Nicoleta J. Economou-Zavlanos
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Nicoleta J. Economou-Zavlanos has a total of 141 co-authors. Here are the top co-authors by number of shared publications:
- Matthew Elmore
Duke University School of Medicine
6 shared publications - Armando Bedoya
Department of Biostatistics and Bioinformatics, Duke University School of Medicine, Durham, NC, United States
3 shared publications Suchi Saria
Johns Hopkins University
3 shared publications- Benjamin Alan Goldstein
Duke University School of Medicine
2 shared publications Christina Silcox
Duke-Margolis Institute for Health Policy , Washington, DC 20001
2 shared publications- Chuan Hong
Duke University
2 shared publications
Latest Publications
- Open Access
- 30 October 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Patient Priorities for Transparency in AI-Enabled Care: A Multi-Panel Study (Preprint)
BACKGROUND Transparency is widely recognized as an important principle for the use of AI in healthcare. However, patient-facing transparency requires decisions about what information should be communicated, and little is known about how patients prioritize different types of information when asked to distinguish what matters most. OBJECTIVE To examine how patients prioritize transparency elements derived from a health system AI governance process, and to identify which types of information are most important for patient-facing communication. METHODS Three patient panels participated in structured discussions and a constrained choice prioritization exercise focused on 17 AI transparency elements derived from Duke Health’s AI governance process. Panels included a demographically sampled advisory panel (n=17), and members of community groups serving predominantly African American (n=12) and Hispanic/Latino (n=7) communities. Participants selected the four transparency elements they considered most important for patient-facing communication. Frequencies and percentages were compared across panels, and qualitative rapid analysis was used to contextualize and interpret prioritization patterns. RESULTS Across the three panels, the most frequently selected transparency elements were risks and limitations (21/34, 62%), how AI outputs are used in clinical care (18/34, 53%), the purpose of the AI tool (14/34, 41%), and intended care population (12/34, 35%). Elements describing the technical and organizational context, including algorithm type, software application, implementation location, and supporting department, were selected less frequently. Qualitative findings indicated that participants valued information that helped them understand how AI may influence care decisions and whether tools were relevant to patients with clinical or demographic characteristics similar to their own. Several priorities were broadly shared across panels, although notable differences emerged. CONCLUSIONS When asked to prioritize AI transparency elements, participants favored information that helped them understand how AI could affect their care and whether tools were appropriate for patients like them. These findings suggest that patient-facing communication should emphasize these concerns while remaining responsive to community-specific priorities. More broadly, the study demonstrates how patient and community perspectives can help translate an existing AI governance process into priorities for patient-facing communication.
- 30 October 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessPatient Priorities for Transparency in AI-Enabled Care: A Multi-Panel Study (Preprint)
BACKGROUND Transparency is widely recognized as an important principle for the use of AI in healthcare. However, patient-facing transparency requires decisions about what information should be communicated, and little is known about how patients prioritize different types of information when asked to distinguish what matters most. OBJECTIVE To examine how patients prioritize transparency elements derived from a health system AI governance process, and to identify which types of information are most important for patient-facing communication. METHODS Three patient panels participated in structured discussions and a constrained choice prioritization exercise focused on 17 AI transparency elements derived from Duke Health’s AI governance process. Panels included a demographically sampled advisory panel (n=17), and members of community groups serving predominantly African American (n=12) and Hispanic/Latino (n=7) communities. Participants selected the four transparency elements they considered most important for patient-facing communication. Frequencies and percentages were compared across panels, and qualitative rapid analysis was used to contextualize and interpret prioritization patterns. RESULTS Across the three panels, the most frequently selected transparency elements were risks and limitations (21/34, 62%), how AI outputs are used in clinical care (18/34, 53%), the purpose of the AI tool (14/34, 41%), and intended care population (12/34, 35%). Elements describing the technical and organizational context, including algorithm type, software application, implementation location, and supporting department, were selected less frequently. Qualitative findings indicated that participants valued information that helped them understand how AI may influence care decisions and whether tools were relevant to patients with clinical or demographic characteristics similar to their own. Several priorities were broadly shared across panels, although notable differences emerged. CONCLUSIONS When asked to prioritize AI transparency elements, participants favored information that helped them understand how AI could affect their care and whether tools were appropriate for patients like them. These findings suggest that patient-facing communication should emphasize these concerns while remaining responsive to community-specific priorities. More broadly, the study demonstrates how patient and community perspectives can help translate an existing AI governance process into priorities for patient-facing communication.
- 21 July 2026
- Published by Taylor & Francis in The American Journal of Bioethics
- p. 1-5
Clear by Design: Consensus-Based Transparency Practices to Foster Trust in Health AI
- 21 July 2026
- Published by Taylor & Francis in The American Journal of Bioethics
- p. 1-5
Clear by Design: Consensus-Based Transparency Practices to Foster Trust in Health AI
- 21 July 2026
- Published by Taylor & Francis in The American Journal of Bioethics
- p. 1-6
A Safety-First Approach to Health AI
In keeping with the Hippocratic commitment to do no harm, the Institute of Medicine has defined patient safety as “the prevention of harm to patients” (Aspden et al. 2004). The health software indu...
- 21 July 2026
- Published by Taylor & Francis in The American Journal of Bioethics
- p. 1-6
A Safety-First Approach to Health AI
In keeping with the Hippocratic commitment to do no harm, the Institute of Medicine has defined patient safety as “the prevention of harm to patients” (Aspden et al. 2004). The health software indu...
- 10 June 2026
- Published by Taylor & Francis in The American Journal of Bioethics
- Vol. 26 (10) , 1-6
Fairness and Bias Management in Health AI: consensus-Based Recommendations for Best Practices Across the AI Lifecycle
In a landscape marked by uneven oversight, consensus-based guidance plays a crucial role in advancing AI fairness and managing bias. It contributes to a shared understanding across disciplines, ena...
- 10 June 2026
- Published by Taylor & Francis in The American Journal of Bioethics
- Vol. 26 (10) , 1-6
Fairness and Bias Management in Health AI: consensus-Based Recommendations for Best Practices Across the AI Lifecycle
In a landscape marked by uneven oversight, consensus-based guidance plays a crucial role in advancing AI fairness and managing bias. It contributes to a shared understanding across disciplines, ena...
- Top Cited PapersOpen Access
- 24 February 2026
- Published by Oxford University Press (OUP) in Journal of the American Medical Informatics Association
- Vol. 33 (5) , 1077-1080
Bridging the governance gap in health artificial intelligence: integrating nursing perspectives into the National Academy of Medicine Artificial Intelligence Code of Conduct
Background and Approach: In 2025, the National Academy of Medicine released an Artificial Intelligence Code of Conduct (AICC). In this commentary, we examine how the AICC introduces governance mechanisms to oversee AI applications and how it can support the ethical development and responsible use of AI in healthcare, paying special attention to the role of nurses. Findings: One shortcoming of the AICC is its lack of explicit acknowledgment of nurses, which risks obscuring their indispensable role in the safe, equitable, and effective use of AI in healthcare. We offer practical steps for health leaders to operationalize the AICC. Conclusion: Implementation of the AICC can support robust AI governance in health systems, but nurse expertise must be incorporated. The AICC offers a framework into which nursing perspectives can be embedded to ensure that AI tools positively transform healthcare delivery and enhance the quality and equity of care.
- 24 February 2026
- Published by Oxford University Press (OUP) in Journal of the American Medical Informatics Association
- Vol. 33 (5) , 1077-1080
Top Cited PapersOpen AccessBridging the governance gap in health artificial intelligence: integrating nursing perspectives into the National Academy of Medicine Artificial Intelligence Code of Conduct
Background and Approach: In 2025, the National Academy of Medicine released an Artificial Intelligence Code of Conduct (AICC). In this commentary, we examine how the AICC introduces governance mechanisms to oversee AI applications and how it can support the ethical development and responsible use of AI in healthcare, paying special attention to the role of nurses. Findings: One shortcoming of the AICC is its lack of explicit acknowledgment of nurses, which risks obscuring their indispensable role in the safe, equitable, and effective use of AI in healthcare. We offer practical steps for health leaders to operationalize the AICC. Conclusion: Implementation of the AICC can support robust AI governance in health systems, but nurse expertise must be incorporated. The AICC offers a framework into which nursing perspectives can be embedded to ensure that AI tools positively transform healthcare delivery and enhance the quality and equity of care.
- 12 February 2026
- Published by Taylor & Francis in The American Journal of Bioethics
- Vol. 26 (2) , 1-6
Bridging the Gap: Consensus-Based Considerations for AI Usefulness in Healthcare
- 12 February 2026
- Published by Taylor & Francis in The American Journal of Bioethics
- Vol. 26 (2) , 1-6
Bridging the Gap: Consensus-Based Considerations for AI Usefulness in Healthcare
- Open Access
- 15 October 2025
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society
- Vol. 8 (2) , 1611-1622
FairPOT: Balancing AUC Performance and Fairness with Proportional Optimal Transport
Fairness metrics utilizing the area under the receiver operator characteristic curve (AUC) have gained increasing attention in high-stakes domains such as healthcare, finance, and criminal justice. In these domains, fairness is often evaluated over risk scores rather than binary outcomes, and a common challenge is that enforcing strict fairness can significantly degrade AUC performance. To address this challenge, we propose Fair Proportional Optimal Transport (FairPOT), a novel, model-agnostic post-processing framework that strategically aligns risk score distributions across different groups using optimal transport, but does so selectively by transforming a controllable proportion, i.e., the top-λ quantile, of scores within the disadvantaged group. By varying λ, our method allows for a tunable trade-off between reducing AUC disparities and maintaining overall AUC performance. Furthermore, we extend FairPOT to the partial AUC setting, enabling fairness interventions to concentrate on the highest-risk regions. Extensive experiments on synthetic, public, and clinical datasets show that FairPOT consistently outperforms existing post-processing techniques in both global and partial AUC scenarios, often achieving improved fairness with slight AUC degradation or even positive gains in utility. The computational efficiency and practical adaptability of FairPOT make it a promising solution for real-world deployment.
- 15 October 2025
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society
- Vol. 8 (2) , 1611-1622
Open AccessFairPOT: Balancing AUC Performance and Fairness with Proportional Optimal Transport
Fairness metrics utilizing the area under the receiver operator characteristic curve (AUC) have gained increasing attention in high-stakes domains such as healthcare, finance, and criminal justice. In these domains, fairness is often evaluated over risk scores rather than binary outcomes, and a common challenge is that enforcing strict fairness can significantly degrade AUC performance. To address this challenge, we propose Fair Proportional Optimal Transport (FairPOT), a novel, model-agnostic post-processing framework that strategically aligns risk score distributions across different groups using optimal transport, but does so selectively by transforming a controllable proportion, i.e., the top-λ quantile, of scores within the disadvantaged group. By varying λ, our method allows for a tunable trade-off between reducing AUC disparities and maintaining overall AUC performance. Furthermore, we extend FairPOT to the partial AUC setting, enabling fairness interventions to concentrate on the highest-risk regions. Extensive experiments on synthetic, public, and clinical datasets show that FairPOT consistently outperforms existing post-processing techniques in both global and partial AUC scenarios, often achieving improved fairness with slight AUC degradation or even positive gains in utility. The computational efficiency and practical adaptability of FairPOT make it a promising solution for real-world deployment.
- Open Access
- 11 August 2025
- Published by Annual Reviews in Annual Review of Biomedical Data Science
- Vol. 8 (1) , 491-507
Algorithm-Based Clinical Decision Support: Evolving Regulatory Landscape and Best Practices for Local Oversight
The potential of algorithm-based clinical decision support (CDS) in healthcare continues to increase with the growing field of artificial intelligence (AI)-enabled CDS. The use of these technologies to support clinicians, patients, and health systems is still quite new, and to date, implementors and regulators are still identifying the best processes and practices to ensure the effective, safe, and equitable use of these technology solutions. To assist individuals and organizations interested in implementation of algorithm-based CDS and AI-enabled CDS in healthcare, this article reviews the important regulatory decisions that form the landscape within which algorithm-based CDS has emerged, modern governance frameworks used to oversee these CDS systems, nuances in evaluation and monitoring throughout the CDS life cycle, best practices for real-world implementation, safety and equity considerations, and avenues for future collaboration and innovation.
- 11 August 2025
- Published by Annual Reviews in Annual Review of Biomedical Data Science
- Vol. 8 (1) , 491-507
Open AccessAlgorithm-Based Clinical Decision Support: Evolving Regulatory Landscape and Best Practices for Local Oversight
The potential of algorithm-based clinical decision support (CDS) in healthcare continues to increase with the growing field of artificial intelligence (AI)-enabled CDS. The use of these technologies to support clinicians, patients, and health systems is still quite new, and to date, implementors and regulators are still identifying the best processes and practices to ensure the effective, safe, and equitable use of these technology solutions. To assist individuals and organizations interested in implementation of algorithm-based CDS and AI-enabled CDS in healthcare, this article reviews the important regulatory decisions that form the landscape within which algorithm-based CDS has emerged, modern governance frameworks used to oversee these CDS systems, nuances in evaluation and monitoring throughout the CDS life cycle, best practices for real-world implementation, safety and equity considerations, and avenues for future collaboration and innovation.
- 10 March 2025
- Published by Oxford University Press (OUP) in Journal of the American Medical Informatics Association
- Vol. 32 (4) , 626-637
Application of unified health large language model evaluation framework to In-Basket message replies: bridging qualitative and quantitative assessments
Objectives: Large language models (LLMs) are increasingly utilized in healthcare, transforming medical practice through advanced language processing capabilities. However, the evaluation of LLMs predominantly relies on human qualitative assessment, which is time-consuming, resource-intensive, and may be subject to variability and bias. There is a pressing need for quantitative metrics to enable scalable, objective, and efficient evaluation. Materials and Methods: We propose a unified evaluation framework that bridges qualitative and quantitative methods to assess LLM performance in healthcare settings. This framework maps evaluation aspects—such as linguistic quality, efficiency, content integrity, trustworthiness, and usefulness—to both qualitative assessments and quantitative metrics. We apply our approach to empirically evaluate the Epic In-Basket feature, which uses LLM to generate patient message replies. Results: The empirical evaluation demonstrates that while Artificial Intelligence (AI)-generated replies exhibit high fluency, clarity, and minimal toxicity, they face challenges with coherence and completeness. Clinicians’ manual decision to use AI-generated drafts correlates strongly with quantitative metrics, suggesting that quantitative metrics have the potential to reduce human effort in the evaluation process and make it more scalable. Discussion: Our study highlights the potential of a unified evaluation framework that integrates qualitative and quantitative methods, enabling scalable and systematic assessments of LLMs in healthcare. Automated metrics streamline evaluation and monitoring processes, but their effective use depends on alignment with human judgment, particularly for aspects requiring contextual interpretation. As LLM applications expand, refining evaluation strategies and fostering interdisciplinary collaboration will be critical to maintaining high standards of accuracy, ethics, and regulatory compliance. Conclusion: Our unified evaluation framework bridges the gap between qualitative human assessments and automated quantitative metrics, enhancing the reliability and scalability of LLM evaluations in healthcare. While automated quantitative evaluations are not ready to fully replace qualitative human evaluations, they can be used to enhance the process and, with relevant benchmarks derived from the unified framework proposed here, they can be applied to LLM monitoring and evaluation of updated versions of the original technology evaluated using qualitative human standards.
- 10 March 2025
- Published by Oxford University Press (OUP) in Journal of the American Medical Informatics Association
- Vol. 32 (4) , 626-637
Application of unified health large language model evaluation framework to In-Basket message replies: bridging qualitative and quantitative assessments
Objectives: Large language models (LLMs) are increasingly utilized in healthcare, transforming medical practice through advanced language processing capabilities. However, the evaluation of LLMs predominantly relies on human qualitative assessment, which is time-consuming, resource-intensive, and may be subject to variability and bias. There is a pressing need for quantitative metrics to enable scalable, objective, and efficient evaluation. Materials and Methods: We propose a unified evaluation framework that bridges qualitative and quantitative methods to assess LLM performance in healthcare settings. This framework maps evaluation aspects—such as linguistic quality, efficiency, content integrity, trustworthiness, and usefulness—to both qualitative assessments and quantitative metrics. We apply our approach to empirically evaluate the Epic In-Basket feature, which uses LLM to generate patient message replies. Results: The empirical evaluation demonstrates that while Artificial Intelligence (AI)-generated replies exhibit high fluency, clarity, and minimal toxicity, they face challenges with coherence and completeness. Clinicians’ manual decision to use AI-generated drafts correlates strongly with quantitative metrics, suggesting that quantitative metrics have the potential to reduce human effort in the evaluation process and make it more scalable. Discussion: Our study highlights the potential of a unified evaluation framework that integrates qualitative and quantitative methods, enabling scalable and systematic assessments of LLMs in healthcare. Automated metrics streamline evaluation and monitoring processes, but their effective use depends on alignment with human judgment, particularly for aspects requiring contextual interpretation. As LLM applications expand, refining evaluation strategies and fostering interdisciplinary collaboration will be critical to maintaining high standards of accuracy, ethics, and regulatory compliance. Conclusion: Our unified evaluation framework bridges the gap between qualitative human assessments and automated quantitative metrics, enhancing the reliability and scalability of LLM evaluations in healthcare. While automated quantitative evaluations are not ready to fully replace qualitative human evaluations, they can be used to enhance the process and, with relevant benchmarks derived from the unified framework proposed here, they can be applied to LLM monitoring and evaluation of updated versions of the original technology evaluated using qualitative human standards.
- 24 February 2025
- Published by Taylor & Francis in The American Journal of Bioethics
- Vol. 25 (3) , 121-123
A Heuristic for Notifying Patients About AI: From Institutional Declarations to Informed Consent
- 24 February 2025
- Published by Taylor & Francis in The American Journal of Bioethics
- Vol. 25 (3) , 121-123
A Heuristic for Notifying Patients About AI: From Institutional Declarations to Informed Consent