Nan Kong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Nan Kong has a total of 142 co-authors. Here are the top co-authors by number of shared publications:
Carolina Vivas-Valencia
Weldon School of Biomedical Engineering, Purdue University, West Lafayette, Indiana, USA
5 shared publicationsZhouyang Lou
School of Industrial Engineering, Purdue University, West Lafayette, IN, USA
5 shared publicationsAditya Sai
Carbon Health
3 shared publications- Bin Du
Nanjing University of Aeronautics and Astronautics
3 shared publications - Dengfeng Sun
School of Aeronautics and Astronautics, Purdue University, West Lafayette, IN, USA
3 shared publications - Mingyang Li
Department of Industrial and Management Systems Engineering, College of Engineering, University of South Florida, 4202 E. Fowler Avenue, Tampa, Florida 33620
3 shared publications
Latest Publications
- Open Access
- 4 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 974-990
Beyond Surveys: Predicting Student Well-being from Open-ended Video Responses using LLMs
- 4 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 974-990
Open AccessBeyond Surveys: Predicting Student Well-being from Open-ended Video Responses using LLMs
- Open Access
- 2 June 2026
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 17, 1843390
Spatial topology and competitive access differentially shape early T cell priming in the lymph node: an agent-based modeling approach
Introduction: Adaptive immune activation in lymph nodes requires rare antigen-specific naïve T cells to locate antigen-bearing dendritic cells within a spatially structured stromal network. Reduced priming efficiency is typically attributed to weak T cell receptor signaling, yet it remains unclear whether failure arises from impaired signaling or from limited access to antigen-bearing dendritic cells during early scanning. Methods: We developed COORDINATE, a spatially explicit agent-based model of lymph node microanatomy that integrates fibroblastic reticular cell topology, chemokine-guided migration, and competition for dendritic cell access. Results: We show that stromal architecture and trafficking biases strongly influence which T cells encounter antigen, while competition for limited dendritic cell access can exclude a substantial fraction of cells from forming any productive contact. Consequently, reduced activation can arise from failed clonal recruitment rather than diminished signaling following contact. Discussion: These results support a view of early T cell priming as an access-limited process and indicate that commonly used endpoint measurements can conflate failure to access antigen with failure to activate. Together, our findings suggest that improving early antigen access, rather than strengthening signaling alone, may represent an alternative strategy to enhance adaptive immune responses.
- 2 June 2026
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 17, 1843390
Open AccessSpatial topology and competitive access differentially shape early T cell priming in the lymph node: an agent-based modeling approach
Introduction: Adaptive immune activation in lymph nodes requires rare antigen-specific naïve T cells to locate antigen-bearing dendritic cells within a spatially structured stromal network. Reduced priming efficiency is typically attributed to weak T cell receptor signaling, yet it remains unclear whether failure arises from impaired signaling or from limited access to antigen-bearing dendritic cells during early scanning. Methods: We developed COORDINATE, a spatially explicit agent-based model of lymph node microanatomy that integrates fibroblastic reticular cell topology, chemokine-guided migration, and competition for dendritic cell access. Results: We show that stromal architecture and trafficking biases strongly influence which T cells encounter antigen, while competition for limited dendritic cell access can exclude a substantial fraction of cells from forming any productive contact. Consequently, reduced activation can arise from failed clonal recruitment rather than diminished signaling following contact. Discussion: These results support a view of early T cell priming as an access-limited process and indicate that commonly used endpoint measurements can conflate failure to access antigen with failure to activate. Together, our findings suggest that improving early antigen access, rather than strengthening signaling alone, may represent an alternative strategy to enhance adaptive immune responses.
- Open Access
- 24 March 2026
- Published by Springer Nature in Health Care Management Science
- Vol. 29 (2)
How much is adequate staffing for a nursing home? Forecasting daily service need of its case mix
Staffing adequately in an economical manner is vital to nursing homes (NHs) in the United States. NHs strive for providing resident-centered differentiated service to their changing and diverse residents and service cases. In this paper, we present a novel Bayesian forecasting method to predict acuity category-specific resident volume and caregiver type-specific staff time on a daily basis for NHs. We utilize the Minimum Data Set (MDS) alongside the Resource Utilization Group (RUG) guidelines according to residents’ acuity and the Patient Driven Payment Model (PDPM) specifications on recommended staff time in response to their staffing need, to generate two time series on daily NH service need, i.e., resident volume of each acuity group and staff time of each caregiver type in the entire facility, respectively. Given that the two multi-dimensional time-series above are nonstationary with potential correlations between dimensions, we propose prediction models with time-varying latent states to capture the dynamic patterns in the data and incorporate shared information between different categories. Specifically, we introduce a generalized mixture model to predict the discrete-valued resident volume and a generalized linear model to predict the continuous-valued staff time. Further, we propose a unified Bayesian estimation framework that allows convenient sequential updating and embed it in a forecasting procedure with rolling window to better capturing the nonstationarity inherent in the data. We demonstrate the superiority of the proposed model-based Bayesian forecasting method by comparing its performance against benchmark methods, using data from representative NHs.
- 24 March 2026
- Published by Springer Nature in Health Care Management Science
- Vol. 29 (2)
Open AccessHow much is adequate staffing for a nursing home? Forecasting daily service need of its case mix
Staffing adequately in an economical manner is vital to nursing homes (NHs) in the United States. NHs strive for providing resident-centered differentiated service to their changing and diverse residents and service cases. In this paper, we present a novel Bayesian forecasting method to predict acuity category-specific resident volume and caregiver type-specific staff time on a daily basis for NHs. We utilize the Minimum Data Set (MDS) alongside the Resource Utilization Group (RUG) guidelines according to residents’ acuity and the Patient Driven Payment Model (PDPM) specifications on recommended staff time in response to their staffing need, to generate two time series on daily NH service need, i.e., resident volume of each acuity group and staff time of each caregiver type in the entire facility, respectively. Given that the two multi-dimensional time-series above are nonstationary with potential correlations between dimensions, we propose prediction models with time-varying latent states to capture the dynamic patterns in the data and incorporate shared information between different categories. Specifically, we introduce a generalized mixture model to predict the discrete-valued resident volume and a generalized linear model to predict the continuous-valued staff time. Further, we propose a unified Bayesian estimation framework that allows convenient sequential updating and embed it in a forecasting procedure with rolling window to better capturing the nonstationarity inherent in the data. We demonstrate the superiority of the proposed model-based Bayesian forecasting method by comparing its performance against benchmark methods, using data from representative NHs.
- Open Access
- 19 December 2025
- Published by Springer Nature in BMC Public Health
- Vol. 26 (1)
Towards an understanding of community-engaged efforts in addressing the youth substance use in the Democratic Republic of the Congo
Youth substance use is a growing concern in the northeastern Democratic Republic of the Congo (DRC). This study aims to enhance understanding of community training programs that prevent youth substance use by designing a comprehensive process map that illustrates the roles and interactions of key entities and individuals, including local churches, community workers, and nongovernmental organizations (NGOs). We also conducted surveys and interviews to further explore community-led programs. Five training centers have been built to help young adults gain professional skills and prevent them using substances. Our findings identified the four key groups in this process: youths, leadership team, teachers, and youth leaders. This study emphasizes the critical role of community engagement in addressing youth substance use and highlights the importance of further research opportunities for using technology to enhance training centers’ operational efficiency and impact. By addressing identified challenges, these training centers can become more effective in empowering youth and preventing substance use in the region.
- 19 December 2025
- Published by Springer Nature in BMC Public Health
- Vol. 26 (1)
Open AccessTowards an understanding of community-engaged efforts in addressing the youth substance use in the Democratic Republic of the Congo
Youth substance use is a growing concern in the northeastern Democratic Republic of the Congo (DRC). This study aims to enhance understanding of community training programs that prevent youth substance use by designing a comprehensive process map that illustrates the roles and interactions of key entities and individuals, including local churches, community workers, and nongovernmental organizations (NGOs). We also conducted surveys and interviews to further explore community-led programs. Five training centers have been built to help young adults gain professional skills and prevent them using substances. Our findings identified the four key groups in this process: youths, leadership team, teachers, and youth leaders. This study emphasizes the critical role of community engagement in addressing youth substance use and highlights the importance of further research opportunities for using technology to enhance training centers’ operational efficiency and impact. By addressing identified challenges, these training centers can become more effective in empowering youth and preventing substance use in the region.
- Open Access
Real-Time Assessment of Bystander Situation Awareness in Drone-Assisted First Aid
Rapid naloxone delivery via drones offers a promising solution for responding to opioid overdose emergencies (OOEs), by extending lifesaving interventions to medically untrained bystanders before emergency medical services (EMS) arrive. Recognizing the critical role of bystander situational awareness (SA) in human-autonomy teaming (HAT), we address a key research gap in real-time SA assessment by introducing the Drone-Assisted Naloxone Delivery Simulation Dataset (DANDSD). This pioneering dataset captures HAT during simulated OOEs, where college students without medical training act as bystanders tasked with administering intranasal naloxone to a mock overdose victim. Leveraging this dataset, we propose a video-based real-time SA assessment framework that utilizes graph embeddings and transformer models to assess bystander SA in real time. Our approach integrates visual perception and comprehension cues--such as geometric, kinematic, and interaction graph features--and achieves high-performance SA prediction. It also demonstrates strong temporal segmentation accuracy, outperforming the FINCH baseline by 9% in Mean over Frames (MoF) and 5% in Intersection over Union (IoU). This work supports the development of adaptive drone systems capable of guiding bystanders effectively, ultimately improving emergency response outcomes and saving lives.
Open AccessReal-Time Assessment of Bystander Situation Awareness in Drone-Assisted First Aid
Rapid naloxone delivery via drones offers a promising solution for responding to opioid overdose emergencies (OOEs), by extending lifesaving interventions to medically untrained bystanders before emergency medical services (EMS) arrive. Recognizing the critical role of bystander situational awareness (SA) in human-autonomy teaming (HAT), we address a key research gap in real-time SA assessment by introducing the Drone-Assisted Naloxone Delivery Simulation Dataset (DANDSD). This pioneering dataset captures HAT during simulated OOEs, where college students without medical training act as bystanders tasked with administering intranasal naloxone to a mock overdose victim. Leveraging this dataset, we propose a video-based real-time SA assessment framework that utilizes graph embeddings and transformer models to assess bystander SA in real time. Our approach integrates visual perception and comprehension cues--such as geometric, kinematic, and interaction graph features--and achieves high-performance SA prediction. It also demonstrates strong temporal segmentation accuracy, outperforming the FINCH baseline by 9% in Mean over Frames (MoF) and 5% in Intersection over Union (IoU). This work supports the development of adaptive drone systems capable of guiding bystanders effectively, ultimately improving emergency response outcomes and saving lives.
- Open Access
- 30 May 2024
- Published by Frontiers Media SA in Frontiers in Physiology
- Vol. 15, 1399374
Seizure prediction in stroke survivors who experienced an infection at skilled nursing facilities—a machine learning approach
Background: Infections and seizures are some of the most common complications in stroke survivors. Infections are the most common risk factor for seizures and stroke survivors that experience an infection are at greater risk of experiencing seizures. A predictive model to determine which stroke survivors are at the greatest risk for a seizure after an infection can be used to help providers focus on prevention of seizures in higher risk residents that experience an infection.Methods: A predictive model was generated from a retrospective study of the Long-Term Care Minimum Data Set (MDS) 3.0 (2014–2018, n = 262,301). Techniques included three data balancing methods (SMOTE for up sampling, ENN for down sampling, and SMOTEENN for up and down sampling) and three feature selection methods (LASSO, Recursive Feature Elimination, and Principal Component Analysis). One balancing and one feature selection technique was applied, and the resulting dataset was then trained on four machine learning models (Logistic Regression, Random Forest, XGBoost, and Neural Network). Model performance was evaluated with AUC and accuracy, and interpretation used SHapley Additive exPlanations.Results: Using data balancing methods improved the prediction performances of the machine learning models, but feature selection did not remove any features and did not affect performance. With all models having a high accuracy (76.5%–99.9%), interpretation on all four models yielded the most holistic view. SHAP values indicated that therapy (speech, physical, occupational, and respiratory), independence (activities of daily living for walking, mobility, eating, dressing, and toilet use), and mood (severity score, anti-anxiety medications, antidepressants, and antipsychotics) features contributed the most. Meaning, stroke survivors who received fewer therapy hours, were less independent, had a worse overall mood were at a greater risk of having a seizure after an infection.Conclusion: The development of a tool to predict seizure following an infection in stroke survivors can be interpreted by providers to guide treatment and prevent complications long term. This promotes individualized treatment plans that can increase the quality of resident care.
- 30 May 2024
- Published by Frontiers Media SA in Frontiers in Physiology
- Vol. 15, 1399374
Open AccessSeizure prediction in stroke survivors who experienced an infection at skilled nursing facilities—a machine learning approach
Background: Infections and seizures are some of the most common complications in stroke survivors. Infections are the most common risk factor for seizures and stroke survivors that experience an infection are at greater risk of experiencing seizures. A predictive model to determine which stroke survivors are at the greatest risk for a seizure after an infection can be used to help providers focus on prevention of seizures in higher risk residents that experience an infection.Methods: A predictive model was generated from a retrospective study of the Long-Term Care Minimum Data Set (MDS) 3.0 (2014–2018, n = 262,301). Techniques included three data balancing methods (SMOTE for up sampling, ENN for down sampling, and SMOTEENN for up and down sampling) and three feature selection methods (LASSO, Recursive Feature Elimination, and Principal Component Analysis). One balancing and one feature selection technique was applied, and the resulting dataset was then trained on four machine learning models (Logistic Regression, Random Forest, XGBoost, and Neural Network). Model performance was evaluated with AUC and accuracy, and interpretation used SHapley Additive exPlanations.Results: Using data balancing methods improved the prediction performances of the machine learning models, but feature selection did not remove any features and did not affect performance. With all models having a high accuracy (76.5%–99.9%), interpretation on all four models yielded the most holistic view. SHAP values indicated that therapy (speech, physical, occupational, and respiratory), independence (activities of daily living for walking, mobility, eating, dressing, and toilet use), and mood (severity score, anti-anxiety medications, antidepressants, and antipsychotics) features contributed the most. Meaning, stroke survivors who received fewer therapy hours, were less independent, had a worse overall mood were at a greater risk of having a seizure after an infection.Conclusion: The development of a tool to predict seizure following an infection in stroke survivors can be interpreted by providers to guide treatment and prevent complications long term. This promotes individualized treatment plans that can increase the quality of resident care.
- 26 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 11 (4) , 3611-3622
Adapting Progressive Hedging for Solving Two-Stage Stochastic Programs Under a Peer-to-Peer Computing Network
Following recent technology advances and increasing applications in peer-to-peer network computing, we explore the potential of using decentralized optimization methods to solve notoriously hard stochastic programs. As the first attempt, we adapt the well-known progressive hedging (PH) method under a peer-to-peer computing network for solving two-stage stochastic programs efficiently. Similar to the existing parallel PH method, our decentralized variant assigns each node within the network to take charge of the computing tasks covering one or few scenarios, and thus it distributes the overall computational burden over the entire network. However, unlike the parallel PH method, the decentralized variant no longer needs a master node to realize central coordination, and thus it improves the scalability of the network computing as well. In this paper, we show the exact convergence of our decentralized method for solving two-stage stochastic programs with continuous variables subject to convex constraints. Further, we investigate several computational issues for the mixed-integer cases to improve the adaptation efficiency. Finally, the efficiency of our method is demonstrated through comparative computational experiments on a set of benchmark test instances.
- 26 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 11 (4) , 3611-3622
Adapting Progressive Hedging for Solving Two-Stage Stochastic Programs Under a Peer-to-Peer Computing Network
Following recent technology advances and increasing applications in peer-to-peer network computing, we explore the potential of using decentralized optimization methods to solve notoriously hard stochastic programs. As the first attempt, we adapt the well-known progressive hedging (PH) method under a peer-to-peer computing network for solving two-stage stochastic programs efficiently. Similar to the existing parallel PH method, our decentralized variant assigns each node within the network to take charge of the computing tasks covering one or few scenarios, and thus it distributes the overall computational burden over the entire network. However, unlike the parallel PH method, the decentralized variant no longer needs a master node to realize central coordination, and thus it improves the scalability of the network computing as well. In this paper, we show the exact convergence of our decentralized method for solving two-stage stochastic programs with continuous variables subject to convex constraints. Further, we investigate several computational issues for the mixed-integer cases to improve the adaptation efficiency. Finally, the efficiency of our method is demonstrated through comparative computational experiments on a set of benchmark test instances.
- 1 March 2024
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Manufacturing & Service Operations Management
- Vol. 26 (2) , 447-464
Shortening Emergency Medical Response Time with Joint Operations of Uncrewed Aerial Vehicles with Ambulances
Problem definition: Uncrewed aerial vehicles (UAVs) are transforming emergency service logistics applications across sectors, offering easy deployment and rapid response. In the context of emergency medical services (EMS), UAVs have the potential to augment ambulances by leveraging bystander assistance, thereby reducing response times for delivering urgent medical interventions and improving EMS outcomes. Notably, the use of UAVs for opioid overdose cases is particularly promising as it addresses the challenges faced by ambulances in delivering timely medication. This study aims to optimize the integration of UAVs and bystanders into EMS in order to minimize average response times for overdose interventions. Methodology/results: We formulate the joint operation of UAVs with ambulances through a Markov decision process that captures random emergency vehicle travel times and bystander availability. We apply an approximate dynamic programming approach to mitigate the solution challenges from high-dimensional state variables and complex decisions through a neural network-based approximation of the value functions (NN-API). To design the approximation, we construct a set of basis functions based on queueing and geographic properties of the UAV-augmented EMS system. Managerial implications: The simulation results suggest that our NN-API policy tends to outperform several noteworthy rule- and optimization-based benchmark policies in terms of accumulated rewards, particularly for situations that are primarily characterized by high request arrival rates and a limited number of available ambulances and UAVs. The results also demonstrate the benefits of incorporating UAVs into the EMS system and the effectiveness of an intelligent real-time operations strategy in addressing capacity shortages, which are often a problem in rural areas of the United States. Additionally, the results provide insights into specific contributions of each dispatching or redeployment strategy to overall performance improvement. Funding: This work was supported by the National Science [Grant 1761022]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2022.0166
- 1 March 2024
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Manufacturing & Service Operations Management
- Vol. 26 (2) , 447-464
Shortening Emergency Medical Response Time with Joint Operations of Uncrewed Aerial Vehicles with Ambulances
Problem definition: Uncrewed aerial vehicles (UAVs) are transforming emergency service logistics applications across sectors, offering easy deployment and rapid response. In the context of emergency medical services (EMS), UAVs have the potential to augment ambulances by leveraging bystander assistance, thereby reducing response times for delivering urgent medical interventions and improving EMS outcomes. Notably, the use of UAVs for opioid overdose cases is particularly promising as it addresses the challenges faced by ambulances in delivering timely medication. This study aims to optimize the integration of UAVs and bystanders into EMS in order to minimize average response times for overdose interventions. Methodology/results: We formulate the joint operation of UAVs with ambulances through a Markov decision process that captures random emergency vehicle travel times and bystander availability. We apply an approximate dynamic programming approach to mitigate the solution challenges from high-dimensional state variables and complex decisions through a neural network-based approximation of the value functions (NN-API). To design the approximation, we construct a set of basis functions based on queueing and geographic properties of the UAV-augmented EMS system. Managerial implications: The simulation results suggest that our NN-API policy tends to outperform several noteworthy rule- and optimization-based benchmark policies in terms of accumulated rewards, particularly for situations that are primarily characterized by high request arrival rates and a limited number of available ambulances and UAVs. The results also demonstrate the benefits of incorporating UAVs into the EMS system and the effectiveness of an intelligent real-time operations strategy in addressing capacity shortages, which are often a problem in rural areas of the United States. Additionally, the results provide insights into specific contributions of each dispatching or redeployment strategy to overall performance improvement. Funding: This work was supported by the National Science [Grant 1761022]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2022.0166
- 14 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28 (2) , 1122-1133
An Improved Statistical Modeling Approach to Individual Anticholinergic Drug Use Trend Analysis
Anticholinergic (AC) drugs are commonly prescribed to older adults for treating diseases and chronic conditions, such as chronic obstructive pulmonary disease, urinary incontinence, gastrointestinal disorder, or simply pain and allergy. The high prevalence of AC drug use can have a detrimental effect on the mental health of older adults. We aim to improve the prediction of future trends of AC drug use at the individual level, with pharmacy refill data. The individual drug use data presents challenges in the modeling, such as data being discrete-valued with excess zeros and having significant unobserved heterogeneity in the trend pattern. To address these challenges, we propose a statistical model of hierarchical structure and an EM scheme for the model parameter estimation. We evaluate the proposed modeling approach through a numerical study with synthetic data and a case study with real-world pharmacy refill data. The simulation study show that our analysis method outperforms the existing ones (e.g., reducing MSE significantly), particularly in terms of accurately predicting the trend pattern. The real-world case study further verifies the out-performance and demonstrate the advantageous features of our method. We expect the prediction tool developed based on our study can assist pharmacists' decision on initiating or strengthening behavioral interventions with the hope of discontinuing AC drug misuse.
- 14 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28 (2) , 1122-1133
An Improved Statistical Modeling Approach to Individual Anticholinergic Drug Use Trend Analysis
Anticholinergic (AC) drugs are commonly prescribed to older adults for treating diseases and chronic conditions, such as chronic obstructive pulmonary disease, urinary incontinence, gastrointestinal disorder, or simply pain and allergy. The high prevalence of AC drug use can have a detrimental effect on the mental health of older adults. We aim to improve the prediction of future trends of AC drug use at the individual level, with pharmacy refill data. The individual drug use data presents challenges in the modeling, such as data being discrete-valued with excess zeros and having significant unobserved heterogeneity in the trend pattern. To address these challenges, we propose a statistical model of hierarchical structure and an EM scheme for the model parameter estimation. We evaluate the proposed modeling approach through a numerical study with synthetic data and a case study with real-world pharmacy refill data. The simulation study show that our analysis method outperforms the existing ones (e.g., reducing MSE significantly), particularly in terms of accurately predicting the trend pattern. The real-world case study further verifies the out-performance and demonstrate the advantageous features of our method. We expect the prediction tool developed based on our study can assist pharmacists' decision on initiating or strengthening behavioral interventions with the hope of discontinuing AC drug misuse.