What real-world data shows about current cancer care

Credit: Adobe Stock Clinical trials, defined by meticulous randomization, bias reduction, and causal proof, are often referred to as the gold standard of medical research. But the rigor and focus that make clinical trials so valuable simultaneously leave most patient experiences out of the picture.

“About 5% of adult patients with cancer are enrolled into a clinical trial,” said Jeff Allen, president and chief executive officer of Friends of Cancer Research. “So our vantage point in terms of the experience of treating cancer is relatively narrow, when you think of the information that we typically see from clinical studies.” The consequences of this narrow viewpoint are evident in day-to-day cancer treatment.

Oncologists are often challenged to provide answers to patients who don’t fit the patient profile of a trial or match the specific trial question; clinicians need to make decisions about sequencing or treatments that haven’t been tested head-to-head or in patients like their own.

In the variable world of patient treatment, it’s necessary to look beyond clinical trials for practical evidence to guide clinicians on the best course of action for unique individuals. In recent years, real-world data (RWD) has provided an important avenue for expanding the understanding of disease progression, treatment effectiveness, and real-world patient experiences across all areas of medicine.

“The opportunity to broaden the aperture of insights into experience with cancer treatment is really important,” said Allen. “Technology is moving so rapidly that hopefully we’ll be able to design data sources that are able to rapidly answer questions in situations where a clinical trial may not be necessary or entirely appropriate.” Examining the expansive use of RWD in a transparent manner, looking at the performance of different data sources, and understanding where opportunities and limitations exist, can inform a familiarity and degree of comfort with applying different sources in clinical practice.

A growing RWD movement Historically, RWD was defined as data generated during routine clinical care outside of clinical trials and consisted primarily of structured claims and pharmacy data. It has evolved to include sources such as electronic health records (EHRs), genomics, patient-reported outcomes (PROs), and data from wearable devices like smartwatches.

The use of AI and large language models has facilitated the curation of unstructured RWD, and by connecting data from different sources (such as EHRs and claims), researchers and clinicians have gained a much more detailed, longitudinal and representative understanding of the patient experience than was traditionally afforded by clinical trials alone.

The FDA has since built frameworks for using RWD and real-world evidence (RWE) to support drug approvals. The 21st Century Cures Act, signed in December 2016, directed the agency to evaluate the use of RWE in regulatory decisions, and the FDA released its RWE framework in December 2018.

“The 21st Century Cures Act really acted as a catalyst in the drug and biologic space,” said Donna Rivera, executive vice president of clinical evidence modernization for Canal Row Advisors, who previously led the Oncology Real World Evidence Program at the FDA’s Oncology Center of Excellence.

“After 21st Century Cures, there was the release of the RWE framework, a suite of guidances from the FDA. This evolution towards evidence for effectiveness has shown in many ways, especially in oncology and rare diseases, that real-world data can serve and has served as pivotal evidence.” Today, clinicians and researchers are using RWD throughout the development lifecycle, from optimizing clinical trials and setting up external control arms, to pharmacovigilance and commercialization.

High-quality RWD and AI accelerate research by enabling faster data and insight generation, allowing researchers to test drug targets and identify biomarkers before taking a drug to trial. The question of representativeness in studies and RWD One common example of how RWD can support informed decision making, and affect the representativeness of a study population, is with setting the eligibility criteria during trial design.

For instance, determining the appropriate hemoglobin level for including patients in the clinical trial. “That’s a very common question that comes up in the majority of the hematological or oncological clinical trials and studies,” said Ankit Kansagra, MD, a hematologist-oncologist and associate vice president of oncology integration and community alignment at the Tampa General Hospital (TGH) Cancer Institute.

“If you lower the hemoglobin from eight to seven-and-a-half [grams per deciliter], what does it really do? Does it really lead to poor survival in the real world? Or, does it lead to enrolling a subset of patients which might have otherwise been excluded?” said Kansagra.

“Real-world data has the power to help make those decisions in day-to-day life for clinicians as they design clinical trials.” The issue of representation is important not only in designing clinical trials, but in evaluating real-world data sources. With RWD, it’s important to know where it’s coming from and what patient population it reflects.

For example, is the data exclusively from large cancer centers or does the dataset cover a range of treatment facilities and patient types? To support viable insights and evidence, the RWD should reflect the broader population. As researchers increasingly turn to RWD to help improve the design of clinical trials, evaluate eligibility criteria, and address questions such as what is the optimal sequencing of new therapies, ensuring the representativeness of the underlying RWD is one of several critical elements to keep in mind.

The importance of context RWD isn’t an automatic panacea for the limitations of clinical trials. In addition to its representativeness, RWD must possess several additional qualities. “From a regulatory perspective, fit-for-use data relevance and reliability are fundamental,” said Rivera.

But even before you evaluate RWD for fitness-for-use, Rivera pointed out that the important contextual question should be evaluated. “And that’s really whether real-world data or real-world data enabled designs are appropriate for the study objective. Before we think about whether the data are fit-for-use, what is the use?

Is the clinical context appropriate? Does it align with real world evidence generation, or is a clinical trial perhaps more appropriate if feasible and ethical?” If RWD is appropriate, then the fit-for-use evaluation centers on the components of relevance and reliability. The technical components in relevance include availability and feasibility, Rivera said, as well as key study variables like exposures, outcomes, and covariates.

“But also, is it reflective of the epidemiology of the disease and the intended use population? This is where you see a lot of the particular uses in oncology, rare diseases, pediatrics, and specific subpopulations where it is particularly helpful.” Reliability: ensuring the data supports decision-making The standard for data reliability in terms of using RWD to support regulatory decision-making centers around the accuracy, completeness, and trustworthiness of the data.

Ensuring the completeness of RWD is important. Capturing the entire patient journey in an RWD source can be difficult since clinical documentation is typically designed for care delivery rather than research purposes. One area where missingness is typical is in the biomarker space, as biomarker testing can vary by practice in the real world.

However, the increasing ability to link different data sources together, from EHR to genomics to claims, offers a means to achieve a more complete picture of the patient journey. One ongoing challenge with RWD relates to variability: the lack of uniform approach across different medical practices, data systems, and types of data, as well as the lack of standardized definitions (for example, “line of therapy”) across different datasets.

Variability in real-world practice isn’t inherently bad. The variability in clinical practice often reflects efforts to tailor the treatment regimen and the approach for the best scenario for the individual patient, Allen said. “But being able to identify those sources of variability and how to manage them, and in the form of performing an analysis, is really important, and that comes back to data transparency,” said Allen.

For organizations, clinicians, and researchers working with RWD, establishing consistent protocols and transparent methodologies is necessary to build confidence in observations. Future outlook Real-world data is helping to overcome the natural limitations of clinical trials, filling important gaps and broadening our understanding of diseases and the patient experience.

The application of RWD to oncology continues to deliver significant benefits to cancer research and care. What does the future hold for RWD in oncology? Challenges and opportunities remain.

RWD highlights significant gaps in standard biomarker testing, with uptake ranging between 70% and 80%. And the lack of standardization across the industry restricts efficient data curation and limits applications. But improved technology and increased collaborative efforts, including an international push to harmonize guidelines and improve the interoperability of datasets, hold promise for continued improvement in RWD.

“I think you see this collaborative community coming together and trying to figure out how we increase standardization and interoperability,” said Rivera. “There are initiatives and ways in which technology and AI can help in this more seamless data collection in clinical settings, bridging clinical care and clinical research.” As technology continues to improve, techniques such as ambient scribing hold potential to streamline documentation and improve data completeness, provided these tools are carefully implemented to deliver structured, usable data for research.

With advances in RWD techniques and application, we’re progressing toward a goal of precision care: real-time, point-of-care recommendations for an individual’s unique condition. And that’s a gold standard of medical practice worth pursuing. Ming He, MD , is senior medical director at Verana Health, where she works with research, analytics and product teams on oncology real-world evidence, particularly in hematologic malignancies.

A board-certified hematologist and medical oncologist, she trained at UT Southwestern Medical School, completed her residency at Presbyterian Hospital Dallas and her fellowship at Baylor College of Medicine. Before joining Verana, which merged with COTA in January 2026, she practiced clinical oncology and served as a clinical lead for IBM Watson for Oncology.

She is adjunct faculty at UT Southwestern Medical School. Filed Under: clinical trials , Data science , Oncology , Regulatory affairs Tagged With: 21st Century Cures Act , biomarker testing , clinical trials , COTA , data standardization , electronic health records , external control arms , FDA , Friends of Cancer Research , oncology , real-world data , real-world evidence , Tampa General Hospital , Verana Health