Why EHR Data is Essential for Real-World Evidence Generation

During a recent Endpoints-hosted webinar: “Lessons from the Real World: Journey from Claims to EHR Data in Real-World Evidence Generation,” we outlined why real-world data (RWD), specifically electronic health record (EHR) data, is becoming a more valuable option for generating real-world evidence (RWE), which is crucial for informed decision-making in therapy development. 

In the webinar, with special guest Dawn Sim, Medical Director of Ophthalmology at Genentech, we:

• Analyzed specific elements that can be utilized in claims vs. EHR data as RWD sources, both individually and when linked
• Reviewed how advanced AI is applied to curate EHR clinical notes and extract key information at scale
• Offered examples and case studies of how life sciences companies select and then leverage these data sources for RWE studies

The Power of Real-World Data: Claims vs. EHR

Real-world data is a broad term encompassing data collected outside of the controlled environments of clinical trials, often through routine clinical practices. The data comes from sources such as EHRs, claims, and images.

Both EHR and claims data bring unique strengths to the table, but claims data has limitations that life sciences companies must navigate, depending on the use case. For example, claims data is highly structured and allows for quick and seamless analysis. However, it lacks clinical nuance, such as the results of patient symptoms, diagnostic tests, or insights into the reasons for particular diagnoses.

This is where EHRs can fill in the gaps. Within the unstructured clinical notes of EHRs lies a rich source of context that can reveal insights into the patient journey.

Frameworks for Data Quality and Selection

One critical challenge in RWD generation is selecting high-quality data that can answer the specific questions at hand. RWD should be assessed using three major principles: depth and validity, source of data, and recency.

1. Depth and validity: Does the data provide enough detail to answer the research question? Are the clinical outcomes captured in sufficient granularity?
2. Source of data: Where is the data coming from, and how reliable is it? One example of quality data is that it’s captured directly from clinical encounters through EHRs, collected through qualified clinical data registries.
3. Recency: How recent is the data, and how important is data refresh timeliness for the research? For instance, newly approved therapies often require recent data collection to assess their impact in the real world.

These considerations should be at the forefront when selecting fit-for-purpose datasets. 

One of the leading medical societies that Verana Health partners with is the American Academy of Ophthalmology. Verana Health curates RWD from the American Academy of Ophthalmology IRIS^® Registry (Intelligent Research in Sight) – the world’s largest specialty society clinical data registry in the United States. The IRIS Registry includes data on 80 million de-identified patients from 15,000 contributing clinicians. This massive RWD network allows life sciences companies to conduct robust analyses that are representative of the U.S. population.

The Role of Artificial Intelligence in Real-World Data Collection from EHRs

Artificial intelligence (AI) has become indispensable in healthcare, particularly for processing large volumes of RWD. Verana Health utilizes AI techniques, such as machine learning and natural language processing, to extract valuable insights from both structured and unstructured data within EHRs.

However, applying AI to RWD comes with significant challenges, which is why it’s important to utilize rigorous quality assurance and quality control processes through the model development lifecycle to prevent any biases or inaccuracies in the data, which can compromise the results.

For example, the AI models developed by Verana Health undergo continuous monitoring for data drift and clinical relevance checks to ensure accuracy and robustness over time.

Case Study: Real-World Data in Ophthalmology

Recent Verana Health and Genentech research, shared by Dawn, proves how curated RWD, utilizing AI, can be analyzed to reveal valuable insights in ophthalmic research. The study examined the real-world use of faricimab to treat neovascular age-related macular degeneration (nAMD). Leveraging RWD from the IRIS Registry, we analyzed treatment outcomes of this anti-VEGF therapy in a large cohort of patients to better assess patient burden, confirm learning in clinical trials, and inform clinician communications.

Patients treated with faricimab in the real world showed improvements in vision, particularly those who were treatment-naive. One of the most striking findings was the reduced frequency of injections performed over time—an insight crucial not only for clinicians, but also for payers and policymakers planning future treatments. A reduction in injections can reduce burden on both patients and caregivers. 

The study highlighted the power of combining RWD with clinical expertise to deliver actionable insights, particularly with regards to real-world patient characteristics, treatment patterns and  outcomes.

The Future of Real-World Evidence Generation

As healthcare continues to embrace data-driven decision-making, the role of EHR data in real-world evidence generation will only grow. However, the key to generating reliable and meaningful insights is to collaborate with a trusted data partner who can combine structured and unstructured data from EHRs, leverage advanced AI tools, and ensure rigorous quality control processes. 

To learn how Verana Health is teaming up with life sciences companies to deliver high-quality insights that are making an impact in healthcare, click here.

If you’d like to learn about this topic in detail, view the webinar recording: Lessons from the Real World: Journey from Claims to EHR Data in Real-World Evidence Generation