage more women to pursue a career in statistics.
How are advanced statistical methods helping to drive innovation and improve outcomes in clinical trial research?
Hebing: Advanced statistics are helping clinical trials move beyond a one-size-fits-all approach. In global studies, it is increasingly important to understand not only whether a treatment works overall, but also whether the evidence is consistent and meaningful across different regions and patient populations.
As the statistical representative for China, I often work at the intersection of global study objectives and local clinical research considerations. For example, regional sample size planning can help ensure that a study is appropriately designed to evaluate the consistency of treatment effects in China and other key regions, while avoiding unnecessary recruitment. I have also worked on interim analysis sample size re-estimation using a Bayesian approach combined with conditional power. This enables the study team to learn from accumulating data and make better-informed planning decisions while maintaining statistical rigor.
For me, the value of these methods lies not only in their technical sophistication, but in their ability to make clinical development more efficient, more inclusive and more relevant to patients in different parts of the world.
Samantha: Following from Hebing’s point that advanced methods help us move away from the one-size-fits-all approach; in early phase we are seeing an almost constant rise in new methodology to ensure we keep up with the changing clinical landscape. Whether we are investigating a new type of radiotherapy, a virus, or a completely new type of agent, the study design must meet the needs of the study and all its nuances. There are various considerations in the early phase design space, with recent advances moving away from solely considering safety, to thinking about the impact of efficacy, biomarkers, PK, PD and more. Ultimately, the more well designed a study, the more likely we reach the correct answer.
Helen: Real-world evidence (RWE) studies are increasingly used to support regulatory submissions and decision-making. However, real-world data (RWD) can bring about statistical challenges generally not observed in prospective, randomized interventional studies. RWD can come from a variety of sources resulting in different observation periods between and within endpoints. This may impact, for example, on the definition of time-to-event endpoints and choice of censoring dates. In addition, data defining patient eligibility or subgroup identification may only be observed during follow-up, rather than at the start, which can introduce a different type of bias called immortal-time bias. Understanding the selected data sources, their limitations and potential challenges and applying appropriate solutions and alternative statistical methodology, are vital to ensure statistically robust RWE studies can be designed and analyzed to effectively support practice-changing regulatory submissions.
Which recent developments in statistics are you most excited about, and why?
Hebing: I am particularly excited by the growing use of adaptive and Bayesian methods in clinical research. These approaches allow us to learn from the data as a study progresses and, when appropriately planned, make clinical trials more flexible without compromising their scientific integrity.
My interest is also closely connected to the increasingly important role of China in global clinical development. China has a large and diverse patient population, a rapidly developing clinical research environment and growing experience in contributing to international studies. This creates both opportunities and challenges for statisticians. We need to ensure that global trial designs are scientifically sound while also considering regional differences in disease characteristics, treatment patterns and recruitment feasibility.
I believe the future of statistics will involve not only developing more sophisticated methods but also applying them thoughtfully across different regions. The most valuable methodology is the methodology that helps generate reliable evidence for patients wherever they live.
Samantha: There are lots of different methodologies for early phase that continue to appear year on year. I am particularly excited to see the ones that have been designed with application in mind. I want to know that the design is robust, reliable, and useable in practice, not just statistically interesting.
What practical steps can individuals, organizations, and the wider industry take to make the field of statistics more inclusive?
Vicky: Creating a more inclusive profession requires action at every level. For individuals, it starts with being conscious of biases, advocating for colleagues, and helping to create environments where everyone feels comfortable contributing their ideas. Small actions, such as inviting quieter voices into discussions or supporting a colleague’s development, can have a significant impact.
For organizations, inclusivity must be built into recruitment, development, and leadership practices. This includes ensuring diverse candidate pipelines, providing equitable access to training and career opportunities, and offering flexible working arrangements. The wider industry also has an important role to play by showcasing diverse role models, increasing outreach to schools and universities, and raising awareness of the exciting careers available, so that a more diverse range of future talent are encouraged into the profession. These are all things that organizations like PSI (Statisticians in the Pharmaceutical Industry) aspire to do, and I’d encourage leaders across the industry to recognize the value of allowing staff to support such initiatives.
Can you share any notable use cases, publications, research projects, or advisory roles that highlight the impact of your work?
Hebing: I contributed to a conference abstract and poster analyzing pooled data based on two phase III trials which explored treatment responses to secukinumab [1] a selective interleukin-17A inhibitor, in patients with hidradenitis suppurativa. The analysis examined how disease duration and disease severity influenced clinical response and highlighted the potential value of earlier treatment. This experience showed me how statistical analysis can help uncover clinically-meaningful insights from trial data and communicate them in a way that is relevant to patients and healthcare professionals.
More recently, as the statistical representative for China, I have supported global clinical development activities involving regional sample size planning, assessment of regional consistency and interim analysis sample size re-estimation. These experiences have given me the opportunity to connect global statistical strategy with local scientific and operational considerations. I find this particularly rewarding because it allows statistical work to contribute not only to robust study design, but also to ensuring that patients and clinical research teams in China are meaningfully represented in global evidence generation.
Samantha: An interesting project I worked on a few years ago was the development of a first-of-its-kind phase I platform using time-to-event methodology and randomization to control or a variety of treatment arms. It was hugely rewarding to see the study take off after many hours of discussions with medics, radiotherapists, trialists, funders and more!
I have been an author of some really interesting early phase papers working with other experts in the field. These include an extension to guidelines for the content of SAPs for early phase studies [2], a roadmap for the design of phase I trials of radiotherapy and novel agents [3] and a discussion of the practicalities of using time-to-event methodology in early phase trials [4].
As the lead for the PSI early phase special interest group, I hosted a session at the PSI 2026 conference updating the statistical community on some of the recent advances in early phase trials methodology. As a group we have plans for further training and advice, as well as publications to help further innovation across early phase studies.
Helen: A few years ago, I was the lead statistician on a highly impactful, “landmark” study, results of which were eagerly awaited at the international San Antonio Breast Cancer Symposium – the anticipation and buzz in the air was palpable! The study was investigating the addition of a bisphosphonate, zoledronic acid, to standard adjuvant therapy in early breast cancer.
Although the results of the overall study population were not statistically significant, a pre-specified subgroup analysis revealed a highly significant treatment benefit in post-menopausal women. These milestone findings were later published in the New England Journal of Medicine [5] and were followed by several other publications in high-impact journals such as the Lancet Oncology and Journal of Clinical Oncology. A collaborative meta-analysis also ensued and was published in the Lancet [6], confirming “clear evidence of benefit” in post-menopausal women – a finding that fundamentally shifted international clinical guidelines.
How do you support and champion women pursuing careers in statistics?
Samantha: I am currently an industrial supervisor for a fantastic MSc student and will be attending an event at the university aimed at promoting careers in statistics for women. The event will be made up of a panel of women working to give the students an idea of potential career pathways, and examples of how women can thrive in the world of statistics.
Vicky: I believe one of the most important ways to support women in statistics and data science is to make the path forward visible. I try to ensure access to opportunities that allow women within my team to grow, build confidence, and demonstrate their expertise. This includes actively encouraging women to take on leadership roles and high-profile projects, and creating an environment where diverse perspectives are valued and heard. Mentorship is also incredibly important. Sometimes all it takes is someone encouraging you to put yourself forward for a new opportunity or reassuring you that you are ready for the next step.
Leaders have a responsibility to ensure that different skill sets and career pathways are nurtured and recognised fairly. By creating inclusive teams and inviting women into different conversations, we can help more women thrive and progress within the field.
What advice would you give to women looking to start, develop, or advance their careers in statistics?
Vicky: Don’t be afraid to take on new challenges, even if you don’t feel completely ready. Career growth often happens when we step outside our comfort zone. Seek out mentors and sponsors who can support your development, build a strong professional network, and look for opportunities to broaden your skills beyond technical expertise, including communication, leadership, and collaboration.
Most importantly, remember that there is no single route to success. Embrace opportunities as they arise, remain adaptable, and trust in the value you bring. The profession needs diverse perspectives and experiences, and your contribution can make a real difference.
References
- Abstracts of the 10th Annual Symposium on Hidradenitis Suppurativa Advances 2025 | Dermatology and Therapy | Springer Nature Link
- https://www.bmj.com/content/bmj/376/bmj-2021-068177.full.pdf
- https://aacrjournals.org/clincancerres/article/28/17/3639/708074
- https://link.springer.com/article/10.1186/s12874-020-01012-z
- https://www.nejm.org/doi/full/10.1056/NEJMoa1105195
- https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(15)60908-4/fulltext
