Introduction
The necessity of US public health data modernization was evident prior to the global COVID-19 pandemic.Reference Hagan 1 Fragmented infrastructure, poor interoperability, and delayed data reception and processing were issues already well-known to public health professionals, and some efforts to modernize were already underway. 2 However, the data challenges experienced during the COVID-19 pandemic exposed the consequences of neglected public health data systems and propelled data modernization into critical necessity, leading to massive advancements in a very short time. 3
In the US, public health data collection and analysis is largely decentralized. Consequently, modernizing public health data systems to enable national public health responses must include governance that addresses interjurisdictional data sharing. Data governance — the laws and policies for who uses and controls public health data, under what terms, and with what safeguards — is vital to navigate the legal, ethical, and relational concerns that different data-sharing partners might have.Reference Weber-Fares 4 Without effective and trusted data governance, friction between public health data partners could stall or impede data sharing and risks the breakdown of data modernization.
Historically, governance of interjurisdictional data sharing has relied on one-by-one negotiated Data Use Agreements (DUAs). 5 Each DUA is crafted between parties — often between the Centers for Disease Control and Prevention (CDC) and Epidemiology and Laboratory Capacity-funded jurisdictions — and sets terms for data access, use, and protection. 6 This governance model is flexible, but it is also resource intensive and time consuming. Additionally, the resulting variation in agreements limits transparency, complicates national coordination, and inhibits timely data sharing. Recognizing these barriers, the CDC aims to address these challenges by simplifying and standardizing DUAs to enable more efficient and effective public health data sharing and use. 7
However, DUAs, like all agreements, require some degree of trust between contracting parties. Put simply, data sharing “moves at the speed of trust.” 8 Without trust — in federal partners, in institutional neutrality, in protections against data misuse — governance efforts will collapse, threatening data modernization broadly. Currently, this trust is in a fragile state; a perceived lack of transparency, politicization of public health, and uncertain data protections erode faith in public health partners.
In this article, we examine the evolving landscape of public health data governance through jurisdictional perspectives, legal frameworks, and interjurisdictional negotiations. Building on these insights, we explore the critical role of mutual trust in data governance and examine recent developments that threaten to erode it.
US Public Health Data Governance: Challenges and Opportunities
Fragmentation in US Data Protection Laws
The patchwork landscape of state and local data-sharing and privacy laws presents a significant challenge to public health data governance. Existing laws often differ by data type, data subject, data use, and level of government, making the question “Can public health data be shared with federal public health authorities?” complex and multifaceted.Reference Schmit, Kelly and Bernstein 9
First, jurisdictions vary on the type of data covered by their laws. Specific categories of data, such as immunization data and syndromic surveillance data, may be covered by different laws with different requirements within jurisdictions, adding a level of complexity to data sharing. Additionally, not all jurisdictions have laws which expressly discuss data sharing. Of those that do, most jurisdictional data-sharing laws establish a voluntary standard, not a mandatory one. 10 This variation presents a challenge for the federal public health authorities to consistently receive data and creates potential friction points in negotiation.
Some jurisdictional restrictions can also affect the quality of data that can be shared. Some restrictions — including data subject consent, de-identification, and aggregation 11 — impact the utility of the data shared to researchers and public health officials. Restrictive laws could mean that jurisdictions are not legally allowed to share data including data elements that are granular enough for what is needed for effective public health activities and response.
Legal variation is a daunting problem for interagency and intergovernmental data sharing. Federal public health authorities must navigate numerous differing state, local, and territorial laws with various levels of restrictiveness to facilitate access to necessary national data.Reference Schmit 12
Transactional Friction in COVID-19 DUAs
In late fall of 2020, the CDC set out to establish immunization DUAs with state and territorial jurisdictions to govern how, when, and which data would be shared as the US positioned itself to fight the COVID-19 pandemic through novel vaccines. Publicly available COVID-19 immunization DUAs reveal foundational patterns and points of friction in interjurisdictional data sharing, informing future efforts even in an era of eroded trust and transparency.
Notably, there were two DUA templates circulated by the CDC: one dated October 21, 2020, and a later template dated November 5, 2020. Both served the overarching purpose of facilitating COVID-19 immunization data exchange between jurisdictions and the federal government, but the November 5 template exhibited greater specificity and restrictiveness than its predecessor. States with publicly available DUAs signed a version of one template or the other, with none drafting completely unique agreements. The emergence of the later template may indicate that the CDC revised its approach in response to jurisdictional concerns about the permissiveness of the original template terms. Lending support to this possibility is the fact that states that signed the October 21 template tended to have more permissive data sharing laws — or an absence of restrictive data sharing laws — when compared to states that signed the later template. 13
Importantly, the publicly available DUAs reveal that most jurisdictions made only minimal changes to the templates. However, identifying jurisdictional modifications offers insight into key points of friction in negotiations. Among the states that modified their templates, many restricted sharing of specific data elements or added protections for certain populations — like minors, immigrants, or members of small communities. Others restricted sharing to the minimum amount necessary or to specific, narrow purposes. A common but highly restrictive modification was a requirement for all data shared to be de-identified or aggregated, typically with citations to state laws. 14
These publicly available DUAs reveal patterns in interjurisdictional data sharing negotiations, with modifications reflecting jurisdictional legal interpretations, statutory constraints, and overall cautiousness regarding public health data sharing and protection. That said, the relative simplicity of the modifications — and the limited number of templates necessary to accommodate jurisdictions — may reflect a moment of opportunity born out of emergent circumstances that has since narrowed. It is certainly conceivable that the issues identified in these early DUAs have only intensified recently, propelled by an erosion of mutual trust and transparency.Reference Wamsley 15
It should be noted that the impact of the variation in COVID-19 immunization DUAs across jurisdictions has not been measured, so it remains untested whether more restrictive terms directly inhibited public health response on a local or national scale. However, there is robust evidence that public health activities are most effective when informed by data that is as timely and complete as possible, which is why legal restrictions such as de-identification and aggregation sometimes run counter to public health ethical guidelines.Reference Schmit 16
Jurisdictional Considerations on Data Sharing
As modernization efforts have continued, recurring jurisdictional concerns about data use, sharing, and governance have been documented. A common desire among jurisdictions is information on how their data will be used, shared, and protected to ensure compliance with state and local laws.Reference Mistry 17 In particular, jurisdictions have raised concerns about federal action without notifying states about the conditions under which data might be shared with other entities — such as other parts of the CDC or the Department of Health and Human Services (HHS), other federal agencies, or external researchers. 18 Jurisdictions also want more information about security standards, de-identification practices, and the potential for reidentification when data are combined with other sources. 19
In terms of process, jurisdictions desire consistent communication and standardized procedures for negotiating data sharing. 20 Because of the diversity of laws across jurisdictions, many states will find conflicts between their own laws and the terms of certain DUA templates. 21 Consistent points of contact could help ensure that the CDC understands each jurisdiction’s laws, their interpretations, and any necessary DUA modifications to maintain compliance. 22
These insights highlight themes of transparency, reciprocity, and shared authority — all tools of trust in intergovernmental data governance. If there is trust that partners will follow all legal and contractual requirements, data sharing projects are more likely to succeed. 23 Notably, these insights reflect a period of relative stability that has since shifted, potentially giving rise to new and more urgent jurisdictional concerns about sharing public health data.
Core DUA: The Proposed Governance Framework for Data Modernization
In order to help address these governance challenges, the CDC is developing the Core Data Use Agreement (DUA): a uniform DUA that could be utilized by public health jurisdictions across states, tribes, localities, and territories. 24 The Core DUA aims to promote clarity, consistency, and trust across jurisdictions by providing standard terms for data sharing and use while allowing for data-specific addenda (e.g., for data on immunizations, syndromic surveillance, case reporting, vital statistics) to accommodate legal diversity among jurisdictions.
The Core DUA framework is comprised of two components: Common Provisions and Addenda. The Common Provisions are applicable to all public health data and outline terms like confidentiality, security, and routine use. The goal of standardization is to reduce administrative burden, promote interjurisdictional consistency, and build clarity and trust among agencies. The Addenda help address variation in data-sharing laws by allowing agreements to be tailored to state-specific laws and regulations, for instance relating to data privacy or use limitations. 25 Ideally, this dual structure balances the efficiency of standardization with the reality of a fragmented legal landscape.
The Core DUA has been under development since at least December 2023, and while CDC officials have reiterated their commitment to advancing it, progress has slowed amid persistent governance challenges. Jurisdictional considerations about data authority, control, and misuse continue to shape its evolution. Underlying these challenges is the fundamental force that will determine the success of standardized data governance: trust.
Trust: A Fragile Foundation for Governance
While the Core DUA is a promising model for more consistent and transparent data governance, its success depends on more than the terms of the agreement itself. DUAs are only effective when parties are willing to cooperate, and this requires mutual trust. This is especially relevant to state and local jurisdictions, whose primary duty is owed to their constituents. Given the CDC’s limited authority over state and local public health data, a breakdown in negotiations has the potential to collapse national governance efforts. Thus, the erosion of trust may be the most pressing threat to long-term success of public health data modernization.
Some current data governance practices can undermine trust even in the absence of external political and legal pressure. The process of negotiating and executing DUAs remains a nonpublic one; even executed agreements are often unavailable to the public (i.e., executed syndromic surveillance DUAs between state and federal agencies are not publicly available). This means that jurisdictions have little visibility into the terms negotiated by others. This lack of transparency produces several trust-eroding consequences: jurisdictions may receive inconsistent information from the CDC, opportunities for collective learning are limited, and accountability is weakened in cases of misuse.
Moreover, the US Department of Government Efficiency (DOGE), the short-lived but hugely consequential US government initiative to reduce government waste, took many controversial actions related to sensitive data that could undermine trust in federal agencies. 26 Extensive cuts to federal public health funding and programs have threatened public health activities across state and local departments, leading to legal challenges and uncertainty about community health impact. 27 Additionally, in its term of operation, DOGE sought — and in some cases was granted access to — sensitive systems in agencies like HHS and the Centers for Medicare and Medicaid Servies that hold health data. 28 This will likely only reinforce concerns among stakeholders about the federal government’s ability to guarantee that public health data is handled legally and ethically.
Nonpublic health uses of public health data also threaten trust. Recent federal actions — such as attempts to transfer Medicaid data to ICE — signal a willingness by the federal government to utilize sensitive health data for law enforcement purposes. 29 In the wake of Dobbs, there is also concern that sensitive data may be used to prosecute individuals seeking controversial reproductive healthcare. This is not a baseless fear; there have already been attempts at the state level to access reproductive and genetic health data for law enforcement use. 30 These heightened concerns are likely to resonate with public health data custodians who are often zealously protective of data subjects.Reference Schmit 31
In addition to fears about inappropriate data management, there are increasing concerns about the independence and neutrality of the CDC itself. The politicization of public health has only increased since the height of the COVID-19 pandemic, with controversy expanding from COVID-specific responses to widely accepted, longstanding public health practices (e.g., the firing of CDC Director Susan Monarez, ostensibly for refusing to support unscientific vaccination policies). 32 Regardless of the factual basis, the mere perception that scientific objectivity is overridden by politics can crumble the credibility — and thus trustworthiness — of scientific agencies like CDC.
The federal government could simply ignore the “trust” problem altogether by requiring DUA signatures as a condition of federal funding, and recent developments suggest that this approach is increasingly plausible. While it may seem expedient, the consequences of such an action could be dramatic. Coercive DUAs would fundamentally undermine the collaborative governance model by creating a power imbalance and further eroding interjurisdictional trust. Moreover, even if jurisdictions are forced into compliance, they retain several avenues to protect their data interests. Jurisdictions could withhold discretionary data elements, many of which are vital to national coordination and public health decision-making, or even disincentivize health providers that voluntarily share their data with public health authorities (i.e., most states lack mandatory syndromic surveillance reporting laws).
Either response would weaken the integrity and completeness of public health data systems at all levels and limit public health response effectiveness. In addition, coercive measures could trigger more restrictive state and local data protection legislation that could limit what jurisdictions are authorized to share with federal public health partners, further fragmenting the legal landscape. Constitutional principles of federalism also limit the ability of the federal government to compel state participation in federal programs, meaning that coercive DUA negotiations are likely to provoke legal challenges and political resistance. 33 In short, whether through compliance, resistance, or litigation, a coercive approach will erode the foundations of public health data governance and could limit public health data abundance. Abandoning trust is not a viable path forward for modernization; it is a path towards the deterioration of the public health system itself.
Trust is not simply a theoretical concern; it is central to collaborative governance frameworks like the Core DUA. Jurisdictional considerations and tensions present in DUAs demonstrate that even well-intentioned and well-designed agreements rely on confidence that data will be handled legally, ethically, and transparently. However, that trust currently faces significant threats as fears grow regarding data security, non-public health use of sensitive data, and political interference in public health agencies. How these threats will truly impact interjurisdictional data sharing remains uncertain. What is clear, however, is that the stability of public health data modernization depends on building — or at least, not undermining — the trust that makes collaboration possible.
Disclosures
The work discussed in this publication was supported, in part, by the Centers for Disease Control and Prevention (CDC) of the US Department of Health and Human Services (HHS). The contents are those of the authors and do not necessarily represent the official views of, nor an endorsement by, CDC/HHS or the US Government.
Research reported in this publication was supported in part by the Gulf Research Program of the National Academies of Sciences, Engineering, and Medicine under award number SCON-10000856.
Open access
