About the Duke-Margolis Center for Health Policy

‍
The Robert J. Margolis, MD, Center for Health Policy at Duke University is directed by Mark McClellan, MD, PhD, and brings together expertise from the Washington, DC, policy community, Duke University, and Duke Health to address the most pressing issues in health policy. The mission of Duke-Margolis is to improve health, health equity and the value of health care through practical, innovative, and evidence-based policy solutions. Duke-Margolis catalyzes Duke University’s leading capabilities, including interdisciplinary academic research and capacity for education and engagement, to inform policymaking and implementation for better health and health care.  For more information, visit healthpolicy.duke.edu.

‍

Acknowledgments

Duke-Margolis would like to thank several individuals for their contributions to this white paper. The paper would not have been possible without the months-long collaboration of the Shared Evidentiary Opportunities Working Group, which was comprised of representatives from both the Duke-Margolis Real-World Evidence Collaborative and its Value for Medical Products Consortium. Their expert perspectives, open discussion, and thoughtful feedback on working drafts were indispensable, and we are grateful for their engagement. We also thank the participants in the expert workshop Duke-Margolis convened held in November 2021 including Center researchers who provided valuable insight—Mark McClellan, Marianne Hamilton Lopez, Morgan Romine, Nitzan Arad, Beena Bhuiyan Khan, and Aparna Higgins. Finally, the authors wish to acknowledge Patricia Green from the Center for her communications guidance and Laura Hughes for her support in developing graphics and design for this white paper.

Any opinions expressed in this paper are solely those of the authors, and do not represent the views or policies  of any other organization external to Duke-Margolis. Funding for this work is made possible through the generosity of the Margolis Family Foundation, which provides core resources for the Center, as well as a combination of financial and in-kind contributions from Real-World Evidence Collaborative members, including Abbvie, Amgen, Bayer, Boehringer Ingelheim, Genentech, GlaxoSmithKline, Eli Lilly and Company, Merck, Novartis, Pfizer, and Teva, and Value for Medical Products Consortium members, including Alnylam, Amgen, BioMarin, bluebird bio, Boston Scientific, Edwards, Gilead, Gore, Medtronic, Merck, Novartis, Pfizer, REGENXBIO, and Sarepta. For more information on the Real-World Evidence Collaborative, visit https://healthpolicy.duke.edu/projects/real-world-evidence-collaborative.  

‍

For more information on the Value for Medical Products Consortium, visit https://healthpolicy.duke.edu/projects/value-medical-products-consortium.
‍

Rebecca Ray
Trevan Locke
Rachele Hendricks-Sturrup

‍

Disclosures

- Rachele Hendricks-Sturrup reports contract work with the National Alliance Against Disparities in Patient Health.

- Mark B. McClellan, MD, PhD, is an independent director on the boards of Johnson & Johnson, Cigna, AlignmentHealthcare, and Prognom IQ; co-chairs the Guiding Committee for the Health Care Payment Learning and ActionNetwork; and receives fees for serving as an advisor for Arsenal Capital Partners, Blackstone Life Sciences, and MITRE.

In the face of rapid innovation, a strong need exists for improved evidence generation capabilities to answer critical questions faced by regulators and payers. Systematically collecting real-world data to generate real-world evidence may provide the means to supplement the evidence available for newly approved novel therapies. Under the auspices of the Duke-Margolis Real-World Evidence (RWE) Collaborative and the Center’s Value for Medical Products Consortium, a multi-stakeholder working group considered how the health care community might align to ensure real-world data is capable of generating evidence that meets the needs of payers and regulators.

Primarily through the lens of three disease use cases (Alzheimer’s disease, cardiovascular disease, and spinal muscular atrophy), this paper explores considerations and provides recommendations to stakeholders for building a robust real-world data ecosystem through seven key themes.Additionally, data governance and resource needs are considered.

To build a robust real-world data ecosystem, policymakers and other key stakeholders should take the following steps:

- Provide resources needed to support electronic health record (EHR) interoperability between clinical research networks and central claims data repositories, such as those funded by Patient-Centered Outcomes Research Institute (PCORI);

- Provide additional funding for the United States Food and Drug Administration’s (FDA) Sentinel Initiative and other existing initiatives focused on safety and effectiveness(e.g., registries or observational studies that monitor the safety and effectiveness of therapies granted accelerated approval); ‍

- Reconsider bans on unique patient identifier funding and encourage the Department of Health and Human Services (HHS) through the Office of the National Coordinator for Health IT to advance unique patient identifier development;

- Harmonize stakeholder goals and initiatives by cultivating and supporting pre-competitive, public-private partnerships.

- If implemented, we believe the recommendations herein support robust post-market data collection that could enable efficient generation of evidence to supplement clinical trial results and further enable learning health care systems.

Introduction

As the pace of medical innovation increases and novel therapies are approved, regulators and payers face unprecedented pressure not only to surveil the safety and efficacy of such therapies, but also help ensure affordable access to those therapies. Indeed, regulators and payers have different roles in facilitating patient access to and clinically vetting and implementing new therapies. Yet, for new technologies and treatments intended to have long-term impacts, evidence of long-term health outcomes, safety and efficacy, and durability of treatment effects is often limited or insufficient to support decisions that ultimately impact patients and health plan beneficiaries.

Differences in evidence generation goals and data capabilities across regulators and payers ultimately affect broader patient access to new therapies

Most notably, available evidence on the real-world efficacy of new treatments at the time of regulatory approval may not be sufficient to payers who cover the costs of such treatments.This is especially true for therapies that have received accelerated regulatory approval. In 2021, 74 percent of all approvals were under expedited regulatory pathways and many of these are high-cost therapies for rare conditions.ⁱ Differences in evidence generation goals and data capabilities across regulators and payers ultimately affect broader patient access to new therapies. For example, U.S. FDA’s evidentiary goals are based on assessments of therapeutic safety and efficacy. On the other hand, payers make coverage and payment decisions based on medical necessity, cost, and the availability of comparator treatments. For this reason, the liminal state between regulatory approval and payers coverage often varies across therapeutics, potentially resulting in limited patient access to newly approved therapies and system- or practice-level variation in the clinical-implementation of such therapies.

Systematic real-world data (RWD) collection and curation may alleviate challenges associated with relying on limited real-world efficacy and safety evidence for certain therapeutics at the time of regulatory approval. Yet, many RWD collection efforts face limitations, including but not limited to misaligned priorities among regulators and payers as well as redundant and uncoordinated data collection, which ultimately lead to administrative burdens for providers and low value delivery to patients.

Further developing a shared RWD ecosystem, pieces of which exist already, can help alleviate these challenges. For example, the Office of theNational Coordinator for Health Information Technology (ONC) created the United StatesCore Data for Interoperability (USCDI) to support interoperability. The FDA has leveraged its Sentinel System for years to conduct safety surveillance of medical products. Related efforts in the FDA’sBiologics Effectiveness and Safety (BEST) Initiative and the National Evaluation System for healthTechnology (NEST) are also underway. In Europe, the Data Analysis and Real-World Interrogation Network (DARWIN-EU) is working to provide timely and reliable evidence on the use, safety, and effectiveness of medical products from real world health care databases across the European Union. Health Level 7 International (HL7) has been leading the effort to expand Fast Healthcare Interoperability Resources (FHIR)-based tools. Additional FDA post-market requirements as well as the Centers for Medicare and Medicaid Services’ (CMS) Coverage with Evidence Developmentⁱⁱ (CED)program, which imposes evidentiary requirements on certain medical products with new FDA approval, also help inform this ecosystem.

The efforts noted above can improve post-market evidence development by supporting the generation of important new knowledge for care decisions and give stakeholders a clearer understanding of the risks and benefits of a new intervention. For example, BEST Initiative has explored rates of myocarditis and pericarditis from COVID-19 vaccines using claims databases.ⁱⁱⁱ

In addition, instances exist in which stakeholders’ evidentiary goals were aligned through a national registry when CMS issued national Medicare coverage for technology with CED.^iv As these efforts advance, it will be important to identify opportunities to scale the ecosystem as well as best practices for the community.

To further explore these concepts and identify opportunities for harmonization between regulators and payers, members of the Duke-Margolis RWE Collaborative and Value for Medical Products Consortium explored the landscape of shared post-market evidence needs, including efforts that are underway to set meaningful data collection standards. This exploration was informed by three disease-based use cases — monoclonal antibody treatments for Alzheimer’s disease, gene therapy treatments for spinal muscular atrophy (SMA), and therapies and medical devices to treat cardiovascular disease (CVD).
‍

‍This paper is informed by a November 16, 2021, private workshop, “Aligning to Address Shared EvidentiaryNeeds in Data Collection and Characterization,”hosted by the Duke-Margolis Real-World Evidence Collaborative and the Value for Medical Products Consortium, by several regular working group and stakeholder calls with members of these groups, and by literature cited throughout this document. Please see the end of this document for a list of participants in the working group.

‍

Use Case Background

SMA is a rare genetic neuromuscular disease that results in muscular weakness and significant disability, leading to a drastically shortened lifespan. The FDA has approved three treatments for SMA, including Zolgensma, a gene replacement therapy, and Spinraza and Evrysdi, two therapies that enhance the SMN2 gene. In particular, the approval of Zolgensma represents the beginning of a new paradigm for treating SMA. As we highlight in this paper, a robust RWD ecosystem can help collect data on the safety and effectiveness of these therapies over time.

CVD are a group of disorders of the heart and blood vessels. It is the leading cause of death affecting approximately 82.6 million people in the United States. There are many types and combinations of drugs available to treat CVD as well as eight types of devices to treat the heart. Given the breadth of available therapies to treat CVD, there is an opportunity to gather data from patients in clinical and prospective studies as well as passive RWD from EHRs since the endpoints are definitive for CVD products (e.g., prevention of myocardial infarction). This data can show the value of new products versus older products and presents a starting point to compare clinical trial data and RWE from an outcomes standpoint. However, several open questions remain, including how to collect data on combinations of therapies and track patient adherence.

In Alzheimer’s disease, aducanumab, a monoclonal antibody treatment, was recently granted approval through the FDA’s accelerated approval program. Under this program, regulatory approvals may be based on a determination that the product has an effect on a surrogate end point or on a clinical endpoint that can be measured earlier than irreversible morbidity or mortality, and that is reasonably likely to predict an effect on irreversible morbidity or mortality. In the case of aducanumab, the surrogate endpoint was reduction in beta-amyloid plaque, an indicator that has been frequently evaluated as a marker for disease severity in patients with Alzheimer’s disease. However, evidence linking this surrogate endpoint to improved cognitive and functional outcomes for Alzheimer’s disease patients is largely inconclusive.^v The need for additional evidence on accelerated approvals like this one represents an opportunity to develop new real-world endpoints and leverage RWD to contribute to the broader totality of evidence for novel therapies.

As a result of this process, seven themes emerged as challenges and opportunities to building a robust data ecosystem meant to leverage RWD and RWE to support regulatory and payer decision-making:

‍

RWD Ecosystem to Meet Shared Evidentiary Needs

This paper explores these themes, discussing important considerations and providing recommendations related to each for building a robust data ecosystem equipped to support both regulatory and payer decision-making through the use of RWD and RWE. Additionally, use-case agnostic considerations for data governance as well as needs for additional resources to support ecosystem development are explored.

Many opportunities in the post-regulatory approval setting exist to build an ecosystem harmonized for payer and regulatory decision making. A single-administration therapy for SMA, like Zolgensma, warrants not only an understanding among regulators and payers about the immediate benefits of Zolgensma but also the long-term efficacy and safety of the therapy, which today remains unknown. Continuous and strategic RWD collection maybe useful to regulators seeking to monitor the long-term safety of early-administered, one-time therapies like Zolgensma as well as for payers who seek to understand if and how such treatments provide immediate impact and long-term value. Likewise, longitudinally following Alzheimer’s therapies in the real-world setting will offer more evidence of the long-term value of these products.

For CVD, therapies and devices often need to be assessed or compared post-regulatory approval to determine the best patient care strategies. This need is especially true given the breadth of available therapies and devices to treat CVD. Indeed, opportunities exist to explore ways in which RWE can be leveraged to determine if existing therapies are efficacious in CVD population profiles excluded from preceding clinical trials. For instance, CVD drug efficacy trials typically exclude certain populations(e.g., geriatric patients and or patients with co-morbid conditions) from pre-market randomized controlled trials or pivotal studies for scientific reasons. RWE-driven studies conducted in the post-market setting, including studies that examine patient-level outcomes (e.g., patient-reported outcomes) can be useful to examine the comparative and/or combined effectiveness of CVD treatments.

A major barrier to building the envisioned RWD ecosystem is the lack of interoperable systems in the health care setting. In the SMA space, while multiple therapies can be captured in one registry, several registries currently exist for SMA.^vi,vii,viii Patients who receive Zolgensma will likely be covered under either public or private insurance and thus, potentially listed in separate registries. It is important for these separate registries to be interoperable to enable larger scale data analysis. Similarly, for Alzheimer’s disease, registries for new monoclonal antibody treatments should be interoperable with data collected in untreated patients at Alzheimer’s Disease Research Centers and at other medical centers to enable valuable means of comparing outcome for treated and untreated patients. In CVD care, RWE gaps tend to emerge when a registry is anchored to a specific clinical care site and a patient either changes provider locations or dies away of the facility, (e.g., emergency department, home, etc.)  Thus, challenges like patient data portability and tracking and linking patient data from multiple sources are to be expected.

In cases where substantial value misalignment exists across these registries and other RWD systems, accompanying variation in the quality, interoperability, and robustness of those systems, is likely, commonly due to resource availability and/or constraints faced by the system owners and/or operators. In such cases, incentives should be structured to facilitate and sustain value alignment across disparate RWD systems.

To begin addressing some of these challenges,ONC and CMS released interoperability and access rules in 2020 that establish policies meant to increase patient access to data and promote interoperability between health data sources. CMS’ rule requires payers to build Application Programming Interfaces (APIs) that would allow payer-to-payer information exchange. Yet, payers are concerned about the operational challenges of implementing this rule and any ensuing risks to data quality in the absence of specific standards.^ix CMS is currently exercising enforcement discretion on this provision of the rule while more conversations take place with key stakeholders. Though initially challenging to implement, the approaches outlined in these ONC and CMS interoperability and access rules are critical to creating a robust data infrastructure that balances access with privacy.

Linking separate RWD sources represents an opportunity to build a more robust ecosystem. For example, in CVD, there are opportunities to leverage the current breadth of data sources, including clinical trial data sources and EHRs, as a strategy to examine the real-world comparative and/or combined effectiveness of therapeutics and devices. This strategy, however, likely will present ongoing challenges (e.g., technical, legal, etc.) associated with data access, linkage and management across multiple systems. To meet these challenges, an ideal RWD structure would include access to real-time data that can be collected over time to address longitudinal data needs. It also would include components to reduce the burden of manual data entry and enhance linkages across other databases to maximize data analysis capabilities further.

Importantly, the lack of unique patient identifiers (UPIs) in the United States is a hurdle to building an interconnected, multi-stakeholder data system. A legislative ban on research funding for UPIs has limited federal-level work on developing a robust solution. Though efforts have been made to overturn this ban, none have been successful. Policymakers should consider the benefits of UPIs, support the mitigation of any risks associated with overturning this ban, and encourage HHS through ONC to advance UPI development in the United States. While not a perfect solution alone, UPIs are useful for linking or combining data sources, supporting the creation of clinical trial infrastructure, and ultimately, empowering decision makers striving to create learning healthcare systems.

‍Collecting longitudinal data is often a challenge as patients change care settings or providers. The lack of longitudinal data impairs the ability to make decisions in a range of disease areas, including cardiovascular care, as well as long-term follow upon gene therapies. Well-implemented UPIs, or other technologies and policies supporting the connection of disparate data sources, have the potential to unlock substantial efficiencies in the ability to collect and use longitudinal data.

Using RWD to generate RWE is hampered in many settings by the lack of standardized endpoints and outcome measures. In Alzheimer’s disease, for example, agreed upon endpoints and outcome measures are needed to investigate therapies in the real world. Using one standardized outcome, like admittance to a nursing home or use of 24-hour in-home care could be a useful outcome to evaluate longitudinal functional outcomes following Alzheimer’s disease treatment.Dementia rating scales to collect and track changes in cognitive status over time, alongside patient-functional outcome measures, might help ascertain a therapy’s real-world value. However, some of the assessment measures used in those treatments’ pre-approval clinical trials, including aducanumab, are complex to administer and are generally not used outside of clinical research settings. For the other monoclonal antibody treatments in the class that may be covered in prospective studies, a need exists to agree on other outcome measures that can be used more easily in real-world settings.Gathering diverse data with definitive endpoints from patients (e.g., prevention of myocardial infarction) is also one strategy to examine the real-world comparative and/or combined effectiveness of a broad array of CVD products.

With SMA and other single-use gene therapies, short-, intermediate-, and long-term outcomes measures to evaluate durability should be considered. In the short- and intermediate-term, emphasis should focus on safety measures to understand how the therapy is attenuating the disease while also allowing the nervous system to recover and respond to facilitate long-term survival. In the long-term, emphasis should focus on quality-of-life measures, how the disease impacts diagnosed children and their families, and the overall cost to society. Ultimately, outcome measures should be simple yet valuable for care providers and families. Past work by Duke-Margolis and its Real-World Evidence Collaborative includes a roadmap that explores key considerations for developing endpoints for use in the real-world.^xii

Related to the development of these standardized endpoints and outcomes measures is the importance of patient and provider education from drug developers to help them understand the validated measures and endpoints used to evaluate the therapy. Differences in capabilities among and across health systems, medical centers, and smaller clinics to collect RWD is likely. Educational efforts around validated measures and endpoints can help mitigate administrative difficulties associated with RWD capture among large and small patient-provider settings. For example, existing efforts to evaluate RWD sources through public-private partnerships, including the ROADMAP project in Alzheimer’s disease, have supported increased disease understanding and may help streamline decision-making processes.^xiii

Stakeholders identified the power of broad partnerships to facilitate the collaborations necessary to address the challenges and take advantage of the opportunities highlighted herein.In Alzheimer’s disease, opportunity exists for stakeholders to collaborate in a pre-competitive environment to build a registry that includes abroad and representative patient population to evaluate new therapies against conventional therapies. However, capturing health outcomes using large amounts of data from a single data source comes with certain challenges. In an example with SMA, a longitudinal cohort study of more than 3 million commercial insurance members reported that one in five members dis-enrolled each year, making it difficult to track long-term patient outcomes following coverage decisions for FDA-approved therapeutics. Therefore, registries that would have the broadest possible population also should harness RWD from a wide variety of sources. To accomplish this, partnerships among registry owners or managers, industry, and other key stakeholders are critical to establish shared short-, intermediate-, and long-term evidentiary standards and goals for one-time therapies like Zolgensma, as well as subsequent and/or concomitant therapies that might be required depending on patient needs.

Data partnerships typically require buy-in and support among and between policymakers, public agencies, and private enterprises. Pilot programs for such partnerships are key to develop best practices and frameworks, inform policy, set grant making priorities, and other important initiatives. Such pilot programs would be conducted best in partnership with health systems and other stakeholders that own and manage relevant or well-suited RWD systems. Some early efforts like the CardioHealth Alliance,^xv The NationalTreatment and Diagnostic Alzheimer’s Registry^xvi and ROADMAP in Alzheimer’s disease,^xvii and theCanadian Neuromuscular Disease Registry^xviii may provide instructive models for further partnerships in other areas.

‍A note on analytic methods:

Traditional RWD analyses employ biostatisticians, clinicians, epidemiologists, economists, and other experts to employ classical descriptive and statistical analytic methods. While this paper does not explore analytic methods issues in depth, it is important to note that it will be necessary for stakeholders to partner to develop and refine advanced analytic methods that aggregate data and to build algorithms for each use case. Advanced analytics have the potential to generate evidence for stakeholders by combining data sources and processing large amounts of data (e.g., genetic and genomic information). Therefore, improving analytical methods knowledge can enable broad consensus/understanding of when RWE is appropriate, what questions RWE can answer that traditional clinical evidence cannot, and how to ensure objective evaluation of the RWE quality stakeholders should build data infrastructure capabilities with these analytic needs in mind. As part of the partnerships discussed here, stakeholders in a given disease space should work together to ensure the appropriate analytic methods are identified.

One foundation for partnerships to build on is existing databases and registries. Existing registries, such as The Alzheimer’s Disease Neuro imaging Initiative (ADNI), the IDEAS study on amyloid PET, and the National Alzheimer’sCoordinating Center (NACC), which coordinate data collection and foster collaborative research among Alzheimer’s Disease Research Centers across the country, could serve as a starting point to build more robust data collection mechanisms for both larger and smaller clinics.In the CVD space, Yale has made promising steps to accomplish many of the goals explored herewith the Yale New Haven Health System Heart

Failure Registry.^xix Similarly, existing registries in the SMA and CVD space likewise could be leveraged.xx, xxi, xxii The FDA’s recently disseminated draft guidance on using registries to inform regulatory decisions covers considerations regarding a registry’s fit-for-use in regulatory decision-making, considerations to link registries to other RWD sources, and considerations to support FDA review of submissions that include registry data.^xxiii Once final, the FDA’s guidance maybe helpful to inform the development of registry initiatives suited to meeting shared evidentiary goals to monitor monoclonal antibody treatment outcomes in Alzheimer’s disease patients.

Collecting high-quality, complete, fit-for-purpose data is critical to support a robust RWD ecosystem. However, collecting and storing such data comes with costs in time and resources that should not be underestimated, especially in terms of determining who would bear the costs. Misalignment among stakeholders around who benefits most from the data and who pays the costs ultimately affects the overall value proposition, especially in instances where long-term data collection is needed or would be best. To address this dilemma as it would occur in real-world practice, effective incentives are needed to fully engage stakeholders, including providers and patients, who can contribute to generation and collection of such data.

Participants in the CVD breakout session considered ways to structure payment models and incentives for long-term RWD collection and analysis, focusing on value-alignment needed among health systems and other key stakeholders, including patient data registries, to accomplish this goal. For instance, there are several existing patient data registries for CVD that may or may not be interoperable with EHR systems.^xxiv As mentioned above, sustainable incentives should be used to facilitate value alignment across disparate RWD systems.

Health are providers who struggle to find time and resources needed to deliver effective care often struggle with the demands of tedious and time-consuming data entry procedures. Much of the data collection burden, especially for EHR and registry data, falls on clinicians or registry owners. Health system workflows should ensure that such burdens are minimized through sufficient staffing and resources. Automating some data collection activities, such as prior authorization determinations, is one option to streamline data collection. In CVD, SMA, and other disease areas, consideration should be given to structuring payment models and incentives for long-termRWD collection and analysis, focusing on value-alignment needed among health systems and other key stakeholders, including patient data registries, to accomplish this goal. For instance, several patient data registries for CVD exist that may or may not be interoperable with EHR systems.

Using high-quality and validated wearables, survey measures, or other instruments to collect patient-reported data is also another area ripe for incentives. Patient-level or reported data collected in real-world settings would disperse the data collection burden, rendering value especially in cases where data might feed into data registries and contribute to the body or totality of evidence on the safety and effectiveness of new therapies granted accelerated approval by the FDA. Therefore, payers and regulators should collaborate to revisit their guidelines, guidance documents, and policies to align on what is needed to cover and monitor patient populations appropriately using RWD.^xxvi

Without proper reimbursement or payment to support the significant amount of clinician time and effort, the goal of collecting and synthesizing high quality, complete, fit-for-purpose data becomes lost

Payers also could explore strategies to reimburse clinicians for the time required to input data of value to the payers. Without proper reimbursement or payment to support the significant amount of clinician time and effort, the goal of collecting and synthesizing high quality, complete, fit-for-purpose data becomes lost.^xxv CMS also might consider additional incentives, like requiring participation in a data registry or CED program, to obtain reimbursement for prescribing or dispensing certain treatments or for using intuitive, structured data systems to help reduce burden or burnout from tedious data entry processes.

‍

ADDITIONAL CONSIDERATION #1: Data Governance and Management

When building a shared evidence infrastructure that considers the teams above, data and system governance must be forethought. Data and system governance are challenging yet rewarding endeavors when the right questions are addressed early in the data system development process. For example, questions might include:

- Who makes decisions about how the data are accessed and used?

- Who is responsible for ensuring the data are fit for use?

- Who owns the data and where do the data live?

- Who analyzes the data and how?

- How are necessary privacy and transparency considerations implemented?

As stakeholders contemplate these questions, they should consider and determine which elements of governance should be centralized either within government agencies or disease-specific organizations and those that should remain decentralized. A distributed data network, like the FDA’s Sentinel Initiative, is one privacy-preserving model to consider--as the data does not leave the environments of individual data partners. Instead, the Sentinel Operations Center queries data partners and then receives de-identified query results. The Sentinel System also uses a common data model to ensure all data is formatted consistently across the distributed data network.

It is vital that patients are central in any RWD ecosystem. While most RWD generated today comes from clinician inputs (e.g., registry, claims, and EHR data), an increasing amount of RWD is likely to come directly from patients through their wearable devices, phones, and other digital health tools. As the RWD ecosystem is built, scalable approaches to collecting data from patients and linking it to other sources of RWD or clinical data are important for building more robust datasets. Furthermore, stakeholders must ensure that data collection from patients does not exacerbate health inequities, so such approaches must consider vulnerable populations including those with limited access to digital health technology. Accomplishing this approach to data collection in a patient-centered way will require new informed consent mechanisms and direct patient engagement in research. It is important that patients know what their data are being used for and see the results and benefits of contributing their data.

As the amount and types of real-world health data grow, additional consideration must be given to ensuring that sensitive personal health data is properly protected. However, many novel sources of insightful health data are not covered or protected by the Health Insurance Portability and Accountability Act (HIPAA). Any effort to build RWD ecosystems must take privacy into account, while allowing patients access to their data and a voice in determining how their data are used. In the absence of clear federal legislation on privacy protections for health-related data, take holders should take it upon themselves to ensure that any use of sensitive data has privacy protections in place.

‍

ADDITIONAL CONSIDERATION #2: Need for Additional Resources and Funding

Federal support for the various data infrastructure considerations discussed above could reduce the burden of data collection costs as a simple solution in the near term, thus enabling easier and broader data collection. For long-term sustainability of such initiatives, more options, possibilities, and opportunities should be explored. Existing federal efforts also could be enhanced with additional funding. Additional resources at PCORI could support EHR interoperability between its clinical research network participants and create a central claims data repository. Within FDA, additional funding for its Sentinel Initiative could accelerate existing efforts to increase the system’s ability to answer effectiveness questions in addition to its more established safety surveillance role. These efforts could be combined with additional resources at CMS to fund registries or observational studies that assess the safety and effectiveness of new therapies approved by FDA accelerated approval. Additionally, increasing access to Medicare data, even at a cost to researchers, could provide valuable support for furthering an RWD ecosystem. Collaboration among these federal efforts will help ensure resources are used efficiently to answer a broader range of questions.

Finally, value-based payment (VBP) models can be leveraged to encourage post-market evidence development among key stakeholders. Ideally,VBP models built for this purpose should be supported on a long-term basis to cultivate a sophisticated data infrastructure that aligns with regulator and payer post-market evidence needs. Key stakeholders involved in the curation of value-based payment models should collaborate to ensure that payment models are feasible both financially and practically and without a significant cost and time burden to a single party or group. Furthermore, payers need to be able to adapt to evolving evidence as clinical care contexts changeover time. Having clear expectations for payer evidence needs in post-market settings will enable medical product developers to make plans forRWE generation well in advance.

‍

The challenges and opportunities summarized herein and in prior work^xxvii indicate a lack of a robust data ecosystem as one of the major barriers to broader adoption and use of high quality RWD. While federal agencies and policymakers have a vital role in supporting the development of this ecosystem, those that generate or use RWD also must offer important contributions to inform meaningful next steps in this process. The seven key themes and additional considerations around data governance and resource provision detailed here indicate that much work remains to build infrastructure for data linkage and interoperability, establish governance structures and new data partnerships, and develop relevant real-world endpoints and incentives for providers and patients, all while protecting and educating patients. While health care systems, drug sponsors, governments, and payers determine how to move forward from the COVID-19 pandemic, now is the time to build the equitable data ecosystem we need for evaluating modern therapies in the real-world. However, no one stakeholder can do it alone. Building this needed ecosystem will take buy-in and alignment between stakeholders across the life cycle of medical products.

Shared Evidentiary Opportunities Working Group Members

The working group was composed of representatives of member organizations of the Duke-Margolis Real-World Evidence Collaborative and the Duke-Margolis Value for Medical Products Consortium as well as additional stakeholder experts. We thank the working group again for informing the development of this paper.

‍Aylin Altan

Optum Labs

Haider Andazola

Foley Hoag

Bruce Artim

Regenxbio

Preeti Bajaj

Genentech

Thomas Barker

Foley Hoag

Ginny Beakes-Read

Amgen

Marc Berger

ISPOR

Lindsay Bockstedt

Medtronic

Mac Bonafede

Veradigm

India Bowman

PatientsLikeMe

Claire Brunken

Novartis

Kenneth Carson

Tempus

Harold Carter

Cigna-Express Scripts

Sree Chaguturu

CVS-Aetna

Robin Christoforides

Boehringer Ingelheim

Francesca Cook

Regenxbio

Bill Crown

Brandeis University

Greg Daniel

Eli Lilly and Company

‍Lisa Ensign

Medidata Solutions

Josh Gagne

Harvard University/ISPE

Aaron Galaznik

Medidata Solutions

Martin Gold

Edwards Lifesciences

Liesl Hargens

Boston Scientific

Heather Haworth

Edwards Lifesciences

Carolyn Hickey

Sarepta Therapeutics

Ravi Iyer

Teva Pharmaceuticals

Christine Jackson

Medtronic

Ashley Jaksa

Aetion

Ritesh Kumar

Merck & Co.

Gracie Lieberman

formerly Genentech

Vadim Lubarsky

Novartis

Nirosha Mahendraratnam Lederer
Aetion

Nicole Mahoney

Novartis

Deepa Malhotra

Pfizer

Nell Marshall

Evidation Health

Monica McClain

Genesis Research

‍Sean Mcelligott

Novartis

Newell McElwee

Boehringer Ingelheim

Annie McNeill

AbbVie/ISPE

Anne-Marie Meyer

University of North Carolina

Stephen Motsko

Amgen

Irene Nunes

Flatiron Health

Megan O’Brien‍

Merck & Co.

Laura Okpala

Gilead Sciences

Sally Okun

Clinical Trials TransformationInitiative

Greg Poulsen

Intermountain Healthcare

Sulabha Ramachandran

GlaxoSmithKline

Sadhya Rao

Blue Cross Blue Shield Massachusetts

Jeremy Rassen

Aetion

Mary Beth Ritchey

Med Tech Epi/ISPE

Andres Rodriguez

GlaxoSmithKline

Gail Ryan‍

Harvard Pilgrim Health Care

Debra Schaumberg

Evidera

‍Kristin Sheffield

Eli Lilly and Company

Lauren Silvis

Tempus

Jason Simeone

Cytel

Maria Stewart

Boston Scientific

‍Aniketh Talwai

Medidata Solutions

Ben Taylor

Aetion

David Thompson‍

OPEN Health

Aracelis Torres

Verana Health

Vani Vannappagari

GlaxoSmithKline - ViiV

Dick Wilke‍

ISPOR

David Wormser

Novartis

2021 Real-World Evidence Collaborative Advisory Group

‍Aylin Altan

OptumLabs

Marc Berger

Retired

Barbara Bierer

Multi-Regional ClinicalTrials Center of Brighamand Women’s Hospital and Harvard

Brian Bradbury

Amgen

William Capra

Genentech

Adrian Cassidy

Novartis

Bill Crown

Brandeis University

Riad Dirani

Teva Pharmaceuticals

Nancy Dreyer

IQVIA

Andrew Emmett

Pfizer

‍John Graham

GlaxoSmithKline

Stacy Holdsworth

Eli Lilly and Company

Solomon Iyasu

Merck & Co.

Ryan Kilpatrick

AbbVie

Lisa LaVange

University of North Carolina

Christina Mack

ISPE

Irene Nunes

Flatiron Health

Sally Okun

Independent Consultant

Bray Patrick-Lake

Evidation

Eleanor Perfetto

National Health Council

Richard Platt

Harvard Medical School

Jeremy Rassen

Aetion

Subhara Raveendran

PatientsLikeMe

Stephanie Reisinger

Veradigm

Debra A. Schaumberg

Evidera

Thomas Seck

Boehringer Ingelheim

Lauren Silvis

Tempus

David Thompson

Syneos Health

Richard Willke

ISPOR

Marcus Wilson

HealthCore

Listed 2021 member affiliations may not reflect current affiliations. For a current roster of the Duke-Margolis Real-World Evidence Collaborative’s Advisory Group, please visit the RWE Collaborative page on the Duke-Margolis Center for Health Policy website.

‍

‍

Value for Medical Products Consortium

This paper was additionally informed by the expert collaborators in the Duke-Margolis Value for MedicalProducts Consortium. In previous publications, this group was referred to as the Value-Based PaymentConsortium. We thank the members of the Advisory Group for informing the development of this paper. The following list reflects the current membership as of June 2022.

‍Alan Balch

Patient Advocate Foundation

Thomas Barker

Foley Hoag

Mary Bordoni

Bristol Myers Squibb

Chuck Bucklar

BioMarin Pharmaceutical

Sree Chaguturu

CVS Caremark, CVS-Aetna

Alexandra Clyde

Medtronic

Francesca Cook

Regenxbio

Mary Coppage

Edwards Lifesciences

Gregory Daniel

Eli Lilly and Compancy

Darin Gordon

Gordon & Associates

Jeffrey S. Guy

Hospital Corporation of America

Esther Krofah

FasterCures

Leigh Ann Leas

Novartis

Steve Miller

Express Scripts

Barbara Minton

IngenioRx

Megan O’Brien

Merck & Co.

‍Gregory Poulson

Leavitt Partners

Rekha Ramesh

Gilead Sciences

Sandhya Rao

Blue Cross Blue Shield Massachusetts

Lewis Sandy

UnitedHealth Group

Michael Sherman

Point32Health

Surya Singh

Singh Healthcare Advisors

Robert Smith

Pfizer

‍Participants at the November 16, 2021, Workshop on Shared Evidentiary Opportunities

This paper was additionally informed by the expert collaborators in the Duke-Margolis Value for MedicalProducts Consortium. In previous publications, this group was referred to as the Value-Based PaymentConsortium. We thank the members of the Advisory Group for informing the development of this paper. The following list reflects the current membership as of June 2022.

‍Samiah Al-Zaidy

Independent Consultant

Alirez Atri

Banner Sun HealthResearch Institute

Danielle Bargo

Flatiron Health

Ralph Brindis

University of California,San Francisco

Julie Chandler

Eli Lilly and Company

Amy Delozier

Eli Lily and Company

Brian Hitt

Edwards Lifesciences

Schuyler Jones

Duke University

Fred Masoudi

Ascension

Leah McGrath

Pfizer

Erik Musiek

Washington University

Matthew Roe

Verana Health*

Keely Scamperle

W. L. Gore & Associates*

Sreek Vemulapalli

Duke University

‍Loukas Xaplanteris

Biogen

Noah Zimmerman

Tempus


Observers:

John Concato

Food and Drug Administration

Khair Elzarrad

Food and Drug Administration

Kenneth Quinto

Food and Drug Administration

Donna Rivera

Food and Drug Administration

*At time of workshop

‍References

i Craven J. “FDA approved more first-in-class drugs, gave more accelerated approvals in 2021.” Regulatory AffairsProfessionals Society. January 7, 2022. https://www.raps.org/news-and-articles/news-articles/2022/1/fda-approved-more-first-in-class-drugs-more-with-a

ii  Coverage with Evidence Development. Centers for Medicare and Medicaid Services. https://www.cms.gov/Medicare/Coverage/Coverage-with-Evidence-Development

iii  Wong H. ”Safety Surveillance of COVID-19 Vaccines in Children and Adolescents.” Vaccines and Related BiologicalProducts Advisory Committee Meeting. U.S. Food and Drug Administration. June 14, 2022. https://www.fda.gov/media/159224/download

iv  Rumsfeld JS, Holmes DR Jr, Stough WG, Edwards FH, Jacques LB, Mack MJ. “Insights from the early experience of the Society of Thoracic Surgeons/American College of Cardiology Trans catheter Valve Therapy Registry.” JACC Cardiovasc Interv. 2015 Mar;8(3):377-381. doi: 10.1016/j.jcin.2014.09.022. Epub 2015 Feb 18. PMID: 25703888. https://pubmed.ncbi.nlm.nih.gov/25703888/

v  Clinical Review Report: Nusinersen (Spinraza): (Biogen Canada Inc.): Indication: Treatment of patients with 5qSMA [Internet]. Ottawa (ON): Canadian Agency for Drugs and Technologies in Health; 2018 Jan. Appendix 7,Clinical Features, Epidemiology, Natural History, and Management of Spinal Muscular Atrophy. Available from: https://www.ncbi.nlm.nih.gov/books/NBK533981/

vi Finkel, Richard S et al. “RESTORE: A Prospective Multinational Registry of Patients with Genetically ConfirmedSpinal Muscular Atrophy - Rationale and Study Design.” Journal of neuromuscular diseases vol. 7,2 (2020):145-152. doi: 10.3233/JND-190451

vii International SMA Consortium Spinal Muscular Atrophy Patient Registry (iSPAC SMA Registry). Nemours. https://www.nemours.org/pediatric-research/clinicaltrials/neurology-1238325.html

viii Hodgkinson V, et al. “A National Spinal Muscular Atrophy Registry for Real-World Evidence.” The CanadianJournal of Neurological Sciences Inc., Volume 47, No. 6. November 2020.

ix Gordon WJ and Rudin RS. “Why APIs? Anticipated value, barriers, and opportunities for standards-based application programming interfaces in healthcare: perspectives of US thought leaders.” JAMA Open, Volume 5,Issue 2. July 2022. https://academic.oup.com/jamiaopen/article/5/2/ooac023/6564215

x Brooks-LaSure, Chiquita. ”Interoperability and the Connected Health Care System.” Centers for Medicare andMedicaid Services. December 8, 2021. https://www.cms.gov/blog/interoperability-and-connected-health-care-system

xi ”Enhanching Patient Matching is Critical to Achieving Full Promise of Digital Health Records.” The Pew CharitableTrusts. October 2, 2018. https://www.pewtrusts.org/en/research-and-analysis/reports/2018/10/02/enhanced-patient-matching-critical-to-achieving-full-promise-of-digital-health-records

xii “A Roadmap for Developing Study Endpoints in Real-World Settings.” Duke-Margolis Center for Health Policy.August 28, 2020. https://healthpolicy.duke.edu/sites/default/files/2020-08/Real-World%20Endpoints.pdf

xiii Gallacher J, et al. “Challenges for Optimizing Real-World Evidence in Alzheimer’s Disease: The ROADMAP Project.”Journal of Alzheimer’s Disease. January 22, 2019. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6398537/

xiv Fang H, Frean M, Sylwestrzak G, Ukert B. Trends in Disenrollment and Reenrollment Within US CommercialHealth Insurance Plans, 2006-2018. JAMA Netw Open. 2022;5(2):e220320. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2789399

xv CardioHealth Alliance. https://cardiohealthalliance.org/

xvi “Alzheimer’s Association Announces National Effort to Collect “Real World” Data on Newly-Approved Treatments. ”Alzheimer’s Association. November 9, 2021. https://www.alz.org/news/2021/alzheimers-association-announces-national-effort

xvii Janssen O, et al. “Real-world evidence in Alzheimer’s disease: The ROADMAP Data Cube.” Alzheimers Dement.2020 Mar;16(3):461-471. doi: 10.1016/j.jalz.2019.09.087. Epub 2020 Jan 6. PMID: 32157788.

xviii Hodgkinson VL, et al. “A National Spinal Muscular Atrophy Registry for Real-World Evidence.” 2020 Nov;47(6):810-815. doi: 10.1017/cjn.2020.111. Epub 2020 Jun 4. PMID: 32493524; PMCID: PMC7656664.

xix Ahmed T and Desai N. R. “Reimagining Evidence Generation for Heart Failure and the Role of Integrated HealthCare Systems.” Circulation: Cardiovascular Quality and Outcomes, Vol. 15, No. 4. American Heart Association.March 11, 2022. https://www.ahajournals.org/doi/full/10.1161/CIRCOUTCOMES.121.008292?af=R

xx Finkel, Richard S et al. “RESTORE: A Prospective Multinational Registry of Patients with Genetically ConfirmedSpinal Muscular Atrophy - Rationale and Study Design.” Journal of neuromuscular diseases vol. 7,2 (2020): 145-152.doi:10.3233/JND-190451

xxi International SMA Consortium Spinal Muscular Atrophy Patient Registry (iSPAC SMA Registry). Nemours. https://www.nemours.org/pediatric-research/clinicaltrials/neurology-1238325.html

xxii Hodgkinson V, et al. “A National Spinal Muscular Atrophy Registry for Real-World Evidence.” The CanadianJournal of Neurological Sciences Inc., Volume 47, No. 6. November 2020.

xxiii  “Real-World Data: Assessing Registries to Support Regulatory Decision-Making for Drug and BiologicalProducts: Guidance for Industry.” Food and Drug Administration, U.S. Department of Health and HumanServices. November 2021. https://www.fda.gov/media/154449/download

xxiv  Cardiac Amyloidosis Registry Study. CARS – Clinical Study. Duke Health. https://www.dukehealth.org/clinical-trials/directory/pro00107308

xxv Zhang J, et al. ”Best Practices in the Real-World Data Life Cycle.” PLOS Digital Health. January 18, 2022. https://doi.org/10.1371/journal.pdig.0000003

xxvi ”Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims.”Guidance for Industry. Food and Drug Administration. December 2009. https://www.fda.gov/media/77832/download

‍

‍

‍

‍

‍