You Chen, PhD
You Chen, Ph.D., is an Assistant Professor of Biomedical Informatics at Vanderbilt University Medical Center. He is the director of the Optimization of Health ProcEsses and Networks Laboratory (OHPENLab), which was established to address the growing need for care coordination research and development for the health information technology sector. He is also the co-director of the Health Information Privacy Laboratory (HIPLab).
Dr. Chen’s research is funded through various grants from the National Institutes of Health (NIH), and National Science Foundation (NSF) to construct methodologies and technologies that optimize healthcare process via learning of healthcare systems.
Dr. Chen’s research foci include medical data mining and machine learning, multi-site based transfer learning, clinical workflow mining, care team identification, disease progression path modeling, predictive analytics, disease profiling and personalized medicine, hospital readmission and patient length of stay analytics, natural language processing, blockchain technology, and health information security and privacy. Dr. Chen’s research has been incorporated into a variety of clinical setting including neonatal care, maternal care, and critical illness patient care.
Dr. Chen holds a doctoral degree in computer science from the Chinese Academy of Sciences. After his graduation, he has been continuously involving in the field of biomedical informatics at Vanderbilt. His ultimate goal in the field is to leverage medical data (e.g., electronic medical records, genetic data, and public findings), health information technology (e.g., mobile health, wearable devices, Restful API, OAuth) and computational modeling (e.g., graph neural networks, network analysis, temporal modeling) to build patient-centered care to improve their care quality and reduce their healthcare cost.
Discovering Virtual Provider Interaction Networks in the EHR: Interpretation and Impact on Patient Outcomes
The research goal of this project is to create data mining models (e.g., social network analysis models), statistical models (e.g., logistical regressions), and interpretation strategies (e.g., surveys and focus group interviews) to i) learn interaction networks of care providers from EHR audit logs and EHR medical data; ii) identify interaction patterns contributing to the improved patient outcomes; and iii) translate effective interaction patterns into actionable criteria. We have been doing many types of research to learn virtual care teams and clinical workflows and measure their relationships with patient outcomes (length of stay).
Chen Y, Lorenzi NM, Sandberg WS, Wolgast K, Malin BA. Identifying collaborative care teams through electronic medical record utilization patterns. Journal of the American Medical Informatics Association. 2017 Apr 1;24(e1):e111-20. This work was reported by VUMC reporter at http://news.vumc.org/2016/10/27/study-tracks-makeup-of-vumc-collaborative-care-teams/
Chen Y, Patel MB, McNaughton CD, Malin BA. Interaction patterns of trauma providers are associated with length of stay. Journal of the American Medical Informatics Association. 2018 Feb 22;25(7):790-9. Editor’s choice paper.
Chen Y, Lorenzi N, Nyemba S, Schildcrout JS, Malin B. We work with them? Healthcare workers interpretation of organizational relations mined from electronic health records. International journal of medical informatics. 2014 Jul 1;83(7):495-506.
TeamWAS: Team-Wide Association Study - discovering care teams to satisfy a patient’s medical needs
EHR data are rich resources in terms of patients’ medical data and providers’ activities. By applying data mining and machine learning, these databases can reveal variations of a patient’s medical needs, as well as corresponding care teams dealing with such variations. The goal of this research is to leverage advanced informatics approaches to align care teams with a patient’s medical needs, which will empower healthcare organizations to choose the most appropriate team for a specific patient. The complete story can be found at https://researchfeatures.com/2018/05/30/using-medical-records-improve-care/
Chen Y, Kho AN, Liebovitz D, Ivory C, Osmundson S, Bian J, Malin BA. Learning bundled care opportunities from electronic medical records. Journal of biomedical informatics. 2018 Jan 1;77:1-10.
Chen Y, Xie W, Gunter CA, Liebovitz D, Mehrotra S, Zhang H, Malin B. Inferring clinical workflow efficiency via electronic medical record utilization. In AMIA annual symposium proceedings 2015 (Vol. 2015, p. 416). American Medical Informatics Association.
Yan C, Chen Y, Li B, Liebovitz D, Malin B. Learning clinical workflows to identify subgroups of heart failure patients. In AMIA Annual Symposium Proceedings 2016 (Vol. 2016, p. 1248). American Medical Informatics Association.
Discovering Risk Factors and Progression Paths of Perinatal Morbidity
Maternal and neonatal morbidity and mortality have been rising in the United States. To improve pregnancy and neonatal outcomes, we aim to leverage advanced data mining (e.g., word embeddings) and machine learning algorithms (e.g., neural networks) along with data in EHRs to identify risk factors (e.g., lifestyle choices, social determinants) and progression paths of perinatal morbidity. Specially, we have been working on the identification of risk factors or progression pathways for preterm birth, neonatal encephalopathy, and several maternal morbidity.
Li T, Gao C, Yan C, Osmundson S, Malin BA, Chen Y. Predicting neonatal encephalopathy from maternal data in electronic medical records. AMIA Summits on Translational Science Proceedings. 2018;2017:359.
Gao C, Yan C, Osmundson S, Malin BA, Chen Y. A deep learning approach to predict neonatal encephalopathy from electronic health records. The 7th IEEE International Conference on Healthcare Informatics. 2019; In press
Gao C, Osmundson S, Yan X, Malin BA, Chen Y. Leveraging electronic health records to learn the progression path for severe maternal morbidity. MedInfo 2019. In press
Gao C, Osmundson S, Yan X, Malin BA, Chen Y. Learning to identify severe maternal morbidity from electronic health records. MedInfo 2019. In press
Zuber C, Chen Y. A temporal pattern discovery algorithm for predicting neonatal mortality in NICU postoperative patients. MedInfo 2019. In press
Transfer learning – developing a common informatics approach to harmonize medical concepts across healthcare organizations
To improve the generalizability and reproducibility of phenotypes learned from EHR data of each individual site, we propose to develop a data mining and machine learning based framework to align phenotypes learned from each site into a common phenotypic space. It will then leverage the common phenotypic space to characterize the medical needs of a patient in each site.
Chen Y, Ghosh J, Bejan CA, Gunter CA, Gupta S, Kho A, Liebovitz D, Sun J, Denny J, Malin B. Building bridges across electronic health record systems through inferred phenotypic topics. Journal of biomedical informatics. 2015 Jun 1;55:82-93. Editor’s choice paper.
Zheng T, Xie W, Xu L, He X, Zhang Y, You M, Yang G, Chen Y. A machine learning-based framework to identify type 2 diabetes through electronic health records. International journal of medical informatics. 2017 Jan 1;97:120-7.
Discovering Effective Hidden Coordination in EHRs to Improve Patient Safety and Outcome in the Neonatal Intensive Care Unit
It has been suggested that virtual care coordination via electronic health record systems (EHRs) can resolve many current neonatal care-related challenges such as poor interpersonal communication, as well as insufficient provider-to-provider handoffs, all of which have the potential to increase medical errors and extend the length of stay. We developed a network-based approach to model virtual care team structures in the neonatal intensive care unit and measure their relationships with non-routine events and length of stay.
Kim C, Lehmann C, Schildcrout J, Hatch D, France D, Chen Y. Learning provider interaction networks in the neonatal intensive care unit and measuring their relationship with length of stay. AMIA. Clinical Informatics. 2019. In press.
Application of Blockchain in Health Security and Privacy
Access to accurate and complete medication histories across healthcare institutions enables effective patient care. Health institutions currently rely on centralized systems for sharing medication data. However, there is a lack of efficient mechanisms to ensure that medication histories transferred from one institution to another are accurate, secure and trustworthy. we introduce a decentralized medication management system (DMMS) that leverages the advantages of blockchain to manage medication histories.
Li P, Nelson SD, Malin BA, Chen Y. DMMS: A Decentralized Blockchain Ledger for the Management of Medication Histories. Blockchain in Healthcare Today. 2019 Jan 4:2(38). https://doi.org/10.30953/bhty.v2.38
Learning Opportunities for Drug Repositioning via GWAS and PheWAS Findings
The first goal of this project is to leverage Genome-Wide Association Studies (GWAS) and Phenome-Wide Association Studies (PheWAS) findings to discover new clinical targets for existing drugs. The second goal of this project is to learn disease relational patterns from electronic medical records, and then validate if such associations are genetic related.
Yin W, Gao C, Xu Y, Li B, Ruderfer DM, Chen Y. Learning Opportunities for Drug Repositioning via GWAS and PheWAS Findings. AMIA Summits on Translational Science Proceedings. 2018;2017:237.
Insider Threat in the Healthcare Setting
Insider threats in electronic medical record systems are increasingly relied upon to manage sensitive information. Healthcare organizations, for example, have adopted EHRs to enable timely access to patient personal data. However, the detail and sensitive nature of the information in EHRs make them attractive to numerous adversaries. This is a concern because the unauthorized dissemination of information from such systems can be catastrophic to both the managing agencies and the individuals to whom the information corresponds. We rely upon social network analysis technologies to detect or prevent patient information from being illegally accessed by hospital employees. We designed community-based anomaly detection systems to detect anomalous insiders insider actions.
Chen Y, Nyemba S, Malin B. Detecting anomalous insiders in collaborative information systems. IEEE transactions on dependable and secure computing. 2012 May;9(3):332-44.
Chen Y, Nyemba S, Zhang W, Malin B. Specializing network analysis to detect anomalous insider actions. Security informatics. 2012 Dec;1(1):5.
Chen Y, Malin B. Detection of anomalous insiders in collaborative environments via relational analysis of access logs. In Proceedings of the first ACM conference on Data and application security and privacy 2011 Feb 21 (pp. 63-74). ACM.
Chen Y, Nyemba S, Malin B. Auditing medical records accesses via healthcare interaction networks. In AMIA Annual Symposium Proceedings 2012 (Vol. 2012, p. 93). American Medical Informatics Association.