In the landscape of cancer research, clinical trials remain the gold standard for evaluating new therapies. Historically, cancer drug development has followed a structured path: phase 1 studies determine safety and dosing; phase 2 studies assess preliminary efficacy, typically in small single-arm cohorts; and phase 3 trials compare the new therapy to the standard of care in randomized controlled settings. Although this model has driven countless advances, the highly controlled nature of these explanatory trials results in selected populations that may not fully generalize to the breadth and complexity of patients with cancer who are treated in routine practice.
The U.S. Food and Drug Administration (FDA) Oncology Center of Excellence and the broader oncology community have long encouraged expansion of cancer clinical trial eligibility, where appropriate.Eligibility criteria are likely to be more stringent in phase 3 trials to reduce confounding and to maximize the ability to isolate a clear and reliable treatment effect. However, registrational trials frequently exclude older adults, patients with significant comorbidities, and other important subgroups that make up a representative real-world U.S. population. This creates a significant challenge in generalizing trial results from selected trial populations to the patients who will ultimately be treated in routine clinical care
To address these gaps, RWD has emerged as a potentially useful tool. RWD are defined by the FDA as “data relating to patient health status and/or the delivery of health care routinely collected from a variety of sources.”In other words, RWD refers to health information collected outside the controlled environment of clinical trials. This data source can include EHRs, insurance claims, and registries. RWD may have the potential to complement data from traditional randomized controlled clinical trials to enhance generalizability of the safety and efficacy results to broader populations in routine care settings.
Unfortunately, a fragmented U.S. health care system and heterogeneous EHR systems lead to challenges in the use of RWD. Retrospective uses of RWD often face challenges, including incomplete or missing data, that are exacerbated by a lack of monitoring for data quality. These limitations significantly affect the reliability of insights derived from RWD. A promising opportunity is designing prospective standardized RWD data collection through observational studies or registries. Prospective approaches can collect data deliberately and systematically, creating more standardized datasets with less missing data while retaining representation of populations treated in routine practice settings
While RCTs are designed to answer the question “Can the drug work?” (i.e. efficacy), observational studies and RWD are more adept at answering the question “Does the drug work?” (i.e. effectiveness). For most tumor types and treatments, surrogacy be- tween efficacy and effectiveness is not established—in other words, even if a drug works in a clinical trial, it may not improve outcomes when delivered in the real world. This efficacy-effectiveness gap is widely recog- nised, and informs daily discussions with patients. If RWD outcomes are not a valid surrogate for clinical trial outcomes, can RWD play a separate role, using distinct endpoints to evaluate important questions and improve the time to drug access for patients? Can efficacy and effectiveness questions have independent value for cli- nicians, regulators, and patients, regardless of surro- gacy? Finally, if a drug has demonstrated efficacy in clinical trials but cannot demonstrate effectiveness when adopted in a real-world population, does it still have value and should it still be prescribed? These questions require careful evaluation as the cancer community develops processes and pathways to inte- grate the growing amount of real-world data into regu- latory, funding, and clinical decision making.
Positive efficacy, negative effectiveness– what to do next?
As RWD continues to emerge as a common source of data, stakeholders (including patients, clinicians, regulators and funding agencies) will face a growing number of scenarios where trials demonstrate positive results, but these do not translate into routine practice. This may prompt important questions from patients such as “If the drug hasn’t been shown to benefit patients in the real world should I still take it?” Clinicians will be faced with increasingly complex discussions as we try and explain the differences between trial and real-world populations, and as we try and analyse reasons for these discordant results. Should such discordant results force clinicians to adhere more rigorously to prescribing the drug only to patients who meet the clinical trial criteria? And what answers can we offer to patients who fall outside of these eligibility criteria? Regulatory bodies may be faced with similarly challenging questions, such as “should drugs that fail to demonstrate population- level benefit in the real-world still be approved?”, and “Should a demonstration of real-world benefit be required as part of drug’s regulatory lifecycle?” Funding agencies will also face questions, such as “If a drug is cost-effective based on clinical trial analyses, but is not cost-effective using real-world data, should it still be funded?”, and “Should drug pricing be informed by real-world effectiveness?” These questions require careful consideration as we increasingly rely on real- world data for treatment, regulatory and funding decisions.
Novel study designs that may bridge RWD and clinical trials
Given the limitations of both RCTs and observational RWD, hybrid and alternative study designs capable of generating RWE may help bridge the gap between clinical trials and RWD. As an example, pragmatic trials offer unique opportunities to adopt the methodological rigour of the RCT with the potential cost savings and practical advantages of real-world studies. Pragmatic clinical trials must have 3 key attributes: (1) an intent to inform decision makers (patients, clinicians, adminis- trators and policy makers); (2) an intent to enroll a population relevant to the decision in practice or representative of the patients or populations and clinical settings for whom the treatment is relevant; (3) and an intent to streamline procedures and data collection, so that sufficient power can be allocated towards informing clinical and policy decisions. Depending on the design, pragmatic trials may generate randomized evidence within the context of a real-world population. Hybrid trials, which use the traditional clinical trial design but incorporate pragmatic trial elements, also have the capability of using RWD to generate RWE.
The FDA has launched Project Pragmatica to promote pragmatic trials capable of generating RWD and promoting trials designed with functional efficiencies, such as fewer eligibility criteria, increased trial flexi- bility, and enhanced patient centricity. The European Organisation for Research and Treatment of Cancer (EORTC) has also voiced support for select RWD studies, prioritizing the execution of clinical trials that produce randomized real-world evidence,of which pragmatic clinical trials are an example. Other examples
of collaborative efforts to increase pragmatic trials include the REACTs collaborativedesigned to compare standard approved treatments in a real-world setting across a broad range of patients, and NIH Pragmatic trials Collaboratory designed to strengthen the national capacity to implement large-scale research studies that engage health care delivery organizations as research partners.
Conclusion
Interest in using RWD to support regulatory sub- missions is growing. However, we believe that the pre- requisite for most approvals should still remain robust evidence regarding efficacy based on a well conducted RCT designed to meet a clinically meaningful endpoint for patients. Observational and RWD cannot answer the question “Can a drug work?” and treating patients before efficacy has been clearly demonstrate places pa- tients at risk of harm and toxicity without proven benefit. However, there may still be value in prospective observational studies evaluating “Does the drug work” in real world practice, though many questions remain regarding how patients, clinicians, regulators, and funding agencies should use the results when there is a lack of effectiveness despite proven efficacy.
Author: Eliza Akhveldiani, MD, Clinical Oncologist, Kristina Kili Oncology Center
