What Does the Future of Clinical Research Look Like in a Data-Driven World?

With trial timelines under pressure, patient recruitment falling short and regulatory expectations climbing, pharmaceutical companies are being pushed to rethink how evidence is generated and applied across the clinical development lifecycle.

Traditional trial models are struggling to keep pace with scientific complexity, cost constraints and calls for broader patient representation.

In this Xtalks Spotlight interview, Alison Hollands, Senior Vice President of Epidemiology and Clinical Research Solutions (ECRS) at Optum Life Sciences, shares how real-world data (RWD), unstructured datasets, external control arms and adaptive trial designs are helping sponsors address these challenges head-on.

From the rapid increase in clinical trial endpoints to the ongoing underrepresentation of minority populations, Hollands offers practical insights into how data-driven strategies are reshaping study design. These approaches are also improving safety surveillance and accelerating time to insight, all without compromising scientific rigor.

The Increasing Data Demands Behind Today’s Drug Approvals

With the FDA approving around 50 new drugs annually, including a growing share in rare diseases, oncology and gene or RNA-based therapies, the need for smarter, more scalable evidence-generation strategies has never been greater.

Each approval not only represents a scientific milestone but often brings new post-marketing requirements that extend a product’s clinical obligations well beyond launch.

“Actually, 30% to 50% of new drug approvals need at least one post-marketing safety study. That means the FDA has asked pharmaceutical companies to follow up and conduct more studies,” noted Hollands. “And around 1,000 active studies are currently being tracked in the FDA’s database.”

These studies are designed to detect long-term safety signals, rare adverse events and pregnancy-related risks, all of which may be difficult to capture through traditional randomized trials.

Understanding how treatments perform in real-world settings, particularly across diverse patient populations, has become a critical component of post-approval evidence generation. Real-world data (RWD) helps bridge the gap between controlled trial conditions and everyday clinical use, offering a more complete picture of a therapy’s value and safety.

But regulatory pressure is just one part of the equation. Trial sponsors are also navigating rising protocol complexity, increasing data volume and growing expectations around inclusion and transparency. The result is a research environment where traditional models are falling short, and where enrollment failure remains a persistent bottleneck.

In fact, Hollands mentioned how 80% of trials are still failing to meet enrollment targets.

“What we are really trying to do is increase that patient pool… so that life sciences is still able to study the right patient population, but they’re not eliminating groups of patients unintentionally,” explained Hollands.

This has brought greater focus to a set of emerging priorities in evidence generation:

• Post-marketing and Phase IV evidence requirements
• Adaptive designs that allow mid-trial adjustments based on emerging data
• Decentralized and hybrid trials that reduce the need for site visits
• Stronger mandates for inclusive recruitment and real-world relevance

These trends signal a move toward more flexible, data-integrated research models that meet regulatory demands while making trials more accessible and resilient.

Simplifying Complexity in Modern Study Design

Trial complexity has increased over the last decade, often unintentionally. Hollands points to a 33% increase in clinical trial endpoints, with the average trial now measuring around 26 endpoints, placing strain on both patients and sites.

This expansion of endpoints not only prolongs study timelines but also creates downstream issues in data management, monitoring and regulatory review. Sites must dedicate more staff hours to data capture and verification, while patients are often asked to complete lengthy questionnaires or undergo repeated testing that may not align with their day-to-day lives.

This complexity contributes to patient dropout and makes it harder to translate Phase III findings into real-world care settings.

“We are really trying to use data and insights to inform protocol optimization,” explained Hollands. “It’s often difficult to translate what you were finding in a Phase III setting to somebody’s ordinary practice.”

To address these pain points, Hollands points to several strategies, including:

• Structured and unstructured data for protocol feasibility
• eCOA tools and wearables for remote patient engagement
• Hybrid trial models including remote components to reduce geographic and logistical barriers
• Targeted strategies to broaden patient pools, ensuring underrepresented groups are not unintentionally excluded

The goal is to build studies that are scientifically rigorous while remaining practical, inclusive and reflective of real-world care.

Epidemiology at the Center of Post-Marketing Safety

Post-marketing studies are often tasked with detecting rare adverse events, which requires large, longitudinal datasets and methodological precision. Epidemiology teams must tackle:

• Confounding factors
• Missing or inconsistently coded data
• Bias from non-randomized real-world settings

“They have to be a historian, a detective and a storyteller,” Hollands said. “They really have to design studies that reflect real-world complexity.”

“It could mean looking at long-term safety, pregnancy outcomes and rare adverse events.”

— Alison Hollands

One increasingly urgent area of study is maternal and infant health outcomes. Because pregnant women are often excluded from trials, data linkage between mother and infant records is critical for understanding exposure risk, a trend Hollands said is gaining momentum.

This type of analysis not only fills a long-standing evidence gap but also supports regulators and clinicians in making more confident decisions about treatment safety during and after pregnancy.

Rethinking Trial Design with Unstructured Data and More

While structured datasets remain the foundation, Hollands emphasized that 80% of healthcare data is unstructured, and unlocking it holds the key to identifying and enrolling more eligible patients. They can reveal early symptoms, family history or subtle clinical observations that structured fields may overlook.

According to Hollands, based on their experience in the Alzheimer’s disease space, they found unstructured data, such as provider notes, symptom descriptions or radiology findings, can expand the potential patient pool by up to 30% compared to structured diagnostic codes alone. She explained that this expansion not only broadens recruitment but also helps create trial populations that more accurately reflect the heterogeneity of real-world patients.

“If you’re not able to tap into unstructured data, you’re not able to access as many patients as you possibly could.”

— Alison Hollands

Another powerful innovation is the use of external control arms. Rather than relying on traditional placebo groups, observational data can now supplement or replace control cohorts in certain indications. Hollands pointed to oncology, pediatrics and life-threatening diseases as key areas where this approach could be applied.

Hollands also mentioned that regulators, including the FDA and EMA, are showing increasing openness to external control data when it is rigorously sourced and validated. By reducing reliance on placebos, trial sponsors can lower costs, shorten timelines and still meet the high evidentiary standards required for regulatory approval.

Looking Ahead: AI, Inclusion and Adaptive Trials

Looking to the future, Hollands is optimistic about how clinical research can evolve in alignment with Optum’s mission to “make the health system work better for everyone.”

She sees five major trends shaping the years ahead:

• Broader use of Phase IV trials to assess long-term impact and value
• Decentralized and hybrid models that bring research to where patients are
• AI and machine learning expanding from recruitment to outcome prediction
• Deeper commitment to inclusive trial design and community engagement
• Adaptive trial designs that respond to interim data in real time

“We need to truly embrace real-world evidence,” Hollands said. “That means trials should take place where patients are already being treated. If they can discuss participation with a trusted provider, they’re more likely to understand the study and feel comfortable joining and staying in the trial.”

As clinical trials become more data-rich and patient-driven, success will center on how effectively life science organizations bridge traditional research with real-world relevance.

For Hollands and the Optum Life Sciences team, the future lies in connecting the right data, tools and communities. The goal is to ensure that every study is not only more efficient but also more equitable and meaningful.

This article was created in collaboration with the sponsoring company and the Xtalks editorial team.