Contributor – Melinda Rubinstein

Nurses have long carried the burden of disconnected health information. As healthcare enters the AI era, nurse designers have a critical role in shaping the data foundation for what comes next.

Nurses rarely have the luxury of caring for a patient with the whole story neatly assembled in front of them. Instead, we piece it together. We reconcile the medication list in the chart with the bottles a family member brought from home. We ask patients to remember what happened during a hospitalization at another facility, wait for records to arrive, search for test results, call pharmacies and caregivers, and try to determine whether a missing piece of information is irrelevant or the detail that changes everything. At the end of a shift, a question can linger: In an era of extraordinary technological capability, why are nurses and patients still doing the work our systems cannot?

The ability to access complete, accurate and timely health information affects everything from clinical decision-making and continuity of care to patient safety and equity. Evidence supports the clinical importance of interoperability as outlined by a systematic review by Li et al. (2022) finding that interoperable electronic health records were associated with improvements in medication safety, reductions in patient safety events and lower costs. Yet our health information infrastructure remains fragmented across hospitals, physician practices, pharmacies, labs, insurers and patient portals. When those systems cannot effectively communicate, the burden of connecting them shifts to patients.

That burden is not distributed equally. A patient who receives most of their care within one integrated health system, speaks English, has reliable internet access, uses a smartphone, and feels comfortable navigating digital portals has a very different ability to assemble a health history. But what if you’re an older adult seeing multiple specialists, a patient receiving care across disconnected systems, a person with limited digital access or literacy, or a child moving among caregivers and healthcare settings? Fragmented information isn’t only an interoperability problem, but it quickly becomes an equity problem. Shaw and Sekalala (2023) describe this through the framework of “health data justice,” drawing attention to how structural inequities shape who is represented in health data, who is excluded, and how those data are ultimately governed and used. Their framework shifts the conversation beyond technical questions of data quality or algorithmic bias to the institutional structures that shape how health data are generated, governed and used.

And now, at this pivotal juncture, artificial intelligence is about to depend on this fragmented data.

AI raises the stakes

The projected capability of AI to do more, see more, predict faster, has been at the forefront of the national healthcare conversation. That potential is exciting and matters very much to the future of healthcare. But while AI has extraordinary potential to help clinicians recognize patterns across years of medical history, synthesize complex information and make information more useful to patients and care teams we need to be examining the data from which these models will draw. Nurses are uniquely positioned to inform these models on what data are missing, and how to capture it.

AI does not eliminate healthcare’s information problem, but instead makes the integrity of our information infrastructure even more consequential. An algorithm can synthesize only the data it can access. Missing, inaccessible, inaccurate, or systematically underrepresented information does not suddenly become whole because a more powerful technology is analyzing it. If you remember the emergence of the internet, cloud computing, data lakes and visualization tools, each delivered powerful new capabilities and insights while revealing the limitations of the data beneath them. AI will inherit those same limitations unless healthcare uses this moment of transformation to address them.

The patient should be the center of the architecture

For decades, digital healthcare has largely been organized around institutions. A hospital maintains its record. A physician practice maintains its record. A pharmacy maintains its record. A laboratory maintains its results. Each organization has legitimate reasons for collecting and protecting information, but the cumulative result is striking in that a person’s health story is distributed among institutions while the patient is the only entity consistently moving among them. Gordon and Catalini (2021) describe this history as a patchwork of fragmented information systems that has created silos across healthcare organizations and care settings, limiting the informational continuity needed for connected care.

Healthcare has spent years working toward greater interoperability between those systems. The transition to AI-enabled healthcare is an opportunity to build on that progress by considering a fundamentally different information architecture, one organized around patients rather than the institutions that care for them.

This doesn’t necessarily translate to patients legally owning every piece of information generated during their care. Health information is more complicated than conventional property because, while it represents the patient’s health and experience, it is also generated through the work of clinicians and institutions, who have a responsibility to use what they learn to improve care. However, patients should have reliable access to the complete digital story of their health. A systematic review by Tapuria et al. (2021) found that access to electronic health records was associated with benefits including greater health awareness, medication adherence, improved patient-clinician relationships and, in some studies, improved health outcomes. Information should also move with patients across settings of care. Ethically, patients should also be able to understand how information about them is being used and have meaningful agency over uses that go beyond their direct care.

The principles are bigger than mere ownership and include access, integrity, interoperability, agency and trust. This framing aligns with the World Health Organization’s call to treat health data as a public good, supported by governance that strengthens individual and community control over health data while enabling their responsible use for broader health benefits (World Health Organization, 2021). These should not be optional features added to AI systems after they are developed. They should be pillars of the infrastructure upon which those systems are built.

Data infrastructure is nursing infrastructure

It may be tempting to leave these challenges to informaticists, technologists, health system executives and policymakers. Nurses cannot afford to do that.

Nurses already serve as human interoperability engines throughout healthcare. We reconcile conflicting information, identify gaps in histories and translate between systems and people, all of which serves to recognize when the data in front of us fails to represent the person in front of us. That work gives nursing a perspective that is essential to the design of AI-enabled healthcare.

Nurses also understand something a dataset cannot easily convey – health information exists in context. A medication marked “not taking” in the chart may represent cost, transportation, health literacy, side effects or a patient’s informed choice. A missed appointment may reflect unstable housing rather than disengagement. A sparse medical record may tell us less about someone’s health than about their access to healthcare. Absence of data is not necessarily absence of need.

That distinction becomes increasingly important when algorithms learn from the information healthcare systems collect. If some patients are less visible within our digital infrastructure because of where they receive care or how consistently their information follows them, AI will inherit those blind spots.

Nursing’s contribution to AI therefore cannot be limited to learning how to use new tools. Nurses should help determine what information those tools depend upon, whose experiences are represented, how patients participate in decisions about their information and what happens when the digital representation of a patient conflicts with clinical reality.

This role aligns with Cleofas’s (2026) Human Rights-Based Stakeholdership Framework, which positions nursing as an active stakeholder in healthcare AI governance rather than a passive observer of technological change. The framework calls for nursing engagement in framing AI-related challenges, assessing risks and benefits, evaluating governance approaches and managing risk, grounded in principles of human dignity, equity, participation, accountability and justice. Nursing’s perspective should therefore help shape not only how AI is used at the bedside, but the information infrastructure and governance decisions upon which AI-enabled healthcare is built.

The goal cannot simply be more intelligent healthcare systems. Our shared healthcare goal should be patient centric, emanating from our shared belief that every person deserves the best care that clinicians, institutions and technology can give them. Ultimately, this will allow nurses to spend less time piecing together the story and more time caring for the person it belongs to.

References

Cleofas, J. V. (2026). Towards a framework for human rights-based stakeholdership of nursing academia in the healthcare artificial intelligence governance ecosystem: A discussion paper. Nurse Education in Practice, 92, 104737. https://doi.org/10.1016/j.nepr.2026.104737

Gordon, W. J., & Catalini, C. (2021). Electronic health record interoperability: How did we get here and how do we move forward? JAMA Health Forum, 2(3), e210253. https://doi.org/10.1001/jamahealthforum.2021.0253

Li, E., Clarke, J., Ashrafian, H., Darzi, A., & Neves, A. L. (2022). The impact of electronic health record interoperability on safety and quality of care in high-income countries: Systematic review. Journal of Medical Internet Research, 24(9), e38144. https://doi.org/10.2196/38144

Shaw, J. A., & Sekalala, S. (2023). Health data justice: Building new norms for health data governance. npj Digital Medicine, 6, 30. https://doi.org/10.1038/s41746-023-00780-4

Tapuria, A., Porat, T., Kalra, D., Dsouza, G., Xiaohui, S., & Curcin, V. (2021). Impact of patient access to their electronic health record: Systematic review. Informatics for Health and Social Care, 46(2), 192–204. https://doi.org/10.1080/17538157.2021.1879810

World Health Organization. (2021). Health data as a global public good – a call for health data governance. World Health Organization.

About Melinda Rubinstein

Melinda Rubinstein is an MSN candidate at the Johns Hopkins School of Nursing and a member of its Policy Honors Program. Before entering nursing, she spent more than 20 years in technology strategy, digital transformation and AI implementation, most recently at Microsoft. Her work explores the intersection of nursing, technology and policy, with a particular interest in how nurses can help shape AI-enabled healthcare that protects patient agency, advances equity and keeps technology in service to humanity.

AI disclosure

The author used ChatGPT (OpenAI) during the development of this article to support brainstorming, literature discovery, organization, and editorial revision. The author independently reviewed and verified cited sources and retains responsibility for the ideas, analysis, accuracy, and final content of the article.