As the role of statisticians continues to change in a rapidly advancing digital landscape, adaptability and continued learning matter now more than ever. To this end, PROMETRIKA supports its team members in attending conferences where the most current methods and industry guidance are discussed. the annual American Statistical Association (ASA) Biopharmaceutical Section Regulatory-Industry Statistics Workshop in Rockville, MD,features representation from the pharmaceutical industry, academia, and the FDA. The most recent conference focused heavily on the evolving role of statisticians in the artificial intelligence (AI) and machine learning (ML) era. A consensus emerged regarding AI and ML, following innovative perspectives shared by several speakers. While both are viewed as powerful tools with great potential to more efficiently deliver life-saving therapies to patients, caution must be exercised in model implementation.
Extensive validation and testing of AI/ML models is increasingly necessary, especially when these models strongly influence decision-making. In alignment with this consensus, the FDA encourages early engagement and an open dialogue for any planned use of AI/ML during a study. Ultimately, the role of a biostatistician is likely to evolve as AI/ML models become prevalent in clinical trial design and execution. Increased implementation of such models will create new opportunities for the statistician to become involved in the design phase of trials. PROMETRIKA continues to adapt to the ever-evolving clinical trial landscape and, when appropriate, leverages nascent methods in doing so. Above all, we continue to provide quality services to our clients with the ultimate goal of providing quality therapies to patients globally.
In addition to many discussions centered on AI/ML, the conference featured learning opportunities via roundtables, poster presentations, and short courses on a wide variety of statistical topics. I attended a short course focused on quantitative decision-making for staging up to Phase III clinical development.
Quantitative decision-making refers to a sponsor’s internal decision-making process regarding the development of go/no-go decision criteria to proceed from one stage of drug development to the next. In the current paradigm, decision-making criteria during a Phase II trial are rarely made explicit. Decisions on whether to proceed to a Phase III trial are made with data in hand and with a trend towards statistical significance used as the rationale. Quantitative decision making explicates the criteria upon which a go/no-go decision is made. Human interpretation can still be built into the decision-making criteria, as some combination of criteria met and not met leads to a decision falling within a “gray zone” where the clinical development decision is then deferred to the relevant decision makers. Overall, the quantitative decision-making framework reduces post-hoc human bias and encourages sponsors to define what makes a successful Phase II trial. In conclusion, this short course provided an informative introduction to the quantitative decision-making framework.
Quantitative decision making is just one of a variety of frameworks through which clinical development decisions may be made. At PROMETRIKA, we pride ourselves on providing a wide variety of information and tools that may assist our clients in making informed decisions in the drug-development process. Attending conferences such as the ASA workshops allows us to grow as statisticians so that we may continue to provide quality services in an ever-changing field.
