How AI is Helping to Handle the Onslaught of Clinical Trials Data

Over the past decade, data volume has increased threefold, according to the Tufts Center for the Study of Drug Discovery (CSDD). All of this data needs to be collected, secured and analyzed — so newer, better medications can be developed to help patients around the world.

Traditional (read: manual) approaches to cleaning and monitoring data simply can’t scale — and the status quo of those processes and systems are coming under greater scrutiny. This has created a bottleneck that hinders the advancement of medicine. Newer methodologies that could improve quality and move drugs to market faster — with fewer resources — are becoming more appealing.

Pharmaceutical companies have been slow to adapt because new solutions must:

• offer quantifiable improvements over existing solutions;
• accelerate or provide efficiencies in trial execution; and
• support a highly regulated environment with little tolerance for error.

Many challenges, revolutionary solutions

One of the foremost challenges facing companies that conduct clinical trials is the industry-wide shortage of skilled Data Managers. Any solution to this issue must improve efficiency and make the best use of the people we do have.

Beyond staffing issues, clinical trials also face slowdowns in particular milestone activities, including study design and review, study startup, site feasibility and initiation, operational oversight, database lock, and more.

However, there are major opportunities to improve — if we’re willing to fundamentally change business processes and marry those changes with new technologies like Artificial Intelligence (AI).

For example, addressing data quality throughout trials can be done in near real-time by augmenting workers — not replacing them — with AI technology. This added tool enables faster, more effective investigations and speeds time to query issuing, leading to resolution.

Ongoing, proactive data reconciliation results in a substantial reduction of activities to be done before the database can be locked, shrinking database lock time and eliminating whitespace that can delay submissions.

What to do with all this data

Forward-thinking companies are embracing AI technologies in handling this large amount of data, getting drugs to market, and then to patients faster.

Using AI as the basis of any data aggregation solution is a next-level approach to metadata-driven data collection. Having access to the numerous data sources across a clinical trial and then aggregating that data in an application-accessible data fabric is a game changer.

With cross-domain data accessible, we can harness the power of analytics and insights into real-time data quality monitoring, powerful operational metrics, and customizable patient data interrogation. In simplest terms, we can run trials more efficiently, improve ongoing data quality, and enable companies to make more informed decisions in near real-time — ultimately accelerating treatments to patients.

Data + AI  = The future

We’re only scratching the surface of AI’s potential in the life sciences industry. A particularly unique application is exploring algorithms to make better use of knowledge from both completed and ongoing trials. There’s much to learn from the data we already have — on top of the data we continue to gather every day.

On a day-to-day basis, if we allow AI to perform routine data investigation tasks and guide or suggest areas for further analysis, we empower Data Managers and other staff to spend their time on more complex investigations and interrogations.

By using AI analytic tools, companies can make sense of the data as it’s collected — from all sources — and in context with the rest of the trial data. They can have cross-domain analysis at speed and scale, so they’re able to monitor data quality in near real-time, make better study operational decisions, and deploy resources as efficiently as possible.

The workflow automation capabilities provided by AI are just as exciting — providing the ability to minimize data management hours and accelerate clinical trials in new and innovative ways. Solutions have already been deployed to greatly reduce the time spent reconciling discrepancies in an EDC system — automatically detecting and identifying reasons for those discrepancies, and pre-generating query text. And throughout the process, these solutions keep the “human in the loop” to manage AI-driven predictions.

In the end, it comes down to gathering these massive amounts of data in a single and accessible place, processing it faster and more accurately, and making smarter decisions — all of which leads to getting new treatments through the process quicker and into the hands of the patients that need them most.