Patient Pathway Mission
Introduction
Rényi AI, the Artificial Intelligence Research Group of the HUN-REN Alfréd Rényi Institute of Mathematics, is a research unit that values not only scientific excellence but also the direct societal utility of artificial intelligence.
We deliberately apply our deep fundamental-research expertise and broad AI competences to solve real societal challenges in practical applications. We see particularly strong potential in healthcare: here, unique opportunities exist to turn available data into tangible improvements in quality of life.
The group is committed to creating real value from Hungary’s healthcare data assets. Modern tools for prevention, prediction, and decision support can not only improve patients’ chances of recovery but also enhance the efficiency of the entire healthcare system. For us, this is not just research and development, but social responsibility and strategic innovation for Hungary’s future.
The starting point and the challenge
Hungary’s healthcare system possesses exceptionally rich data assets. NEAK (the National Health Insurance Fund) has been collecting care events for more than fifteen years—from outpatient and inpatient reports, through laboratory and imaging examinations, to prescription redemptions. Since 2017, EESZT (the National eHealth Infrastructure) has been recording healthcare documents: test results, discharge summaries, and outpatient reports. Together, these could provide a complete picture of Hungarian patients’ healthcare pathways; in reality, however, the picture is far more fragmented.
The structured data available at NEAK follow a financing logic and therefore lack many clinically important details. Documents recorded in the EESZT exist only in unstructured, free-text form, often using medical and Latin terminology, which means they cannot be analyzed on their own: structured data must first be extracted before they can be used to track and analyze patient pathways. This duality is why a high-quality patient pathway is not assembled either in individual care or system governance—and this is currently the greatest barrier to the meaningful utilization of healthcare data assets.
With advances in AI systems, tools are now available to make previously fragmented and unanalyzable healthcare documents structured, linkable, and usable. Thus, a mass of documents formerly unsuitable for analysis can be transformed into a genuine, analyzable patient-pathway database. The Patient Pathway Mission seizes this opportunity: its goal is to transform Hungary’s rich pathway-level data assets into a true resource for prevention, prediction, and decision support.
Essence of our Mission
Our mission is to organize patient data into coherent patient pathways and to put these data to practical use at the levels of individual patient care, clinical decision support, healthcare administration, and public health.
Utilization objectives
Our mission does not pursue general research aims; it creates direct impact. It offers tangible solutions to concrete problems that are usable in the short and medium term. The following table summarizes the key problems and pain points, the solutions we propose, and the outcomes the mission will deliver:
| Problem | Solution | Resulting outcome |
|---|---|---|
| Fragmented data sources available for research. | Structured, nationwide patient-pathway database. | Structuring of critical elements of patient documentation. |
| In public health programs, accurately defining and reaching target populations is challenging, limiting effectiveness. | Data-driven approach: prediction and prevention based on a patient-level database. | Algorithms implemented for screening programs (effective algorithms for population-level high-risk conditions such as diabetes and cancers). |
| Ensuring patient safety — Quality of care may vary by institution and practitioner because clinical guidelines are not followed consistently; this can jeopardize patient safety. Local and national professional governance and care organization currently have limited visibility into local healthcare processes (patient pathways). | Built-in process controls and retrospective data analysis to assess compliance. Patient-pathway synthesis. | Automated systems to monitor adherence to clinical guidelines and national comparative reports for selected processes. |
| Lack of data-driven control of medications; issues with adherence and drug interactions. | Data-driven medication analysis; adherence and interaction monitoring. | Data-driven medication-use and adherence monitoring system providing real-time alerts to physicians and patients. |
| Failure to recognize rare diseases in time—diagnoses often take years because patterns are not identified in the data early enough. | Machine-learning-based early detection systems that filter signals in care data in time. | Predictive module for early identification of rare diseases, supporting general practitioners and specialists. |
| Estimation of life expectancy and disease progression is performed using outdated tools not tailored to the Hungarian population. | Data-driven information tools and modern predictive models for life expectancy and disease progression. | Data-driven models for life expectancy and disease progression that provide a reliable basis for therapeutic decisions and can serve as foundations for care scheduling. |
| Despite existing data assets, there are no widely used predictive models for chronic diseases (e.g., diabetes, malignant tumors) or hospitalization risks. | Creation of a nationwide predictive framework integrating NEAK and EESZT data; development and validation of machine-learning models to forecast chronic diseases (e.g., diabetes, malignant tumors) and hospitalization risk. | National Predictive Module Package with three main components: Predictor module for chronic diseases (diabetes, malignant tumors). Hospitalization risk-assessment module. Delivery of predictions directly to physicians’ screens. |
| Physicians do not have time to read through 30–50 historical documents per patient. | Patient-pathway summaries and alerts for physicians. | AI-based patient-pathway summary. |
| National R&D&I capacities are fragmented and uncoordinated, while the EHDS (European Health Data Space) mandates the sharing and secondary use of healthcare documentation. Fragmentation greatly limits the effectiveness of R&D&I activities and erodes the competitive advantage offered by domestic data assets. | The mission brings stakeholders together in a complex, multidisciplinary R&D&I project focused on this critical and well-defined domain, generating valuable experience and creating a common platform for future research, innovation, and industrial collaborations that can build on the mission’s results. | A national, structured patient-pathway data platform that serves as a navigational element and foundation for domestic innovation activities and ensures that Hungary builds a strategic position in data-driven healthcare development. |
Impact summary
The mission’s results will simultaneously improve individual patient care, strengthen system governance, and open new horizons for research and public health. In other words, along this single, focused line of development, it delivers breakthroughs on five levels—for patients, physicians, research, decision-makers and administration, and public health:
Patients: precision screening, personalized diagnostic pathways, better follow-up.
Physicians: transparent patient pathways and decision-support alerts on screen, saving time.
Public health: predictive and preventive models, fewer avoidable hospital admissions, more targeted screenings (greater effectiveness and cost-efficiency).
Research: large-scale, unified database for clinical studies, national-level analyses by discipline, recognition of new patterns.
Administration/governance: modern monitoring and process-support tools (e.g., adherence tracking, built-in controls), more cost-effective processes.
The impact is measurable in the short term, and in the longer term this system can provide the foundation for nationwide prevention and prediction—i.e., the strategic capability that can place Hungary at the forefront of 21st-century healthcare innovation.
Roadmap
Successful implementation of the mission depends on securing the necessary support. If the program receives the required support, we will deliver according to the following schedule:
Year 1 – Foundation and initial working solutions — tangible pilots and proof of feasibility
Operation of the products and functionalities specified in the utilization objectives in pilot mode within clinical test environments, with narrower, well-chosen clinical focuses. (For example: a county hospital already runs a pilot version of the patient-pathway summary; in one region of the country, a data-driven screening invitation program operates in three cancer focus areas.)
Year 2 – First phase of nationwide scale-up — impact on the lives of millions
Nationwide extension of validated Year-1 results, and expansion of the R&D component to new clinical areas with new pilot programs.
Year 3 – System-level deployment and further phases of functionality expansion
Subsequent phases of functional development and nationwide rollout; embedding the full toolkit in the everyday operation of healthcare delivery.
The research group and institutional collaboration ecosystem
Rényi AI has for years been conducting research and development in this field and building the professional capacity required for large-scale, data-driven healthcare innovation.
A 12-member, multidisciplinary expert team is currently working on the mission’s preparation and development. Alongside mathematicians and AI developers, the team includes a physician, experts experienced in healthcare informatics systems (e.g., NEAK, EESZT), and researchers with industrial backgrounds who previously gained experience at international companies. This diversity of competences ensures that the mission is not only scientifically well-founded but also practically feasible.
Institutional collaboration
Through a long-standing collaboration with the Health Services Management Training Centre of Semmelweis University, access to financing data provides the basis for the healthcare research underpinning the mission.
The Rényi Institute also has a decades-long partnership with Eötvös Loránd University in mathematics. Within our mission, this primarily serves talent development and the preparation of future professionals: our aim is to involve young researchers and students at the intersection of healthcare and AI, who can later contribute to the success of joint research and innovation.
These institutions can bring important complementary capacities to the program, while Rényi is among the few domestic centers where the full competence set required for the mission has been comprehensively developed in recent years. The mission’s success will thus be ensured by a collaboration ecosystem integrating multiple institutions.
International dimension
The research group has established active relationships with international partners working on similar research directions. These collaborations enable the transfer of best practices, validation of models in international contexts, and direct alignment with European-level initiatives (e.g., EHDS). In this way, the Patient Pathway Mission can become not only nationally significant but also a project of international reference value.
Other strategic aspects
Secondary, strategic outcome: During the pilots, the most critical quality gaps in healthcare data will become visible. The mission can therefore also serve as a “data-quality compass,” indicating where and how structured data collection and processing should be developed in Hungary.
While comprehensive improvement of data quality is not within the narrow scope of the mission, insights generated in processing PDFs and analyzing patient pathways will contribute critically to long-term system development.
Secondary utilization: The databases and methodologies created during the mission will be used exclusively with domestic—non-foreign—partners. This ensures that the project’s results strengthen Hungary’s healthcare and innovation ecosystem and do not leak to foreign actors. healthcare-en.md Displaying healthcare-en.md.