TVTibor V. Varga

Tibor V. Varga

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Tibor V. Varga has a total of 1000 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Academic Publishing as a Complex System

    by Tibor Varga, Jeroen Uleman, Stéphane Tuffier, Morten Hulvej Rod, Mehdi Osooli

    Academic publishing is facing growing concerns related to reproducibility, peer review, and accessibility, yet proposed reforms have produced limited systemic change. We examined academic publishing as a complex adaptive system by conducting a thematic analysis of stakeholder interviews, developing a Causal Loop Diagram, and using the Action Scales Model to identify potential leverage points. From 13 stakeholder interviews, we synthesized nine interrelated challenges and mapped structural drivers and feedback loops shaping the system. Reinforcing dynamics linking prestige, evaluation metrics, and commercial incentives appear to sustain high publication volumes and constrain individual agency, even when dysfunction is widely recognized. Event-level actions were largely limited to transparency measures, whereas greater leverage emerged at the levels of system structures, goals, and beliefs. Applying this framework to the reproducibility, peer review, and accessibility crises illustrates how tensions are maintained. These findings highlight the value of complex systems approaches for understanding institutional behaviour and guiding coordinated change.

  • Top Cited Papers
    Open Access

    Navigating fairness aspects of clinical prediction models

    by Kaustubh Chakradeo, Inchuen Huynh, Sedrah B. Balaganeshan, Ole L. Dollerup, Hjørdis Gade-Jørgensen, Susanne K. Laupstad, Mikkel Malham, Tri-Long Nguyen, Adam Hulman, Tibor V. Varga

    Background: Algorithms are increasingly used in healthcare, yet most algorithms lack thorough evaluation and impact assessment across diverse populations. This absence of comprehensive scrutiny introduces a significant risk of inequitable clinical outcomes, particularly between different demographic and socioeconomic groups. Main body: Societal biases—rooted in structural inequalities and systemic discrimination—often shape the data used to develop these algorithms. When such biases become embedded into predictive models, algorithms frequently favor privileged populations, further deepening existing inequalities. Without proactive efforts to identify and mitigate these biases, algorithms risk disproportionately harming already marginalized groups, widening the gap between advantaged and disadvantaged patients. Various statistical metrics are available to assess algorithmic fairness, each addressing different dimensions of disparity in predictive performance across population groups. However, understanding and applying these fairness metrics in real-world healthcare settings remains limited. Transparency in both the development and communication of algorithms is essential to building a more equitable healthcare system. Openly addressing fairness concerns fosters trust and accountability, ensuring that fairness considerations become an integral part of algorithm design and implementation rather than an afterthought. Using a participatory approach involving three clinicians and three patients with lived experience of type 2 diabetes, we developed a set of guiding questions to help healthcare professionals assess algorithms critically, challenge existing practices, and stimulate discussions. Conclusions: We aim to direct healthcare professionals on navigating the complexities of bias in healthcare algorithms by encouraging critical thinking about biases present in society, data, algorithms, and healthcare systems.


  • Open Access

    User journeys in cross-European secondary use of health data: insights ahead of the European Health Data Space

    by Rachel B Forster, Eva Garcia Alvarez, Adrian G Zucco, Enrique Bernal-Delgado, Gayo Diallo, Francisco Estupiñán-Romero, Andrea Ganna, Madeleine Gorman-Asal, Christina Hilmarsen, Petr Holub, et al.

    The European Health Data Space (EHDS) regulation aims to facilitate cross-border sharing of health data across Europe. However, practical challenges related to data access, interoperability, quality, and interpretive competence remain, particularly when working with health systems across countries. This study aimed to evaluate and report the user journey of researchers accessing and utilizing health data across four European countries for secondary research purposes prior to implementation of EHDS. We conducted a narrative reflection of individual and collective experiences on key aspects of the user journey—discovery, access, use, and finalization. Data were gathered from various structured and unstructured sources, including an online log, prospective questionnaires, regular meetings, and interviews. Researchers faced challenges at different steps of the user journey, which included lack of data quality in national metadata catalogues (discovery stage). Differences in national regulations led to inconsistent timelines for gaining access to data (access stage), with approval times ranging from a few months to over a year. At the use stage, researchers experienced challenges in harmonizing health data due to variations in coding practices and data quality. Issues related to computational capacity caused further delays. Substantial challenges must be addressed for EHDS to succeed. Establishing knowledge hubs, fostering collaborations, and streamlining access processes are essential. Close collaboration with experts will likely be essential for an effective user journey. This analysis underscores the importance of collaboration, analytical reproducibility, and clear documentation to ensure the success and timely delivery of cross-border projects.


  • Open Access

    Algorithmic Fairness of QPrediction Cardiometabolic Risk Prediction Models

    by Inchuen Huynh, Denise Utochkin, Alexandros Katsiferis, Tri-Long Nguyen, Tibor V Varga

    Background: There is a lack of fairness evaluation of clinical prediction models used in routine practice. This study aimed to evaluate the fairness of three established risk prediction models belonging to the QPrediction family: QRISK3, QDiabetes, and QStroke, which estimate 10-year risks for cardiovascular disease, type 2 diabetes, and stroke, respectively. Methods and findings: We used data from the UK Biobank, a large-scale prospective cohort study comprising 502,131 participants. To assess calibration, the predicted risks were compared to the observed cumulative incidence rates, and survival Brier scores were calculated as a summary metric of model calibration. To assess discrimination, we calculated overall and group-specific true positive, true negative, false positive, and false negative rates (TPR, TNR, FPR, FNR). We compared the rates between subgroups of demographics and considered prediction parity when values were close to each other to a prespecified degree. Overall, all models showed systematic overprediction of risks for the UK Biobank validation cohorts, with the degree of overprediction varying between the models. Prediction parity was not reached in FNR and TNR in terms of ethnicity, education level, average household income before taxation, and Townsend deprivation score. For immigration status, there were no major differences in the discrimination metrics, and prediction parity was achieved. For QDiabetes, the trends were similar, apart from ethnicity, where the opposite was observed as for QRISK3. Conclusions: Inclusion of ethnicity and deprivation in the models may have contributed to more consistent calibration across subgroups. Variables were measured only at baseline, while many social and economic conditions can shift over time. Although including social determinants of health can improve predictive performance, the models are typically deployed in a clinical setting to guide individual-level decision making, which can inadvertently shift responsibility for health outcomes onto patients.

  • Open Access

    Racial disparities in continuous glucose monitoring-based 60-min glucose predictions among people with type 1 diabetes

    by Helene Bei Thomsen, Livie Yumeng Li, Anders A. Isaksen, Benjamin Lebiecka-Johansen, Charline Bour, Guy Fagherazzi, William P. T. M. van Doorn, Tibor V. Varga, Adam Hulman

    Non-Hispanic white (White) populations are overrepresented in medical studies. Potential healthcare disparities can happen when machine learning models, used in diabetes technologies, are trained on data from primarily White patients. We aimed to evaluate algorithmic fairness in glucose predictions. This study utilized continuous glucose monitoring (CGM) data from 101 White and 104 Black participants with type 1 diabetes collected by the JAEB Center for Health Research, US. Long short-term memory (LSTM) deep learning models were trained on 11 datasets of different proportions of White and Black participants and tailored to each individual using transfer learning to predict glucose 60 minutes ahead based on 60-minute windows. Root mean squared errors (RMSE) were calculated for each participant. Linear mixed-effect models were used to investigate the association between racial composition and RMSE while accounting for age, sex, and training data size. A median of 9 weeks (IQR: 7, 10) of CGM data was available per participant. The divergence in performance (RMSE slope by proportion) was not statistically significant for either group. However, the slope difference (from 0% White and 100% Black to 100% White and 0% Black) between groups was statistically significant (p=0.02), meaning the RMSE increased 0.04 [0.01, 0.08] mmol/L more for Black participants compared to White participants when the proportion of White participants increased from 0 to 100% in the training data. This difference was attenuated in the transfer learned models (RMSE: 0.02 [-0.01, 0.05] mmol/L, p=0.20). The racial composition of training data created a small statistically significant difference in the performance of the models, which was not present after using transfer learning. This demonstrates the importance of diversity in datasets and the potential value of transfer learning for developing more fair prediction models.


  • The European Health Data Space can be a boost for research beyond borders

    by Andrea Ganna, Angel Carracedo, Christian F. Christiansen, Emanuele Di Angelantonio, Pearl A. Dykstra, Angel M. Dzhambov, Roland Eils, Sara Green, Katharina L. Schneider, Tibor V. Varga, et al.


  • Open Access

    ­­­­Single-Dose Psilocybin Therapy for Alcohol Use Disorder: Pharmacokinetics, Feasibility, Safety, and Efficacy in an Open-Label Study

    by Mathias Ebbesen Jensen, Dea Siggaard Stenbæk, Catharina Dragsted Messell, Emil Deleuran Poulsen, Tibor V Varga, Patrick McDonald Fisher, Marie Katrine Klose Nielsen, Sys Stybe Johansen, Nora D Volkow, Gitte Moos Knudsen, et al.

    Background Psilocybin, a serotonin 2A receptor agonist with psychedelic properties, shows promise as a novel treatment for alcohol use disorder (AUD). While current studies involve two dosing sessions, the effects a single dose have not been investigated. Aims To investigate the pharmacokinetics, feasibility, safety, and efficacy of single-dose psilocybin therapy in AUD. Methods This open-label, single-group study investigated single-dose psilocybin therapy in ten treatment-seeking adults (eight men and two women; median age 44 years) with severe AUD. The treatment involved two preparation sessions, a high-dose psilocybin session (25 mg), and two integration sessions. Pharmacokinetics were determined by noncompartmental analysis, and changes in alcohol consumption, craving and self-efficacy, were assessed with a linear mixed model. Results Notable between-participant pharmacokinetic variations were observed, with peak plasma psilocin concentrations ranging from 14-59 µg/L. Alcohol consumption significantly decreased over the 12 weeks following psilocybin administration. Heavy drinking days were reduced by 37.5 percentage points (95% CI, -61.1, -13.9, p = 0.005), and drinks per day decreased by 3.4 units (95% CI: -6.5, -0.3), p = 0.035). This was corroborated by reports of rapid and sustained reductions in craving and increases in self-efficacy. Conclusions Despite pharmacokinetic variations, a single 25 mg psilocybin dose was safe and effective in reducing alcohol consumption in AUD patients. Larger randomised, placebo-controlled, single-dose AUD trials are warranted. Funding This work was supported by The Novo Nordisk Foundation (NNF19OC0058412), The Lundbeck Foundation (R-355-2020-945), The Health Foundation(21-B-0358) and The Ivan Nielsen Foundation. Clinical trial registration: NCT05347849


  • Tracked and self-reported nighttime smartphone use, general health, and healthcare utilization: results from the SmartSleep Study

    by Henning Johannes Drews, Christoffer Sejling, Thea Otte Andersen, Tibor V Varga, Andreas Kryger Jensen, Naja Hulvej Rod

    Study Objectives: Nighttime smartphone use is an increasing public health concern. We investigated whether nighttime smartphone use is associated with general health and primary healthcare utilization. Methods: Four thousand five hundred and twenty individuals (age 35.6 ± 9.7 years, 35% male) provided self-reported information on smartphone use frequency, symptoms of depression, and general health (one-item perceived health and cross-symptom composite score). A subset of the study sample (n = 3221) tracked their nighttime smartphone use. Primary healthcare utilization, i.e. the number of weeks in which at least one service from the patient’s general practitioner (GP) was billed in 2020, was extracted from Danish population registries. Statistical analysis comprised logistic and multiple linear regression, controlling for sociodemographics. Results: Three hundred and nineteen individuals (7%) reported using their smartphone almost every night or more. More frequent self-reported nighttime smartphone use was associated with poor general health across all measures. Using the smartphone almost every night or more was associated with 2.8 [95% CI: 1.9, 4.1] fold higher odds of reporting poor health and with an average of 1.4 [95% CI: 0.7, 2.1] additional GP utilizations per year compared to no use. Associations were also found for the cross-symptom composite score across all symptoms. Further adjustment for symptoms of depression attenuated some associations. Smartphone use towards the end of the sleep period (sleep-offset use) was associated with poorer self-reported general health, but not with healthcare utilization. Conclusions: Nighttime smartphone use frequency is associated with poor general health and healthcare utilization. Further studies should investigate the underlying causal structure and nighttime smartphone use as a transdiagnostic intervention target.


  • Open Access

    Acute and Long-Term Changes in Blood-Borne Biomarkers in Response to Dynamic Standing in Nonambulant Children With Cerebral Palsy

    by Tibor V. Varga, Åsa Andersson, Katarina Lauruschkus, Åsa B. Tornberg

    Purpose: To investigate acute and long-term changes in hormonal and inflammatory biomarkers in nonambulant children with cerebral palsy in response to dynamic standing exercise. Methods: Fourteen children with severe cerebral palsy were recruited. Anthropometrics and body composition measures were obtained. Physical activity levels before the study were assessed using hip-worn accelerometry. All children underwent a 30-minute dynamic standing exercise using the Innowalk standing aid. Respiratory data during exercise were collected using indirect calorimetry. Blood samples were collected before and after exercise. Blood samples were also obtained after two 16-week exercise protocols, in a resting state. Hormonal and inflammatory metabolites were measured from blood serum/plasma, and acute and long-term changes in biomarker levels were assessed using Wilcoxon signed-rank tests. Results: Of the 14 children at baseline, all had slightly/moderately/severely elevated C-reactive protein and cortisol levels. C-reactive protein levels were decreased following a 30-minute bout of dynamic standing (before exercise: 53 mg/L [interquartile range: 40–201]; after exercise: 39 mg/L [interquartile range: 20–107]; P = .04). Conclusions: We show that several hormonal and inflammatory biomarkers are dysregulated in children with cerebral palsy. Our preliminary results from a small, but deep-phenotyped prospective cohort indicate acute and long-term alterations of several biomarkers in response to exercise.