banner_data numbers

Biostatisticians bring incredible skill and knowledge to the drug development life cycle.  The statistical strategies and analyses they provide uncover keen insights and trends that enable more productive and efficient development paths. This field continues to generate new insights and methodologies every day that can directly impact on the quality and efficacy of the clinical trial process. However, these innovations don’t always come to the fore, because biostatisticians often operated in isolated niches where their voices aren’t heard and their expertise goes untapped. This is causing the industry to miss out on badly needed opportunities to improve decision-making and cut time and cost from these trials.

If the industry is going to make the clinical trial process more efficient, biostatisticians need let their ideas be heard. When they take on greater leadership roles they are able to drive development of novel trial designs, harness data from new sources, and enrich the information-gathering process across the development program lifecycle.

These were the overall themes covered at the first annual Innovation and Collaboration Workshop organized by the Center for Quantitative Decision Strategies and Analytics (now Advisory Services Analytics) at Quintiles Advisory Services in October. The event, held in Raleigh, North Carolina, brought together academics and executives in biostatistics and biometrics from across biopharma and academia to discuss new solutions for analyzing trial statistics, leading-edge applications for modeling & simulation to improve decision-making, and opportunities for biopharma to educate and mentor the next generation of biostatistics leaders.

And at the heart of every presentation was the idea that the most successful innovations are driven by biostatisticians themselves.

Gaining agility and hedging your bets

Some of the highlights of the event included José Pinheiro’s presentation on how a model based drug development framework can drive innovations in drug development. The head of Statistical Modeling for Janssen argued that modeling and simulation (M&S) gives researchers greater agility in the trial environment by providing a framework for combining assumptions, prior information, and observed data to explore alternative scenarios and evaluate uncertainty. “M&S is the cornerstone of modern development program and trial design,” he said.

Similarly, Stephen Ruberg, distinguished research fellow and the head of advanced analytics at Eli Lilly, discussed how hedging statistical bets can help researchers use initial results to create decision rules based on the probability of success. “There is amazing resistance to the benefits of spending a little extra money to hedge your bets,” he said. But when companies make these investments upfront, they can speed the development process by identifying the best path to follow early on. He shared three examples of how this approach can be used to compute clinical meaningful effect much earlier in the research – after the first cohort of patients complete the first trial – potentially shortening the overall duration of the program by months. “It’s all about changing the way you make decisions,” he said.

Catherine Munera, head of biometrics for Cara Therapeutics talked about how analytics strategies can be used to predict responses in long term pain research studies. Pain research efforts face multiple obstacles, including high dropout rates, treatment abuse, and highly individualized need, but analytics can lessen this risk through the review of biomarkers, patient survey data, and previous trials to see who stayed in and who dropped out, in order to hone trial design and reduce abuse and attrition. “All of this data exists,” she said. “The question we should be asking is ‘what else can we do with it?’” 

Speak up!

Nicky Best, head of the statistical innovation group at GSK, talked about how the use of prior elicitation methods enable uncertainty in trials to be captured and communicated in a more rigorous way. She argued that biopharma companies should be taking greater advantage of this prior knowledge, which exists in every project, to inform their research. Biostatisticians at GSK are already leading several efforts to translate some of this information into quantitative prior distributions, which include determining assurance by looking at predictive probability of success for future studies, informing clinical trial design through strategically staggered investments, and drawing statistical inference through the analysis of existing study data. She also pointed out the importance of bringing in outside experts to reduce the risk of bias in the data analysis process. “Having trained facilitators involved in this process is key.”

That lead to the other broad theme that was frequently touched on at the event: that the potential contributions on biostatistics experts are often overlooked in the broader drug development environment, in large part because biostatistical don’t always take the initiative to bring their ideas to the table, but they must speak up if they want their ideas to be taken seriously, argued Sonia Davis, professor in the Collaborative Studies Coordinating Center at the University of North Carolina. “There is much more to a trial than planning and results,” Davis said. If a clinical trial is not implemented well patient safety can be at risk, the plan’ can be compromised, and the cost and timeline can be increased. Her solution: Statisticians need to be involved in all stages of the study, including protocol design, data collection, analysis and reporting, and regulatory submissions. “We have the needed analytical skills to help protect the plan and reduce risks,” she said. “Our teams need us to be informed, engaged, proactive, and influential.”

Pandurang Kulkarni, vice president and head of biometrics from Lilly, talked at length about the impact analytics can have across the organization, and the importance of applying qualitative thinking at every phase of the drug development lifecycle, and for biostatisticians to reach out to other business leaders and help them understand how analytics can improve their performance outcomes. “There are many opportunities for statistical innovation in preclinical, marketing and manufacturing,” he said. “Statistical leaders need to be bold and explore these areas of influence.” When these experts take a more proactive role, it will benefit the company, the individuals, and the overall healthcare community.