by Brigitte Scott PhD in discussion with the StatisticaMedica team
Where does biostatistics fit in clinical research, what makes a biostatistician tick and what do you they do all day?
Read on to find out what the StatisticaMedica team had to say.
The importance of biostatistics in clinical research
“Dedicated biostatisticians ensure the rigor of the clinical research data, thereby helping the progress of the drug or device to the clinic, where these products can make a difference to patients’ lives.”
Biostatistics is vital in clinical research to ensure that the evidence used to support the claims for drugs and devices is statistically valid. This is an essential step in the drug/device development pathway. Biostatisticians choose the statistical methods carefully and implement the statistical plan to generate a series of evidence-based results, from which realistic conclusions can be drawn about the effectiveness and safety of the drug/device. If the statistical methods are not appropriate, the drug/device may not be as valid as it appears. If the results are too broad, no proper conclusions can be made.
There are many small steps involved in biostatistical analysis in clinical research and each step is there for a reason: to maximise the quality of the analysis and provide clean and reliable data.
“The conclusions that are drawn on the clinical study data are only as good as the methods used to draw those conclusions.”
The changing landscape of biostatistics in clinical research
The landscape of biostatistics in clinical research is changing in the era of electronic data capture and digital endpoints:
“Digital endpoints are changing the landscape of biostatistics in clinical research as measurements can be taken over hours and days rather than at a single timepoint, and devices can take more than one type of measurement. The clinical data are changing; therefore, biostatistics is changing in parallel.”
Is there a typical day in the life of a biostatistician?
The answer to this question was a resounding, and unsurprising, no. A morning meeting is a consistent fixture at StatisticaMedica to plan projects and review the current state of the data evidence to be managed. Beyond that, each day is very much dependent on the current projects, deadlines and priorities rather than having a predictable routine. However, within this variation is a typical framework in which the biostatistician always needs to dive into the data, ask questions and programme and validate the statistical analysis. Furthermore, there is a standard project process that is followed, where applicable, including:
· Drafting of the statistical sections of the study protocol
· Production of the statistical analysis plan
· Programming and validation of statistical methods
· Production of mock tables, figures and listings
· Development of the study data tabulation models
· Draft and final tables, figures and listings
· Writing of statistical reports
· Input into the clinical study report (statistical methods and results)
Beyond this typical framework, StatisticaMedica also provides a statistical consulting service, to support the strategic assessment of the regulatory requirements of a project. In this role of regulatory biostatisticians, the StatisticaMedica team advises the client on how to address the statistical regulatory issues of a project and provides solutions on how to meet regulatory requirements without compromising the highest standard of data quality.
The scale and complexity of regulatory biostatistics cannot be underestimated. StatisticaMedica biostatisticians develop a sophisticated knowledge of the regulatory landscape, they keep themselves updated on new guidelines, and upskill quickly when necessary. These biostatisticians learn how to share their knowledge of the complex biostatistical regulatory landscape with the client in a clear and concise manner, so that, together, they can apply this knowledge effectively to resolve a new issue.
“Every day is different. Every project has its quirks and requires a unique thought process. We use and develop lots of different skills in any one day – programming, writing, taking client meetings, quality control, problem-solving – this keeps everything fresh.”
The involvement of biostatisticians in the clinical research process
Biostatisticians might only be involved in the clinical research process after the study is almost complete and the data are ready for analysis. Ideally, biostatisticians should be involved much earlier in the process, at the study design stage, to ensure that all aspects of the study are statistically valid from the beginning. Involving the biostatistician early in the study process will influence how the data are collected and maximise the strength of the data.
“The sponsor will have ideas about objectives, endpoints and statistical methods for the study. The role of the biostatistician is to ensure that these are statistically meaningful and relevant, while understanding the requirements for clinically meaningful and relevant results in an achievable timeframe.”
How biostatisticians contribute to clinical study design
Biostatisticians can contribute to decisions on clinical study design in many different areas:
· Endpoints
· Sample size
· Timepoints for analysis (e.g., change from baseline, value at week 24)
· Variables
· Analysis methods (how/what/when/order of analysis)
· Setting the hypothesis
· Setting the p-value
· Data collection method (paper, electronically)
· Statistical regulatory requirements
“The biostatistician adds value throughout the study design process and helps ensure the rigor and quality of the resulting clinical study data.”
How biostatisticians ensure data quality
Biostatisticians ensure data quality through the following activities:
· Adhering to regulations, guidelines and standard operating procedures
· Following statistical and validation steps, procedures and rules
· Independent double programming
· Statistical review of data output by an independent biostatistician
· Peer review of data and outputs
· Keeping informed about guidelines and guideline updates
“Peer review of data is a key component of StatisticaMedica’s quality control processes. This review is conducted to check consistency and highlight any suspicious trends, which are then investigated.”
Challenges for the biostatistician
Poorly structured data is a challenge for the biostatistician because it can be difficult to visualise the steps needed to get the data into shape. Breaking down the steps in the statistical process to provide a workable dataset is an effective approach. Taking the time to understand the real context of the data (e.g., the measurement/disease/patient population/timepoint) will also help define the structure and interpretation of the data.
The non-statistical challenges that biostatisticians face are universal: planning and organisational challenges, deadlines and time management. These challenges can be overcome by planning in advance and allowing enough time for each step in the statistical analysis process.
“Communication is key. If you explain your challenge to others, they may help you to think in a different way and take you a step closer to the solution.”
Reasons to become a biostatistician
Common reasons to become a biostatistician include:
· Staying in scientific research
· Working in an area with a tangible impact
· Blending pharma and maths
· Putting statistics into context – the ‘bio’ part of biostatistics
· Caring and curiosity about patient data
· Always learning and extracting something useful
· Developing soft skills as well as technical skills
“Statistics applies to everything. Biostatistics means you can develop an interest in pharma while keeping the maths.”
The qualifications, strengths and qualities needed to be a biostatistician
The following list is a snapshot of what is needed to be a biostatistician:
· A STEM* background and experience of working with data are a good foundation
· Biostatisticians are detail-orientated, problem-solving, lateral thinkers who are confident exploring data in different ways
· A biostatistician needs to be adaptable, versatile, flexible, vigilant, responsible, accountable and patient
· Further desirable qualities include an ability to learn quickly, a strong interest in drug development, an understanding of industry standards and guidelines, and an openness to constructive criticism
· A foundation in soft skills, such as communication, teamwork and adaptability, is advantageous
“There are a range of qualities needed to be a biostatistician, including a genuine understanding that there are real people behind the numbers.”
*Science, technology, engineering and mathematics
Enjoyable aspects of being a biostatistician in clinical research
The StatisticaMedica biostatisticians enjoy their work in clinical research for a variety of reasons. Here are a few examples:
· Learning about a range of clinical studies and therapeutic areas
· Exposure to many different statistical methods
· Seeing mistakes as a learning opportunity and creating solutions
· Working methodically, taking all the steps in the process and seeing the reward in the final output
· Moving up a steep learning curve, and seeing other projects benefit from that learning
· Ensuring high quality data is a task that aligns with personal logic and standards
· Attending congresses and webinars to keep up to date with developments in biostatistics in clinical research, particularly in the era of electronic data capture, digital endpoints and artificial intelligence
“Being a biostatistician feels like you are contributing to something bigger than yourself – a greater purpose. The work we do helps clinicians interpret the data and find out what the results tell us about real people taking part in clinical research.”