Real-World Evidence

What Are Real-World Evidence and Real-World Data?

The terms are often used synonymously. However, they are different; every source of RWD can be used as RWE—a combination of real-world data and analytics.

The FDA defines RWD as: “data relating to patient health status and/or the delivery of health care routinely collected from a variety of sources. Examples of RWD include data derived from electronic health records, medical claims data, data from product or disease registries, and data gathered from other sources (such as digital health technologies) that can inform on health status.”²

The FDA describes RWE as the “clinical evidence about the usage and potential benefits, or risks of a medical product derived from analysis of RWD.”²

Sources of RWD

RWD is often heterogeneous since it is collected from various sources. Figure 1 depicts some of these potential sources.

Figure 1: Sources of RWD 

Electronic health records (EHRs) are digital versions of patients' medical records stored in a central database. They contain information on patients’ diagnoses, treatments, medications, and lab results. Health records in this format are a valuable source of real-world data as they provide a comprehensive view of patients’ health over time. EHR data can be used for a wide range of research purposes, such as identifying trends in disease incidence and prevalence, evaluating the effectiveness of treatments, and assessing the safety of medications.

Claims data are records of healthcare services provided to patients and submitted to insurance companies for reimbursement. Besides the services offered to patients, they also contain the diagnoses and treatments received and the costs of care. Claims are often used to study healthcare utilization, charges, and outcomes. They are also helpful in evaluating policy impacts, such as changes to reimbursement rates or coverage.

Administrative data are collected by organizations or agencies for non-research purposes, such as billing or regulatory compliance. Such data can provide information on healthcare utilization, costs, and outcomes, as well as on patient characteristics and demographics. Administrative data can be used to monitor population health, evaluate the impact of policies, and identify areas for improvement in healthcare delivery.

Disease Registries are databases that collect and store information on patients with specific conditions or who have received treatments. They can help monitor the safety and effectiveness of medical products and interventions and identify patients eligible for clinical trials. Registries can also be used to track outcomes over time and evaluate the impact of interventions on patient health.

Social media data are generated by the users of platforms such as X (formerly Twitter), Facebook, and Instagram. They provide insights into patients' healthcare experiences and attitudes toward healthcare products and services. Social media posts can reveal trends in public opinion, sentiment, and behavior and monitor adverse events related to medications and medical devices.

Survey data are collected through surveys of patients, healthcare providers, or other stakeholders and provide insights into patients' experiences with healthcare and healthcare providers' attitudes and practices. Survey data is also valuable for assessing patient satisfaction, identifying barriers to healthcare access, and evaluating the impact of interventions.

Consumer data are generated by consumers' interactions with healthcare products and services. This data can include information on purchases, preferences, and behaviors and is used to inform healthcare companies' marketing and product development strategies.

Wearables include electronic devices worn on the body, such as fitness trackers, smartwatches, and medical devices. These devices collect an abundance of data, including physical activity, sleep patterns, heart rate, blood pressure, and blood glucose levels. Data from wearables provides insights into patients' health and wellness outside of a clinical setting and information on the effectiveness of medical interventions. Wearable data can help personalize treatments, monitor chronic conditions, and improve patient outcomes. However, wearables may not always provide accurate or reliable data, and data privacy and security concerns must be addressed when using wearable data in healthcare research

Differences Between RCTs and RWE Studies

While randomized controlled trials are the “gold standard” for evidence and provide robust product support, RWE helps fill the gaps.

Real-world Evidence, Clinical Trials, and Observational Studies

Real-world evidence is often compared to clinical trials and observational studies. However, governmental bodies and peer-reviewed papers agree that RCTs and RWE are critical components of an integrated evidence approach, where data from each study type can complement the other.

The FDA proposed a framework for RWE in December 2018 as part of the 2016 Cures Act. This framework evaluates RWE that helps support the approval of a new drug indication or satisfy post-approval study requirements.³ In its "EMA Regulatory Science to 2025 Reflection” paper, the European Medicines Agency said one of its strategic targets was to encourage high-grade RWD to make informed decisions.⁴ A recent study published in JAMA concluded that RWE studies can reach conclusions similar to those of RCTs when the RCTs’ designs and measurements can be closely emulated. The study also demonstrated how RWE can be leveraged to broaden understanding of the effectiveness, safety, and value of medications, complementing, or extending what can be demonstrated by RCTs.⁵

RWE in Pharma and Healthcare

In the past, RWE was used most frequently around the time of a drug's launch and less frequently as the drug matured. Today, many more opportunities exist to utilize RWE studies across the product lifecycle, ranging from understanding diseases and unmet needs to reducing the time and money spent in clinical development. RWE studies also support pricing, commercial effectiveness, pharmacovigilance, and label extension.⁶

Figure 2 below depicts the potential usage of RWE across a product’s lifecycle.

Figure 2: RWE Usage Examples Across Product Lifecycles

Two main factors drive the recent surge in the use of RWE studies: 1) the ability to collect large amounts of data from patients who use wearable tech, digital platforms, and other devices, and 2) advances in analytics that use ML, AI, and data analytics to process vast amounts of data.

Various stakeholders in the healthcare ecosystem can potentially use RWE to address several questions:

• Patients and physicians: Why is a new treatment preferred over existing treatments?
• Payers: What is the value for money offered by the new treatment?
• Industry: What situations in the industry are likely to result in the most benefit for clients when they use the new treatment?
• Regulators: Can RWE inform product labeling aligned to patient and physician needs?⁷

Figure 3 below depicts how RWE can help answer these questions.

 Figure 3: RWE Usage by Different Stakeholders

RWE Use Cases in Regulatory Decision-making

RWE is being increasingly used to expand product labels to include new indications. The tables below provide selected examples of these use cases.

Tokenization in RWE

Tokenization in healthcare protects patient privacy by replacing sensitive information with unique symbols called “tokens.” These tokens retain all the necessary data without revealing the patient’s identity. This technique helps researchers connect patient data from different sources without breaking the patient's confidentiality. Tokens also make it easier to recognize the same patient across multiple data sources. Connecting diverse patient data across multiple datasets provides opportunities, such as analyzing long-term healthcare trends for individuals or using advanced AI/ML analytics to find potential biomarkers for diagnosis or outcome prediction.

RWE Analytics

RWE analytics leverages advanced technology and tools to uncover patterns and insights not easily discernible through traditional data analysis methods. Analytics can automate data ingestion and cleaning to reduce errors. It can ensure consistent and standardized data analysis and reporting approaches, improving the efficiency and cost-effectiveness of evidence generation throughout drug development. Advanced analytics generally uses more complex analysis techniques to go beyond “what” happened and “why” it happened to answer “what will” happen and provide solid recommendations for the future. Some examples of advanced RWE analytics and the questions they answer include:

• Descriptive – What happened?
• Diagnostic – Why did it happen?
• Predictive – What is likely to happen next?
• Counterfactual – What might happen if we change something?

How RWE Can Improve Health Inequity

The growing consensus among healthcare stakeholders is that social determinants of health (SDOH) strongly influence health outcomes, healthcare utilization, and cost. SDOH are non-medical factors that the World Health Organization describes as

“…the conditions in which people are born, grow, work, live, and age, and the wider set of forces and systems shaping the conditions of daily life. These forces and systems include economic policies and systems, development agendas, social norms, social policies, and political systems.”¹⁶

Linking RWD with SDOH can provide valuable insights into how social factors impact healthcare outcomes and inform targeted interventions to improve health equity.

Limitations of RWD

Real-world data and RWE have limitations that must be carefully considered when interpreting and using the data to inform clinical decision-making. Some examples include:

• Lack of standardization: RWD is often collected from various sources using multiple methods, making it challenging to standardize and integrate into analyses.
• Selection bias: RWD is often collected from specific populations or healthcare systems, which can introduce selection bias and limit the general usefulness of findings.
• Data quality: RWD is often collected for clinical or administrative purposes rather than research, which can result in incomplete or inaccurate data.
• Incomplete data: RWD may not capture all relevant data points or may have gaps that limit the ability to analyze the data or draw conclusions.
• Data privacy and security concerns: RWD may contain sensitive information, such as patient health information, which raises concerns about privacy and security.
• Confounding factors: RWE may be influenced by factors other than the intervention being studied, such as patient characteristics or co-morbidities, which can make it challenging to isolate the effects of the intervention.
• Lack of causal inference: RWE cannot establish causality between interventions and outcomes in the same way that randomized controlled trials can.
• Limited ability to study rare outcomes: RWE may not provide enough data to study rare outcomes, hindering conclusions about the safety or efficacy of interventions for those outcomes.

Conclusion

Real-world evidence stems from RWD collected outside of controlled clinical trials from myriad sources. This trove of empirical data, alongside RWE-influenced studies, plays a crucial role throughout the product lifecycle. It also offers valuable insights that complement clinical trials and support regulatory and reimbursement decisions. Stakeholders, including physicians, pharmaceutical companies, payers, regulators, and patients, increasingly recognize the value of RWE in improving health outcomes.

This article is contributed by Javed Shaikh, Director at Axtria.