This industry is poised on the brink of a new era in clinical development in which access to global data, advanced analytics and trial execution could help us fix the problems that have plagued clinical research for years.

And it’s about time.

Every year, drug developers spend lots of money on the clinical research process because they don’t have the tools or access to global data to support evidence-based decision-making. That forces them to make decisions based on limited information that may or may not be accurate.

Consider protocol amendments. According to the Tufts Center for the Study of Drug Development, a single protocol amendment in a phase IIII trial costs an average of $535,000 and sponsors implement at least one substantial global amendment on nearly 60 percent of all clinical trial protocols. Not only does this add costs to the trial, these amendments lead to longer clinical trial durations, which can impact its exclusive time on the market and may affect profits.

Clinical studies also face rates of failure, which could cost sponsors an estimated $800 million to $1.4 billion per failed trial. Some of these failures may have been avoided – or shut down sooner, if sponsors had robust analytical tools and global data from which to base their decisions.

Fortunately, these tools are increasingly becoming available. While we can’t solve every problem with analytics, when sponsors work with experts who can apply machine learning and data analysis to their clinical research decision-making, they could be able to eliminate a lot of the mistakes that add unnecessary time and cost to their projects.

Bigger than big data

The use of data to support decision-making in clinical trials has been discussed for a long-time. For years we have talked about big data and how all of this information may be the answer to our problems. But as we have learned, it’s not enough to just have information. We need tools that allow us to mine those data sets, learn from what we find and translate those results into meaningful actionable insights.

This new generation of faster, more efficient trials is starting to happen today. We are using the industry’s richest data to increase our precision in trial feasibility and site selection, and we are starting to help sponsors design better trials based on insights rather than assumptions to optimize enrollment and speed trial delivery. We believe this may lead to game-changing results.

In a recent study in France, we used advance analytics and patient-level data from multiple sources to validate the protocol and identify treating physicians so that we could more accurately pinpoint the targeted patient population. Through these insights, we were able to predict the top 30 percent of investigators, who typically enroll about 80 percent of patients in phase II and III clinical trials, which is significantly higher than trials that rely on investigator data alone. 

In another example, while working on a study we found two potential sites that both had access to more than 350 patients diagnosed with Crohn’s Disease – a key criterion for the study. However, when we conducted a more detailed analysis of the data, focusing on which sites had patients who had been exposed to biologic therapy – another key inclusion criteria -- the number of potential patients dropped eight-fold, leaving one site with more than 100 patients, and the other with just 16. Being able to conduct that more detailed analysis meant we could avoid choosing a site that had little chance of meeting its enrollment goals.

These are just a few examples of how the right data and advanced analytics can define more precise enrollment plans, predict the right sites and study timelines, and speed recruitment.

We are also using machine learning tools to overcome the challenge of customizing clinical decision making for global studies. Anyone who has ever conducted a trial in multiple countries has experienced the tremendous complexity of addressing the unique needs of different regulatory bodies, and mining data that follow unique guidelines, structures and rules. By centralizing our analytics capabilities, we are leveraging global insights on country-specific challenges and tapping a library of analytic approaches to accelerate insight development. We are deploying technological advancements and data aggregation capabilities that allow us to create visualizations in key countries in a fraction of the time before, enabling much faster decision-making on country, site selection and study start-up.

We believe over time our ability to assess this data will improve, as machine learning algorithms get smarter with every search. That means each project will benefit from previous efforts, building efficiencies over time.

Why you should care

As we hone these tools and skills, QuintilesIMS envisions a future where enrollment rates and budgets become more predictable, where there are less protocol amendments, and where trials secure patients faster and at a lower cost.

These benefits won’t happen overnight but they are starting to happen, and I believe we are the turning point. The application of advanced analytics and machine learning has put these solutions within reach, and I, for one, am very excited to be a part of this transformation and what it means for patients.

Topics in this blog post: Real-world data, Biopharma, R&D, Strategic Partnerships