Smarter Ways to Evaluate and Apply Real-World Data

Real-world data (RWD) is no longer scarce. In fact, it’s everywhere. But as experts from Snowflake, Medtronic, and Inovalon emphasized during a recent live session, access alone does not guarantee value. Without a clear objective, even the most robust datasets can fall short – wasting time, resources, and credibility.

During Smarter Strategies for Working With Real-World Data, industry experts explored how real-world data sources are evolving, what “regulatory grade” really means in practice, how artificial intelligence (AI) is reshaping discovery and stewardship, and the most common pitfalls life sciences organizations should avoid. The recurring theme was clear: success with real-world data depends on being intentional, not reactive.

Real-world data needs a strategy to succeed

The session opened with a simple but critical reminder: Data is only powerful when it’s applied with purpose. Many organizations start with the data they can access, rather than the evidence they actually need. That mismatch often leads to gaps later, especially when regulatory, clinical, and commercial requirements begin to deviate.

Experts stressed the importance of clearly defining the question first. Whether the goal is a regulatory submission, market access support, or internal decision-making, teams need alignment on what they’re trying to prove before selecting a dataset. No single source can do everything, and assuming otherwise can be a costly misstep.

The definition of real-world data keeps expanding

The panel noted how dramatically RWD has changed in just the past five years. Traditional sources like claims and EHR data are now commonly integrated to provide both longitudinal views and clinical depth. Beyond that, newer inputs like registries, lab data, biomarkers, wearables, and unstructured physician notes are becoming part of everyday evidence strategies.

This expansion brings opportunity, but also complexity. As real-world data becomes more abundant and varied, teams must think carefully about how different sources complement one another and where their limitations lie. The boundaries of what qualifies as real-world data will continue to stretch, making thoughtful source selection even more important over time.

What “regulatory grade” really means

“Regulatory grade” is a term that is used frequently, but the panel cautioned against treating it as a vague label. In practice, regulatory-grade data must meet very specific expectations around accuracy, completeness, and transparency. It’s not enough for RWD to be directionally correct; it needs to be defensible.

Speakers emphasized that fit-for-purpose evaluation should be the first step when assessing any dataset. That means understanding what’s missing, how gaps can be filled, and whether the data can support the level of confidence regulators expect. In many cases, reaching that bar requires combining multiple sources, such as linking claims with EHR data or leveraging registries to strengthen rare disease evidence.

AI’s growing role in discovery and stewardship

AI was a major focus of the discussion, particularly its role in accelerating discovery across complex, multimodal datasets. Panelists highlighted how AI can make it easier to surface life sciences data from sources like free text, audio, and wearable data, which were previously difficult to analyze at scale.

At the same time, they cautioned against overreliance. Faster discovery doesn’t eliminate the need for precision. Human oversight remains essential to ensure insights are meaningful, appropriate, and responsibly applied. As AI adoption grows, leaders will increasingly focus on governance – managing how tools are used, not just what they can do.

Market pressures, data loss, and responsible access

Recent market disruptions, including data breaches and shifts in data availability, have made stewardship a top concern. The panel addressed the reality that having many vendors does not necessarily mean having many unique data sources. Repackaging is common, and organizations must understand what they are truly buying.

To address shrinking access in some areas, many teams are shifting their strategies from open claims to closed claims, EHRs, and lab data to maintain continuity. Modern stewardship approaches, such as clean rooms and cohort-level outputs, were highlighted as ways to balance access with privacy, particularly for sensitive or rare disease cases.

Common pitfalls and how to avoid them

When asked about the biggest mistakes organizations make with life sciences data, the panelists were aligned: lack of clarity is the root cause of most problems. Other red flags include assuming one dataset can answer every question, underestimating data gaps, and prioritizing speed over suitability.

Their advice was straightforward: Don’t lose sight of the bigger picture. Start with a well-defined plan, be honest about limitations, and engage data and technology partners early. These experts can help teams avoid missteps before they happen.

Develop your real-world data strategy

Real-world data will continue to grow in volume, variety, and importance. But as this session made clear, outcomes depend far more on strategy than scale. For deeper insights, practical examples, and expert guidance, watch the on-demand recording of Smarter Strategies for Working With Real-World Data and hear from the experts themselves.

Inovalon and design® and Inovalon® are trademarks of Inovalon, Inc.