https://orcid.org/0000-0002-3969-4279
Exponential scientific data growth presents challenges and opportunities for addressing complex public health issues like the opioid epidemic and chronic pain management. Despite the vast amount of research conducted globally, many datasets remain inaccessible or underutilized due to publication access policies and stringent data use agreements. The amount of data generated through research activities is enormous. Limited access to scientific datasets stifles discovery and delays the translation of proven scientific advances into real-world applications.1 To address these challenges, we argue for the critical importance of open science ecosystems, using the National Institutes of Health Helping to End Addiction Long-term® Initiative (NIH HEAL Initiative®) as a case study. We discuss how building community around data can accelerate scientific discovery by enabling dataset integration, increasing statistical power, and fostering interdisciplinary collaboration.
Open science ecosystems represent a fundamental shift in how research is conducted, shared, and utilized. An open science ecosystem is a comprehensive research environment that combines technological infrastructure, standardized protocols, and collaborative networks to enable transparent sharing and integration of scientific data, methods, and findings. These interconnected networks integrate data collection, storage, processing, analysis, and use across organizations while adhering to FAIR (Findability, Accessibility, Interoperability, and Reusability) principles. The ecosystem encompasses not just the technical components for data sharing, but also the human elements: Researchers, institutions, funding bodies, and community partners who work together under shared governance frameworks and data standards to accelerate scientific discovery. The NIH HEAL Initiative® demonstrates how such ecosystems can accelerate discovery in critical public health areas through transparent research methods, open access to data and findings, collaborative approaches to complex problems, and standardized data elements enabling cross-study analyses.
Shifting culture and confronting barriers
Data sharing is crucial to HEAL's mission of addressing the interconnected public health crises of chronic pain and opioid use disorder (OUD). While appropriate medical use of opioids remains important for pain management, the rise in OUD presents distinct challenges requiring comprehensive research approaches. By distinguishing between therapeutic opioid use and OUD, researchers can better target interventions and support both pain management and addiction treatment needs.
HEAL researcher and community partner data dashboards actively identify and monitor emerging threats in the drug supply, such as highly potent fentanyl and xylazine.2 This capability enables rapid resource deployment where needed, potentially saving lives.
More broadly, combining data sets can increase sample size, provide more statistical power to make conclusions, help test hypotheses that a single data set alone would be insufficient to generate, and increase result generalizability. Data can be analyzed differently by multiple groups to answer numerous research questions and consolidate trust in the validity of reported results.3 Secondary analysis of shared data can also uncover patterns in disparate datasets, such as signatures of different chronic pain types,4 genetic factors that may underlie similar pain phenotypes (or symptoms), or qualities likely to make an individual respond to treatment.5
The impact of data inaccessibility on research progress is particularly evident in pain management studies. For example, when clinical trial data remains siloed, researchers cannot identify subtle patterns in treatment response across different patient populations. This limitation has historically hampered the development of personalized pain management approaches and delayed the identification of risk factors for OUD development. Through the HEAL Data Ecosystem, researchers have begun breaking down these barriers. In one instance, the combination of data sets from multiple pain management clinics revealed previously unrecognized patterns in treatment outcomes across diverse demographic groups, leading to more targeted intervention strategies.6
The competitive nature of scientific research, characterized by a culture of individual studies and siloed networks, has hindered widespread acceptance and adoption of open science ecosystems. Additionally, legitimate privacy concerns present complex challenges to carefully navigate. Many scientists and research teams remain skeptical or reluctant to participate in open science ecosystems, due in part to the current passive data sharing culture. Active data sharing, in contrast, requires adjusting to externally imposed standards. Creating an inclusive scientific community around data requires both interdisciplinary blending of people and unique approaches. Technical aspects of a functional open science ecosystem include tools for data ingestion, integration, and visualization; real-time data pipelines; cloud access; metadata management; and open-source code-sharing tools. People are essential drivers of such ecosystems by interacting and collaborating through investigator meetings, webinars, and other interpersonal strategies. For HEAL, these offerings have enabled highly productive synergies, joining researchers working in a range of diverse settings, such as neonatal intensive care units and emergency departments, jails and prisons, Indigenous communities, and basic research laboratories.
Identifying strategies to maximize the value of the HEAL data ecosystem
HEAL research will achieve its greatest scientific impact when researchers leverage the Data Ecosystem to analyze large-scale datasets across studies of chronic pain, opioid use disorder, and co-occurring conditions. Building and sustaining a functional data ecosystem presents numerous challenges beyond cultural barriers. These challenges span from intellectual property concerns to the complexities of data curation and secondary data analyses. Researchers must grapple with the need for a priori harmonization of data collection across distinct studies, as well as operational issues related to sharing, managing, and analyzing large-scale heterogeneous datasets. Ethical considerations regarding data use and community involvement add additional complexity. Pragmatic limitations tied to NIH grant funding cycles and budgeting for data management further complicate the landscape. Additionally, researchers often lack understanding and acceptance for NIH processes to obtain metadata, case report forms, and protocol-level data, which can hinder effective participation in open science initiatives.
There is no one-size-fits-all solution to support full participation in an open science ecosystem. Unsurprisingly, perhaps the most urgent task for the scientific community is to fully embrace the use of data standards,7–9 such as common data elements, data dictionaries, and data repositories compliant with FAIR (findability, accessibility, interoperability, and reusability) principles.10 Embracing data standards will broadly advance the value of data across the research spectrum, including basic research with model organisms, genomic and imaging analyses, wearables, social media monitoring, clinical studies, and multimodal research projects (Figure 1). Similarly, data harmonization efforts will advance a diverse array of projects ranging from imaging studies to analyzing registry data. Expected data use agreement norms are needed to break the logjam of access to public health data, corporate data, and electronic health record (EHR)/medical claims data.
NIH HEAL initiative data management challenges and opportunities. A comprehensive overview of data management challenges and strategies across 13 key data types in the HEAL Initiative ecosystem. The figure illustrates specific challenges for each data type (clinical, corporate, EHR, genomic, imaging, medical claims, public health, qualitative, registry, social media, software, wearables, and animal models) and corresponding strategies to address these challenges. For each data type, challenges related to standardization, privacy, integration, and technical implementation are matched with specific solutions focused on improving data sharing, harmonization, and utilization within the research community. Abbreviations: EHR = electronic health record; NIH HEAL = National Institutes of Health Helping to End Addiction Long-term® Initiative.
The era of data overload has evolved, with data sets now serving as crucial inputs for AI and advanced data science methods. Machine learning, natural language processing, and other algorithms show promise in analyzing diverse data sources, revealing clinical trends absent in traditional research trials and enabling continuous health monitoring for early interventions. HEAL investigators can advance the mission of open science by prioritizing transparency in response to the opioid and overdose crisis. Building community around large-scale data analysis is essential, involving creative activities like hackathons and question-focused sessions that encourage multimodal data collaboration. HEAL has developed resources for metadata registration and tools to guide investigators toward appropriate FAIR repositories, facilitating effective data sharing while adhering to usage agreements. Available tools also help investigators navigate appropriate FAIR repositories where data will most likely be found and used effectively and collaboratively. These repositories typically ensure data use agreement adherence.
This paper’s authors, as members of the HEAL Data Ecosystem Collective Board, contend that the benefits of an expansive and open data ecosystem outweigh the perceived challenges. By embracing open science principles, the research community can more effectively leverage diverse datasets to generate insights, inform policy, and ultimately improve outcomes for individuals affected by chronic pain and addiction. (Figure 1). Growing and sustaining a researcher driven HEAL Data Ecosystem will not only advance health equity but also provide other broad benefits for data contributors and users. Individuals living with these complex and often stigmatized conditions deserve to understand their effective health options based on what rigorous research shows. Policy makers need scientific proof to support policies aiming to help individuals and communities gain access to new treatments and prevention interventions.
A researcher-driven HEAL Data Ecosystem serves as a catalyst for scientific innovation directly impacting millions affected by chronic pain, addiction, and overdose (Table 1). Research institutions can support these efforts through resources for data standardization, recognition of data sharing contributions, and clear policies protecting scientific integrity and patient privacy. By embracing open science principles, the research community can more effectively leverage diverse datasets to generate insights, inform policy, and ultimately improve outcomes for individuals affected by chronic pain and addiction.
Accessing and contributing to the ecosystem.
|
|
Accessing and contributing to the ecosystem.
|
|
The HEAL Data Ecosystem represents a powerful tool for advancing our understanding and treatment of chronic pain and addiction. By fostering a culture of open science and data sharing, we can accelerate discovery, improve patient outcomes, and address the urgent public health crises of opioid addiction and chronic pain. As members of the scientific community, it is our responsibility to embrace these principles and actively contribute to building a more collaborative, transparent, and impactful research environment.
Acknowledgments
The authors appreciate the editorial contributions of Alison F. Davis, PhD. The authors are current and past members of the NIH HEAL Initiative Data Ecosystem Collective Board, which guides the overall strategy and direction of HEAL’s data management and sharing efforts. The authors also recognize the diligence and dedication of Jessica N. Mazerik, PhD, Anthony Juehne, and the HEAL Stewards.
Funding
Research reported in this publication was supported by National Institutes of Health HEAL Initiative and National Institute on Drug Abuse under grant numbers U24DA058606, U24DA057612, R24DA055306, R25DA061740 (Meredith Adams). Additional funding was provided by the National Institutes of Health through the following grants: U24AR076730 (Kevin Anstrom, Micah McCumber); PL1HD101059 (Carla Bann); UG3AR076387 and UH3AR076387 (Emine Bayman); R61MD018333 and R33MD018333 (Maria Chao); R61NS113329 and R33NS113329 (Georgene Hergenroeder); UM1DA049394 (Charlie Knott, John McCarthy); U54DA049110 (Martin Lindquist); R01DE029202 and R01DE029202-01S2 (Z. David Luo); U01DA050442 (Rosemarie Martin); U24NS135547 and OT2OD030541 (Maryann Martone); U24NS113844-05S2 (Sharon Meropol); U24DA050182 (Ty Ridenour); R24DA057611 (Lissette Saavedra); R01DA057599 (Abeed Sarker); and U24DA055330 (Wes Thompson).
Conflicts of interest: The authors include Dr Adams (Chair) and members of the NIH HEAL Data Ecosystem Collective Board. The views expressed in this manuscript represent those of the authors and do not necessarily reflect the official views of the National Institutes of Health or the NIH HEAL Initiative.
