Unlocking US Health Data: Challenges, Solutions, and the Future

Unlike stock market data, which sees large investments in data systems to gather insights, health data, though similarly structured around people, lacks these foundational systems and investments.

The absence of a robust data infrastructure leads to severe consequences, including patients struggling to access their records, doctors facing difficulties in viewing patient histories, and researchers being unable to access crucial data for medical advancements.

Government struggles to build computer systems, often relying on committees of lawyers and bureaucrats who may not fully grasp the detailed, painstaking work involved in developing and maintaining modern data systems, leading to persistent underfunding and a crippled infrastructure despite decades of bipartisan support.

America's health data system is severely constrained by regulations, chronic underfunding, a lack of enabling legislation, and technical weaknesses, universally acknowledged by officials and experts.

The complexity of arcane regulations often dictates why straightforward technological improvements cannot be implemented, turning health data expertise into a mastery of these restrictive rules.

Overcoming these challenges necessitates enabling legislation across all 50 states, the creation and nurturing of technical organizations, new funding mechanisms, and addressing countless legal, structural, and bureaucratic hurdles.

Advocates propose a joint state-federal cooperation model that leverages existing state health information exchanges to provide data to federal agencies, a model currently demonstrating success.

The Department of Health and Human Services (HHS) is running a pilot program between the National Institutes of Health (NIH) and the state of Nebraska, which has successfully generated data for chronic disease studies, serving as a proof of concept for this state-federal hybrid approach.

The success of this model in Nebraska is attributed to the state's prior investments in data governance infrastructure; its broader expansion will require similar commitments from other states.

Working at the state level offers 50 opportunities to identify successful models for expanding America's health data infrastructure, potentially demonstrating a clear path forward even if only a handful of states achieve success.

Due to stringent security protocols, all attendees will be required to remain in the room until around 4:00 p.m., or until Secretary Kennedy has departed the building.

The MAHA Institute, though new to open health data, is drawing high-caliber participants to the summit, reflecting growing interest in addressing the broken health system, with the current speaker, Dr. Donald Rucker, Chief Strategy Officer of OneUp Health, being introduced next.

The 1942 Stabilization Act, a wartime law, made healthcare a pre-tax benefit, effectively shifting control from patients to third parties, who, regardless of their intentions, prioritize their own interests.

The 1965 establishment of Medicare further distorted the market by introducing administratively set prices, eliminating genuine market dynamics for healthcare services.

Cumulative errors in service mispayment have compounded over the years, influencing critical decisions such as medical students' specialty choices and hospitals' focus on high-margin services, often at the expense of patient needs.

Electronic medical records and delivery systems have self-organized to prevent interoperability, optimizing for high-margin procedures and creating systems that resist open data exchange.

Incumbent entities prevent healthcare data interoperability through various tactics, including extensive lobbying, maintaining secretive and convoluted fee structures for major electronic medical record (EMR) vendors, and imposing technical barriers such as ancient APIs and programming paradigms.

The Trusted Exchange Framework and Common Agreement (TEFCA) is criticized as an outdated set of protocols designed to entrench existing practices and protect incumbents, contrasting sharply with the widespread adoption of RESTful APIs in the consumer smartphone market, where Americans interact with thousands of such calls daily.

These practices, some of which resemble blatant antitrust violations, create an anti-competitive landscape where vendors actively block third-party integrations, further hindering the open exchange of health data.

Stricter enforcement of the Cures Act regarding information blocking is crucial, as the current lack of enforcement undermines the law's intent to outlaw such practices.

Achieving 'app parity' would allow patients to use any app of their choice to access their health data, rather than being confined to a specific provider's proprietary application, promoting greater patient control and data portability.

Integrating immunization registries with Health Information Exchanges (HIEs) is a low-hanging policy fruit that could resolve public controversies around vaccine safety by providing comprehensive and accurate data on vaccine effectiveness and comorbidities.

The current healthcare system is not primarily designed for the patient's benefit but rather for the system's own structural needs, leading to a fundamental innovation problem.

Despite advanced medical capabilities, such as reprogramming immune cells to fight cancer, the system still relies on outdated bureaucratic processes like clipboards and fax machines.

This system's failure to integrate basic patient information has severe consequences, exemplified by cases where critical time is lost in diagnosing rare diseases due to fragmented data across multiple providers and systems.

Patients are forced to act as the 'integration layer,' managing numerous patient portals and enduring long waits for appointments, highlighting the absurdity and patient burden created by a system that has digitized broken processes into digital silos.

The agency launched a challenge to accelerate progress within a year, engaging approximately 60 major tech companies, Electronic Health Record (EHR) vendors, provider groups, and patient advocates to prioritize patient-centric design.

Key goals include the "kill the clipboard" initiative, enabling patients to use their identity for one-time verification, and accessing their medical records through any chosen third-party app.

This approach aims to bypass the delays caused by waiting for industry laggards and instead empower those ready to move faster, ensuring privacy and security validation for patient-chosen apps through a new app library.

This progress allows patients to share their medical history with providers via QR codes, eliminating the need for manual paperwork.

The initiative involves efforts to expand data access, including integrating conversational AI for insights and launching a national provider directory to ensure information accuracy.

Other additions include price transparency tools, real-time benefits checking, modern scheduling, clinical trial matching, and the 'ditch the disc' initiative to eliminate physical CD transfers.

Although the CMS healthcare ecosystem is voluntary, financial incentives are being introduced to encourage market participation. The 'Access Model,' active since July, allows certified technology-enabled services to become Medicare Part B providers, enabling them to bill CMS and receive enrollment fees for patient participation. This model ties payments to outcomes, such as a reduction in hemoglobin A1C, and is designed as a 10-year model.

Separately, the program provides a rural health transformation fund of $50 billion to support technology adoption in underserved areas.

Dr. Nim Absar stated that 20% of health outcomes are determined by healthcare delivery systems, while 80% is influenced by factors such as air quality, nutrition, stress, genetics, and environment.

To move upstream and prevent disease, it is necessary to unify all health data, bringing together these clinical and social determinants.

The proposed data architecture must be person-owned and person-driven, applying intelligence to the combined data.

It is suggested that economic models across stakeholders, including providers, payers, retail, pharma, and food and beverage companies, should evolve to align mission and margin around the individual.

Dr. Raymond Ouie highlighted the current state of healthcare data as primarily syntactic interoperability, advocating for a shift to semantic interoperability.

He noted that medical records are often filled with free-text notes, making it difficult to extract concepts for patient care, and emphasized initiatives like USCDI.

Dr. Ouie also pointed out the disconnect between evolving clinical guidelines, such as those for colorectal cancer screening, and the ability of providers to stay updated or for systems to prove patient adherence.

He suggested that this issue often results in vital information not being accessible at the right time.

She explained that health data collection largely begins at the prenatal period, neglecting the months, years, and decades before conception that shape a child's health trajectory.

She advocates for a longitudinal data capture system that extends beyond biology and medical records to include the 'exposome'—the totality of environmental exposures.

Her goal is to capture a legacy archive of the exposome for parents before pregnancy to track environmental factors and potentially understand trajectories of childhood health.

Dr. Nim Absar recounted a personal experience with a patient, Debbie, who was repeatedly labeled 'non-compliant' by the healthcare system for uncontrolled diabetes.

It was revealed that Debbie's adherence challenges stemmed from social factors, including job loss, homelessness, and the inability to afford recommended foods, none of which were captured in her medical records.

This case illustrated a disconnect between acute care and holistic patient care that considers a patient's living circumstances.

The failure to integrate these 'missing pieces of data' resulted in repeated hospitalizations.

This data-driven approach enabled her daughter to present evidence to her doctors, facilitating a clinical breakthrough where AI analysis corrected a 16-year-old misdiagnosis of ulcerative colitis to microscopic lymphocytic colitis.

Initially, doctors were resistant to patient-generated data but became more receptive once they observed the tangible benefits and data-driven insights.

The speaker noted that the healthcare system is often a digitized version of a broken one, and true progress requires fixing fundamental issues and leveraging technology like AI to personalize care and empower patients to engage with their health data.

Carmen emphasized that innovation, particularly from early-stage private companies, is crucial for addressing fragmented health data by capturing multimodal information, including environmental, behavioral, lifestyle, sleep, wearable, personal, emotional, and psychological stress factors.

The goal is to leverage AI to create precision medicine and prevention opportunities, moving beyond reactive care to monitoring and modifying long-term health trajectories for individuals.

Integrating clinical dashboards into these technologies would allow physicians to continuously monitor patient data, bridging the gap between patient-collected information and professional review.

Empowering patients to own and manage their data through personal devices, apps, and APIs is seen as fundamental to systemic improvement, with the private sector playing a vital role in developing solutions for chronic conditions and overall health.

Raymond highlighted a critical workforce gap in the United States: a shortage of clinical and public health informaticists, professionals who bridge the divide between IT and clinical practice.

He noted that clinicians often face a 'swivel chair healthcare' scenario, having to navigate multiple unintegrated software systems, which impedes efficient data use and workflow.

A key issue is the reliance on coding systems like ICD-10, which are anchored on billing rather than comprehensive data representation, failing to capture holistic patient information such as social determinants of health.

This lack of semantic interoperability and standardized data representation prevents the effective merging and aggregation of different datasets, making it difficult for clinical decision support systems to function algorithmically.

Isaiah argued that the continued use of technologies like fax machines symbolizes a lack of trust within the healthcare system, encompassing distrust in laws, federal agencies like the ONC, and among providers regarding prescribed medications.

He highlighted a paradox where individuals trust technology in their personal lives but reject it for health information, leading to fragmented data and poor medication reconciliation.

For Isaiah, interoperability requires a single source of trust, enabling a team-based care approach where experts can validate information for patient health.

Amy added a humorous but telling anecdote: her daughter discovered that faxing a doctor yields a faster response than phone calls, patient portals, or even in-person visits, demonstrating how patients adapt to systemic inefficiencies.

This includes developing foundational tools such as provider directories and identity management systems.

Security and privacy certifications, however, should be delegated to the private sector, enabling innovation to flourish while the government sets the groundwork for secure data access.

Raymond emphasized that trust in health data is impeded by issues of data ownership and regulatory environments.

He advocated for formalizing data governance structures within organizations, suggesting increased cross-departmental communication between chief medical officers and chief information officers.

Achieving consensus on data sharing pathways is a slow process that requires ongoing dialogue.

Participants underscored the importance of integrating patients into these decision-making processes.

Amy highlighted a prevalent misconception regarding HIPAA, noting that it is frequently misused as an excuse to block data sharing.

She pointed out that existing patient portals often fall short, failing to provide comprehensive, longitudinal data.

Raymond stressed the necessity of standardized semantic representation for all health data to achieve genuine interoperability.

Amy encouraged audience members to utilize the Medicare app library on the CMS website to test their access to personal health data.

Patients should identify which healthcare providers are failing to share their information and report these instances to CMS.

This feedback from patients is used to address instances where health systems are not providing data to patients.

The discussion highlighted that individuals are not HIPAA-covered entities, meaning their personal health data falls under different regulatory protections.

The FTC regulates personal data, and large technology vendors have faced multi-billion dollar fines from the FTC for violating terms of service.

These examples demonstrate that regulatory protections for personal health data exist beyond HIPAA, particularly in the consumer-facing digital space.

Dan noted that the Medicaid agency pays about $20 billion a year to states, with a 90% federal match for system implementations, yet many systems remain custom and state-specific across the 50 states.

She noted that the value participants receive is dependent on their engagement level in the program, illustrating with her husband's experience.

The 'All of Us' program currently hosts the largest linked genomic and EHR dataset in the world, with over 800,000 participants.

The program is part of NIH's efforts to create an interoperable real-world data network that integrates various data sources.

Former ONC head Don Rucker described the agency's historic focus on feature-function regulation for EHRs as internally-looking.

He praised the Cures Act provisions on APIs for their approach to freeing up data, which marked a shift towards an API-driven performance model.

Bulk FHIR allows for population-level data access.

Despite progress, performance gaps persist among vendors, with some vendors' APIs functioning more slowly than others.

Don Rucker and Dan discussed how states often procure ineffective and non-interoperable Medicaid technology, frequently buying duplicative code already owned by the federal government.

A significant problem is that vendors are rarely penalized for failed implementations, leading to costly and unusable systems for states.

CMS is leveraging its substantial funding, which ranges from 50% to 90% of state Medicaid technology costs, to demand better procurement solutions and improve accountability.

Increased federal oversight is being implemented in response to a history of failed implementation cycles, aiming to move towards outcome-focused projects and clear standards.

States often lack the technological expertise to develop and implement complex systems, leading them to seek federal guidance for establishing clear standards and governance.

New governance structures are being implemented to ensure interoperability and define the functional requirements of health systems, aiming to prevent recurring failures and wasteful spending.

This shift includes a request for information (RFI) to gather market input on how to build reusable, interoperable systems from scratch, ensuring that future investments are tied to tangible results rather than just project completion.

Dan explained that current patient matching in healthcare relies on complex algorithms attempting to compare 'garbage data' across disparate systems to identify individuals.

He highlighted the advancement of secure digital identity verification through mobile driver's licenses, now available in 26 states, which demonstrate a more reliable standard than existing healthcare methods.

Digital identity verification, using device-based security like Face ID, offers a significantly more secure and efficient method for establishing identity compared to the legacy systems currently in use in healthcare.

Don Rucker emphasized that robust digital identity mechanisms are crucial for reducing friction in open data and patient-controlled data environments, especially as new identity tools rapidly evolve.

The speaker, Don, highlighted concerns that the TEFA global patient record system could create significant privacy risks due to its broad exposure of patient data.

True compliance with privacy regulations should focus on covered entities where patients, providers, and payers already have established relationships, anchoring data sharing within known and trusted networks.

Standardization efforts, such as through USCDI APIs, are crucial for reducing operational friction and ensuring secure, efficient data exchange without compromising privacy.

The NIH's 'Care for Health' program is expanding to embed clinical research hubs within primary care settings in rural communities across all 50 states, aiming to reach underserved populations with high disease burdens.

NIH is also leveraging crowdsourcing through 32 open challenge prize competitions, offering approximately $100 million in potential awards to incentivize innovation from a diverse range of stakeholders.

These initiatives seek to engage not only academic researchers, who traditionally receive 80-90% of NIH's budget, but also patients, patient groups, biotech companies, small pharma, and advocacy groups.

Recent challenge topics include digital twins, synthetic populations for chronic diseases, quantum computing for toxicological assessments, nutrition research for autoimmune diseases, and a real-world evidence innovation challenge from ENCATs.

This initiative aims to bring advanced diagnostic capabilities, often lacking in traditional primary care settings, directly to providers, bridging the technology gap for those without in-house MRI capabilities.

The broader goal is to facilitate early diagnosis and intervention, potentially leading to better treatment outcomes and enabling timely referrals to specialists.

This approach reflects a shift towards leveraging AI to optimize health outcomes at the grassroots level, particularly in primary care, where over 70% of Americans receive their healthcare.

Secretary Kennedy emphasized that medical records belong to patients and should be accessible via smartphones, allowing individuals to carry and share their data with doctors.

The proposal includes using 'super intelligence' (SI) to help patients understand complex medical reports and laboratory markers, enabling them to become "CEOs of their own health."

HHS, under President Trump, is focusing on investigating the root causes of chronic disease by auditing factors such as food, chemicals, and vaccines, aiming to transition the healthcare system from treating illness to prioritizing prevention.

The current system, where insurance companies, pharmaceutical companies, hospitals, and providers often profit more when people remain sick, creates incentives that need to be realigned towards preventative care and curing disease.

Secretary Kennedy revealed that federal agencies like NIH, CDC, and FDA previously hoarded data and even sold it to each other, creating significant barriers to information access and collaboration.

Upon his arrival at HHS, 32 distinct data systems were identified across these agencies, none of which could communicate with each other due to disparate operational frameworks.

HHS is now building new infrastructure to integrate these health systems and standardize data collection methodologies, aiming for a unified and accessible health data ecosystem.

This modernization includes efforts to overcome past instances where Congress-mandated data access, such as for the Vaccine Safety Database, was deliberately obstructed by disagregating data and imposing complex IRB approval requirements across 13 different HMOs.

Robert F. Kennedy Jr. cited Sam Altman's perspective, suggesting that doctors who fail to use SI for checking prescriptions, diagnoses, and assessments might soon be considered engaging in malpractice.

The agency is implementing SI to dramatically reduce the administrative burden on physicians, aiming to cut the time spent on paperwork from 50% to approximately 5%, thereby allowing more focus on patient care.

This initiative is part of President Trump's priority for American leadership in SI, with HHS now using SI more than any other federal agency to accelerate biomedical research and uncover the causes of chronic disease.

The agency is conducting comparative studies of vaccine adjuvants through the National Toxicity Program to evaluate vaccine risks and benefits.

HHS aims to increase transparency by utilizing U.S. and international data to improve vaccine safety monitoring and research.

This involves transforming fragmented 'vertical' medical records into a 'horizontal' format, enabling longitudinal studies to track exposures over long periods, particularly leveraging Medicaid data that covers millions of children.

Jamie Bland highlighted the difficulty of governance, particularly when privacy officers can act as barriers to data sharing.

Nebraska's success in data accessibility is attributed to investments in governance focused on opening up data while ensuring individual access.

Community organizations implementing health interventions outside traditional hospital systems require accessible data to evaluate the effectiveness of their programs.

A disparity exists between large, well-resourced health systems and small, rural providers who may lack access to legal resources, highlighting a need for education and support.

Achieving success involves creating a network where participating entities agree to foundational commitments to ensure connectivity and shared understanding.

The core challenge lies in building a structure where all stakeholders operate under shared expectations regarding data use and oversight.

Financial arguments must be redirected to prioritize preventative care to ensure long-term cost reduction.

Patients should be consulted, as many feel current systems protect institutions rather than empowering individuals with ownership of their health data.

At the MAHA Open Data Summit, it was noted that providers and patients should not see their success based on whether their health systems have access to all data.

Large health systems currently hold a competitive advantage in accessing and utilizing large data sets and analytics models.

Many providers work in small, rural, and frontier environments, where access to usable information is a barrier to success.

The objective is to ensure that a provider's ability to deliver exceptional care is the determinant of success, rather than their size or data resources.

AI is only as good as the data fed into it, according to Jonathan.

A risk involves rushing to 'point solutions' without first considering the underlying infrastructure.

AI offers a technological way to address healthcare's historically labor-intensive, runaway cost problem, referred to as 'Baumol's cost disease'.

There is a 2-3 year window to capitalize on momentum from the current administration and private sector to successfully implement AI.

Health Information Exchanges (HIEs), also known as health information networks or data utilities, serve as multi-stakeholder governance platforms.

They serve as a potential proving ground for AI in healthcare because they aggregate downstream workflow problems and data quality challenges.

Leveraging HIEs through consistent investment in governance and data quality could make it easier to test diverse AI applications, using real-time feedback loops to scale safe deployments.