Introduction and Aim:As AI tools are increasingly integrated into everyday life, it is imperative to understand how young adults perceive and use these systems to ensure safe and informed adoption. The current study explored attitudes, trust, usage patterns, and perceived risks related to AI tools for health information and lifestyle management among university students in Hungary, and identified the factors predicting their use of AI for health.Methods:Students at the University of Debrecen (N=148; 58.1% Hungarian, 41.9% international; mean age = 28.9 years) took part in an online cross-sectional study that measured digital skills, general AI attitudes, frequency of using AI, trust in AI for health, perceived AI-related health risk, privacy concerns, and basic demographics. Descriptive statistics, internal consistency (Cronbach's alpha), Pearson correlations, multiple linear regression for composite attitude scales, and ordinal logistic regression for Likert-type outcomes were among the analyses. A simple classifier and inductive thematic coding were used to analyze open-ended questions about what would boost confidence in health AI.Results:ChatGPT was the most popular tool. Students reported high digital skills (M=3.84/5) and frequent use of AI. 86.5% said they used AI to look up or confirm health information. Misinformation or inaccuracy was the most commonly expressed concern (116 responses). The strongest predictors of using AI for health purposes were more frequent searches for general health information and greater trust in AI for health. While lower trust in AI and greater privacy concerns were the main drivers of perceived AI-related health risk, positive attitudes toward health AI were linked to positive general attitudes toward AI and lower perceived AI-related health risk. The most cited requirements for greater trust in health AI were accuracy, dependability, and evidence-based information; followed by professional supervision, accountability, and clear data source identification.Conclusion:The results indicate that university students' acceptance of AI for health is primarily influenced by preexisting health information-seeking behavior and attitudinal factors, such as trust and perceived risk. Targeted AI-health literacy interventions that focus on source verification, transparency, privacy protection, and the complementary role of human professionals may facilitate the safe adoption of AI tools for health and lifestyle in young adults.

Témavezető: Bata Róbert

Introduction : Heart disease remains a leading cause of mortality in the United States, yet traditional clinical risk scores are seldom available in population surveys. Machine learning (ML) models offer the potential to improve population-level risk estimation using widely collected behavioral and sociodemographic variables.Objectives : To compare multiple ML algorithms for predicting heart disease risk using the 2022 CDC Behavioral Risk Factor Surveillance System (BRFSS) dataset and to position these results within the broader evidence base through a systematic review and meta-analysis of published ML heart disease prediction models.Methodology : The adults with valid heart disease data from the BRFSS 2022 were modeled using demographic, behavioral, and socioeconomic predictors. Eight ML algorithms ( namely Logistic Regression, Decision Tree, Random Forest, LDA, Naive Bayes, KNN, Gradient Boosting, and XGBoost) were trained with stratified 10-fold cross-validation and evaluated for AUC, accuracy, calibration, and SHAP explainability. A random-effects meta-analysis examined eight external ML studies found by a systematic search.Results : XGBoost has achieved the highest cross-validated AUC of 0.798, along with AUCs of 0.73-0.80 on the test set, and good calibration across the models. SHAP has pointed out age, diabetes, smoking, activity, and socioeconomic status, among others, as major predictors. A meta-analysis of 8 external studies yielded a pooled AUC of 0.83 (95% CI: 0.75–0.91), indicating strong ML performance worldwide.Conclusion : Machine learning enables reliable, interpretable prediction of heart disease risk using large-scale survey data. BRFSS-derived ML models show competitive accuracy relative to published evidence and may support population-level risk stratification and public health planning.

Témavezető: Oláh Gergő

Introduction: The emergence of artificial intelligence in healthcare has a significant impact on epidemiological forecasting. What is changing the government's capabilities and approach to health forecasting and planning. Although AI offers excellent opportunities for pandemic prediction and resource allocation, its use raises important ethical and legal questions. The key issues: data privacy, fairness and accountability, accuracy, and transparency. What determines people's trust in governments.Objective: This study examines the legal and ethical framework for using AI in public health forecasting. The goal is to identify risks and assess existing regulations in order to propose mechanisms that can help governments balance innovation with the protection of fundamental human rights. The speed of technological development these days often outpaces the capabilities of regulation and oversight, making this topic particularly relevant.Methodology: During the study, we conducted a complex, comprehensive analysis of international, regional, and national guidelines and legal documents: the WHO document "Ethics and Governance of AI in Healthcare", the EU AI Act, and the FDA guidelines on AI/ML medical devices. Secondly, to further explore these documents in detail, we decided to analyze four existing COVID-19 forecasting models: WHO Hub, Pandemic LLM, α-Satellite, and SIDARTHE. By using existing proven models created by top programmers as a basis, we were able to clarify the practical problems and ethical issues when using them.Results: After conducting our research, we concluded that the current regulatory framework is incomplete. The main risks identified relate to ethical and legal norms. Bias arising during data processing. Also lack of transparency in some AI algorithms and a currently unclear process for holding them accountable for errors. The lack of clear documentation, explainability of algorithms, and public oversight of models undermines trust in modern society. Conclusion: The introduction of AI into public health forecasting requires strict oversight and governance structures. Standards must be closely aligned with the protection of human rights. In this study, we propose: conducting mandatory preliminary impact assessments (DPIA/AIIA), establishing independent oversight bodies, and ensuring public participation in the monitoring of all processes. This principle can underpin the reliable use of AI and strengthen trust in public decision-making.

Témavezető: Dr. Bányai-Márton Gábor

Background:One of the most effective ways of preventing cervical cancer is the Human Papillomavirus vaccination. A considerable part of the human population doesn’t have an equal access to the vaccine, in spite of the fact that many countries have implemented appropriate policies, coverage still varies significantly between social groups. It is critical to understand how vaccination policies impact these disparities to ensure that public health initiatives are equitable and accessible to all. ObjectivesBy secondary data analysis and a comparative policy assessment across chosen countries, this study aims to find how national HPV policies affect health disparities. Methodology Five countries with distinct vaccination programs were chosen: USA, Australia, Brazil, India and Tanzania, to show policy approaches, coverage and equality outcome variations. Data was taken from publicly available sources, such as published literature, national health reports and WHO/UNICEF databases. Socioeconomic gaps and coverage trends were examined using tables and descriptive comparisons. This study offers unique comparisons and interpretations of the data retrieved.ResultsBased on preliminary trends, countries with universal programs such as Australia achieve greater coverage and fewer disparities, while countries with voluntary or unevenly implemented programs (e.g., Tanzania and India) show more disparities and lower coverage.ConclusionThe results demonstrate how vaccination policy design and health equality outcomes are interrelated. Global health measures to lessen HPV-related inequities can be developed and improved by the findings of the analysis.

Témavezető: Márton-Bányai Gábor

The Urban Heat Island (UHI) effect poses growing environmental and public health concerns in society. It is defined as elevated temperature in urban areas, especially in city centers, compared to surrounding rural regions. This thesis is based on a systematic review of UHI effect in large cities worldwide using studies published between 2000 and 2025. We explored the variations in climate, urban structure, and measurement approaches. The reports show an increase in global mean temperature values in arid and tropical regions, and consistently stronger effects at night. Key factors of increased UHI intensity include high built-up density, and low vegetation cover. These factors are especially characteristic to cities with large population size, and high population density. Green infrastructure and proximity to water bodies were identified as mitigating factors. This suggests the importance of increased public awareness. Public health implications are substantial as elevated temperature can cause increased heat-related morbidity and mortality, particularly among vulnerable populations. The study of these regional disparities in UHI may help to mitigate these health impacts. By summarizing global patterns and associated health risks our findings support the need for urban planning strategies that incorporate heat mitigation and promote population health resilience.

Témavezető: Prof. Dr. Béla Tóthmérész

Introduction: Prescription drug misuse affects millions globally, and genetic factors contribute substantially to vulnerability, yet these findings are rarely integrated into clinical decision-making. Physicians may prescribe potentially addictive medications without considering individual genetic susceptibility, limiting opportunities for prevention.Objective: This study aimed to synthesise genetic evidence across prescription drug classes, including opioids, stimulants and benzodiazepines, through meta-analysis. Secondary objectives included initiating the development of a polygenic risk score and outlining a prototype for a web-based tool to support clinical risk assessment.Methodology: A systematic search was conducted following PRISMA guidelines in PubMed, GWAS Catalog and Embase from 1990 to 2024. Studies with at least 500 participants that examined prescription drug misuse were included. Extracted data comprised effect sizes, standard errors and p-values for significant variants. Meta-analysis used inverse-variance weighting with random-effects models, and alleles were harmonised to ensure consistent effect directions. Heterogeneity was estimated using the I-squared statistic, and potential publication bias was explored using funnel plots and Egger’s test. A preliminary polygenic risk score was constructed using meta-analysed effect sizes with ancestry-specific considerations. Work on a web-based platform was initiated but remains under development.Results: Fourteen studies met inclusion criteria, representing more than 20 million participants across multiple ancestry groups. Meta-analysis identified several genome-wide significant variants, with notable signals in OPRM1, ADH1B, ALDH2, DRD2 and CHRNA5. The preliminary polygenic risk score demonstrated moderate predictive performance, with area under the curve values between 0.65 and 0.70 across drug classes.Conclusion: This project contributes to translating addiction genetics into potential clinical use. By combining meta-analysis with early-stage software development, we initiated the creation of a tool intended to support genetic risk assessment before prescribing potentially addictive medications. Further work is required to complete the platform, validate it.

Témavezető: Dr. Balazs Anita

Medical education aims to develop academic excellence and clinical skills, but it often unintentionally reinforces systemic inequities that disproportionately affect international and neurodiverse students. This paper explores the concept of a pathogenic curriculum, which describes how hidden aspects of medical training contribute to stress, moral injury, burnout, and mental health challenges. The culture of medical education still treats the demanding training process as a rite of passage, where anxiety, sleep deprivation, and self-sacrifice are seen as necessary for becoming a doctor. Despite evidence showing that physicians face higher risks of addiction and self-harm, the idea of delayed gratification remains deeply rooted. Students learn that showing emotions is a weakness, patients are problems to solve, and self-care is a luxury they cannot afford. High-stakes exams, competitive grading, and overwhelming workloads add to the pressure, pushing students to seek opportunities abroad. The international nature of medical education of the 21st-century adds another layer of complexity. While Western models dominate, local cultures, especially in the Global South, shape how students experience and manage stress. Cultural differences in family expectations, the prestige of becoming a doctor, and attitudes toward seeking help create uneven experiences. Access to mental health support also varies widely between institutions. Academic self-efficacy, that is, a student's own belief in their capability to succeed in a professional environment, is adjacent to the above-mentioned dilemmas of international medical education. Constant evaluations, fear of failure, and imposter syndrome can create a cycle where low confidence worsens mental health, leading to poor performance and even dropping out, harming not only students but also the institutions they belong to. This study examines the pathogenic curriculum through qualitative and quantitative data, including student surveys, academic interactions, assessment practices, institutional policies, and physician narratives. We argue that the current system disproportionately blames students rather than addressing structural flaws. Our proposed solution emphasizes an ethics of care, sustainable pedagogical practices, and systemic reforms to prioritize student well-being, equity, and inclusion, advancing the broader goal of decolonizing medical education for all, particularly international and neurodiverse learners.

Témavezető: Dr. Ureczky Eszter

Background: Disparities in population health outcomes, including life expectancy and healthy lifeyears, exist between Hungary and Canada, with Canada demonstrating stronger performance. Thisstudy examines whether structural and organizational differences in each country’s public healthsystem contribute to these gaps.Methods: A qualitative document analysis was conducted using key resources from both countries.Four primary areas were reviewed: (1) national public health strategies, (2) competencies expectedin postgraduate public health physician training, (3) requirements for leadership positions in publichealth, and (4) systems for auditing healthcare providers.Results: In Canada, public health uses a coordinated federal-provincial model. The government establishesstandards and supports provinces with funding and guidance, while provincial authorities develop andimplement their own aligned strategies thus ensuring consistency and preparedness nationwide.In Hungary, national planning is not fully embedded at the governmental level. Instead, public healthorganizations (such as the National Public Health Center and the Hungarian Association of PublicHealth Training and Research Institutes) are responsible, hence affecting coordination.(2) Both countries require high-level qualifications and standardized specialist training for publichealth physicians. However, Canada places stronger emphasis on public health advocacy and strategiccommunication, representing a meaningful difference in professional expectations.(3) Although both systems value advanced degrees and specialist qualifications, Canada consistentlymandates relevant public health credentials for leadership positions. In Hungary, significantleadership roles can be held without formal public health education, suggesting weakerprofessionalization of management within the system.(4) Canada applies multi-level accountability, which contribute to performance evaluation, resultingin robust external auditing of public health institutions.Hungary’s regulations assign responsibility for external audits to public health organizations, whichare often inconsistent or insufficiently implemented. As a result, comparative assessment ofhealthcare provider performance is limited.Conclusion: Hungary and Canada share some structural similarities in professional training pathways;However, there are notable differences. These partially explain the public health performance and population health indicators disparities

Témavezető: Dr. Sandor Janos

Introduction: Approximately 9–10 million people die from cancer each year. The lungs are the second most common site of metastases, underscoring the relevance and potential impact of AI-driven approaches in this area. In present times, the use of AI has become extremely widespread, extending from laypeople to healthcare professionals, including clinicians. We highlight the applications of AI in oncology, with a specific focus on lung metastases.Patients and Methods: Our experiment aimed to evaluate the responses of different AI engines to prompts regarding treatment modalities. We analyzed management options for several fictional patients presenting with solitary, oligo-, and multiple lung metastases, using prompts of varying complexities pertaining to different educational levels, such as a layperson, a medical student, a medical doctor, and an expert oncologist. We then compared the findings after noting the evidence levels of the responses. Results: The results were mostly similar, containing controversial data as well, and there were minor discrepancies based on the quality of the prompt. We found misleading answers, too, that only an expert can recognize based on clinical routine. The data was correct, but the context was inadequate. Conclusion: The regulation of AI has not kept pace with its development, posing global concern. Although AI is increasingly mainstream, it cannot replace medical experts due to regulatory issues and challenges such as AI hallucinations. Experimenting various prompting techniques offers an excellent opportunity to develop a sound understanding of the appropriate use of AI systems.

Témavezető: Dr. Furka Andrea

Introduction: With the rapid expansion of robot-assisted surgery, patient demand for information is rising. However, the technology remains surrounded by misconceptions, such as the belief that the robot operates autonomously. Restricted access to operating rooms limits patient understanding of surgical workflows and may exacerbate pre-operative anxiety. Virtual reality (VR) and 3D visualisation offer a solution by providing "safe insight" into the operating theatre. Objective: To assess patients’ knowledge, fears, and demand for visual information regarding robotic surgery, and to develop and pilot a portable VR-based education prototype.Methods: We conducted a cross-sectional study involving 50 patients at the Department of Surgery, University of Debrecen Clinical Center. An anonymous survey assessed awareness, specific misconceptions (e.g., autonomy level), perceived fear (Likert scale 1–5), and openness to VR-based education. Based on these findings, we developed a "pop-up" education system using spatial 3D video recordings and standalone VR (Meta Quest) headsets to demonstrate robotic arm mechanics and operating room logistics. Results: While 58% of respondents had heard of the "Da Vinci system", self-rated understanding was low (mean 2.7/5), with 40% reporting minimal understanding. Critically, 46% of patients held the misconception that the robot performs surgery autonomously. The mean fear level was 2.8/5, with 20% reporting high anxiety (score 4–5). Regarding solutions, 70% indicated that seeing the surgical view "with their own eyes" would help alleviate fear, and 80% expressed willingness to use a VR headset. Most notably, 62% stated that a 3D demonstration would make them "much more confident" in their decision to undergo surgery.Conclusion: This pilot identifies a clear gap between patient awareness and understanding of robotic surgery, characterised by the prevalence of myths regarding autonomous operation. The results demonstrate high demand for visual information and strong acceptance of VR technology, even among older demographics. The developed portable VR solution is a feasible tool for dispelling misconceptions and improving patient decision-making confidence.

Témavezető: Zsolt Varga

A lillafüredi klímaterápiás helyek turisztikai és gyógyászati auditjához készült kérdőív célja, hogy átfogó képet adjon a látogatók demográfiai hátteréről, egészségi állapotáról, a barlangterápiával kapcsolatos ismereteikről, eddigi tapasztalataikról és jövőbeni érdeklődési szándékaikról. A felmérés alapvetése, hogy a Szent István-barlang – különösen a Fekete-terem – speciális mikroklímája olyan környezetet biztosít, amely légzőszervi betegségek kezelésére és stresszcsökkentésre is alkalmas, ezért fontos megérteni, hogyan látják ezt a potenciális és jelenlegi látogatók.A demográfiai blokk – életkor, nem, lakóhely típusa, iskolázottság és jövedelmi helyzet – lehetővé teszi a célcsoportok meghatározását, valamint annak vizsgálatát, mely rétegek érdeklődnek leginkább a barlangterápia iránt. Az egészségügyi rész a krónikus légzőszervi betegségekre, allergiákra, asztmára és más releváns problémákra vonatkozik, továbbá felméri a betegség szubjektív súlyosságát. Ezek az adatok fontos információval szolgálnak arról, hogy mely egészségügyi csoportok profitálhatnak legjobban a barlang mikroklímájából.A kérdőív külön szekciót szentel annak, hogy a kitöltők mennyire ismerik a barlangterápiát, korábban részt vettek-e ilyen programon, és milyen hatást tapasztaltak. A válaszok rávilágítanak arra, hogy a kezelések milyen mértékben hatékonyak megítélésük szerint, mennyi ideig tart a pozitív hatás, illetve milyen csatornákon keresztül értesülnek a szolgáltatásokról.A jövőbeni érdeklődés vizsgálata kulcseleme az auditnak. A kérdések azt térképezik fel, hogy a válaszadók mennyire nyitottak klímaterápiás kezelésekre, milyen tényezők (ár, időtartam, kísérők, szolgáltatások) befolyásolják döntéseiket, és mennyi pénzt lennének hajlandók rászánni egy terápiás alkalomra vagy akár többnapos programra. Emellett információt gyűjt arról is, ismerik-e a lillafüredi barlangot, milyen benyomásaik vannak róla, illetve szívesen kapnának-e további tájékoztatást.A kérdőívből származó adatok elengedhetetlenek a turisztikai és gyógyászati szolgáltatások fejlesztéséhez. Segítségükkel pontosabban meghatározhatók a szükséges infrastrukturális, programbeli vagy kommunikációs fejlesztési irányok, továbbá támogatják egy hosszú távú, fenntartható és látogatóközpontú stratégia kialakítását Lillafüred klímaterápiás kínálatának bővítéséhez.

Témavezető: Pataki Jánosné

A gazdasági és társadalmi tényezők jelentős hatást gyakorolnak a környezetvédelmi szabályozások kialakítására, azok szigorúságára, valamint a környezetszennyezés és az egészségi állapot alakulására. A kutatás célja annak feltárása volt, hogy a gazdasági változók, például az infláció és a társadalmi fejlettséget jelző mutatók, mint a humán fejlettségi index (HDI) miként befolyásolják a környezetvédelmi előírások szigorúságát (EPS). Emellett vizsgálta azt is, hogy e szabályozások milyen mértékben hatnak a károsanyag-kibocsátásra, illetve az egészségügyi indikátorokra, például a légúti betegségekhez köthető halálozásokra.A kutatás 1990 és 2019 közötti időszakra vonatkozó adatokat használt fel több adatbázisból (EDGAR, European Health for All Database, Our World in Data, World Bank). A statisztikai értékelés Spearman-féle rangkorrelációval és vektor autoregressziós modellezéssel történt.A Spearman-féle rangkorreláció eredményei szerint a HDI és az EPS között erős, pozitív kapcsolat figyelhető meg (ρ=0,829; P<0,001), ami arra utal, hogy a társadalmi-gazdasági fejlettség növekedésével párhuzamosan szigorúbb környezetvédelmi szabályozások jelennek meg. Ugyanakkor a szén-dioxid-kibocsátás (ρ=-0,791; P<0,001), az infláció (ρ=-0,783; P<0,001), a légúti betegségekből eredő halálozás (ρ=-0,472; P=0,009) és a korai halálozás miatt elvesztett életévek száma (ρ=-0,823; P<0,001) mind negatív korrelációt mutattak a szabályozások szigorúságával.A vizsgálat arra enged következtetni, hogy magasabb HDI-értékek elősegítik a környezetvédelmi előírások szigorítását, míg a gazdasági instabilitás, például az infláció emelkedése, hátráltathatja az ilyen jellegű intézkedések bevezetését. Továbbá a negatív összefüggések azt jelzik, hogy a szigorúbb környezetvédelmi szabályozások hozzájárulhatnak a károsanyag-kibocsátás mérsékléséhez és a kedvezőtlen egészségügyi mutatók javulásához, ezáltal a fenntarthatóság és a lakosság egészségi állapotának erősítéséhez.

Témavezető: Dr. Szőllősi Gergő József

Egy tömegszerencsétlenség, legyen az természeti katasztrófa, közlekedési baleset, vagy terrorcselekmény, szinte bármelyik ország állampolgárait érintheti. Az áldozatok azonosításának nehéz, gyakran körülményes és összetett feladata ezekben a rendkívüli esetekben meghaladhatja az érintett ország katasztrófaelhárítási lehetőségeit. A DVI a modern kriminalisztika kevésbé ismert különálló szakterülete, ami 1980 óta az Interpol önálló részlegét képezi. Legfontosabb feladata a hatékony és sikeres áldozatazonosítás nemzetközi együttműködés keretében, egy esetlegesen bekövetkezett tömegszerencsétlenség vagy katasztrófa esetén. A DVI szervezetének egységes irányelvek mentén történő működését az Interpol egy nemzetközileg elismert kézikönyv és egységes formanyomtatványok biztosításával teszi lehetővé. Ezen túlmenően fontos szerepet vállal a rendszeres szakmai továbbképzések biztosításában is.Egy tömegszerencsétlenség áldozatainak azonosítását számos körülmény nehezítheti. Gyakran van szükség az egységesített protokollok alkalmazásán túl, speciális technikák és műszerek használatára, valamint további szakértők bevonására. Az azonosítási folyamat több állomásra bontható le. Az első a katasztrófa helyszínén végzett munka, ami magába foglalja a holttestek, testrészek, és maradványok, valamint az ezekhez tartozó személyes tárgyak összegyűjtését, és szállítását. Ezt követően bonctermi vizsgálat történik, ahol dokumentálásra kerülnek az elsődleges illetve másodlagos azonosító jegyek. A holttestek adatainak a lehető legalaposabb részletességgel történő rögzítése után, a tömegszerencsétlenségben valószínűleg érintett személyek eltűnése is bejelentésre kerül. Ezek után megindul a személyes adatok és az elsődleges azonosító jegyek összehasonlító mintáinak begyűjtése az eltűnt személyek családtagjaitól, barátaitól. Végül a bonctermi vizsgálat során nyert, és a hozzátartozóktól származó adatok összehasonlításra kerülnek, egyezés esetén pedig a személyazonosság kimondásra kerül. Az azonosítási protokollok, az egységes dokumentáció és jegyzőkönyv használata teszi lehetővé a gyors, hatékony munkavégzést. Továbbá biztosítja a közös szakmai nyelven történő kommunikációt a különböző nemzetiségű és anyanyelvű szakemberek között.Tömegszerencsétlenség bárhol, bármikor történhet, és ez alól egy állam sem kivétel. Éppen ezért rendkívül fontos, hogy minden fejlett ország rendelkezzen egységes, nemzetközi elvárásoknak megfelelő áldozatazonosító eljárással.

Témavezető: Dr. Rácz Kálmán