Previous Informatics Research Seminars

2022  |  2021  |  2020  |  2019  |  2018

Duke University

Broadcast Link: Seminar

Abstract:

The Duke Institute for Health Innovation (DIHI) was founded in 2013 with a mission to catalyze innovation in health and health care. In the last eight years, the transdisciplinary team at DIHI has completed over 85 innovation pilots and has facilitated investment in over 12 companies. Through an annual request for applications to select priority projects for investment, Duke Health has emerged as a national leader in the design, development, and integration of machine learning and augmented intelligence into clinical care. This presentation will describe DIHI’s approach to innovation and will share principles and practices that enable repeated value creation, with a focus on AI / ML product development.

Biosketch:

Mark Sendak, MD, MPP is the Population Health & Data Science Lead at the Duke Institute for Health Innovation (DIHI), where he leads interdisciplinary teams of data scientists, clinicians, and machine learning experts to build technologies that solve real clinical problems. He has built tools to transform chronic disease management within Duke Health’s Accountable Care Organization, detection and management of inpatient deterioration within Duke Health hospitals, and community health programs that empower community based organizations and community health workers during the COVID-19 pandemic. Together with his team, he has integrated dozens of data-driven technologies into clinical operations and is a co-inventor of software to scale machine learning applications and real world evidence generation across health systems. He leads the DIHI Clinical Research & Innovation scholarship, which equips medical students with the business and data science skills required to lead health care innovation efforts.

He and his team have published in technical venues such as the Machine Learning for Healthcare Proceedings and Fairness, Accountability, and Transparency in Machine Learning Proceedings and clinical journals such as Plos Medicine, Nature Medicine, and Nature Digital Medicine. Their work has been featured in MIT Technology ReviewWired, STAT News, and The Wall Street Journal. He has served as an expert advisor to national organizations, including the American Medical Association, AARP, American Academy of Family Practice, and National Academies of Medicine, on matters related to health innovation, machine learning, and policy. He is a program chair for the annual Machine Learning in Health Care Conference (MLHC). He was named a STAT Wunderkind in 2020 for his efforts to responsibly build and integrate AI into clinical practice. He obtained his MD and Masters of Public Policy at Duke University as a Dean’s Tuition Scholar and his Bachelors of Science in Mathematics from UCLA.

UNC-CH

Broadcast Link:  Seminar

Abstract:

 NC DETECT, North Carolina’s statewide syndromic surveillance system, provides near real time access to data from emergency departments, urgent care centers, EMS, the statewide health information exchange (NCHealthConnex), and North Carolina Poison Control. Users from health departments and data providers access NC DETECT to monitor trends in infectious disease, injury, chronic disease, and mental health. NC DETECT is designed, developed, and maintained by the Carolina Center for Health Informatics in the Department of Emergency Medicine at UNC Chapel Hill with funding from the North Carolina Division of Public Health. This presentation will focus on efforts to improve mental health surveillance using NC DETECT emergency department visit data, specifically at the county level.  

Biosketch:

Amy Ising, PhD, is the Associate Director of the Carolina Center for Health Informatics in the Department of Emergency Medicine, UNC Chapel Hill. Her work and research focus on public health surveillance, specifically syndromic surveillance, and she is a co-investigator on CDC-funded efforts to improve the use of syndromic surveillance for non-fatal overdose and firearm-related injury. Dr. Ising has a PhD in Health Informatics and a MS in Information Science, both from UNC Chapel Hill.

Presented from UNC-C

Broadcast Link:  Seminar

Abstract:

The American senior population is continuing to increase rapidly. By the year 2040, we expect the number of citizens aged 65 and older to reach 80 million. Due to the growing demand for “age-at-home” and shortage of human resources (i.e., medical staff), there is an essential need for technologies that can provide health monitoring for seniors who are living independently. This talk discusses Artificial Intelligence (AI)-enabled health monitoring system that is capable of detecting any potentially harmful condition and notifying the responsible individuals. The aim is to create AI-based pose detection and predictive model for real-time detection and behavioral analysis of human health-related activities (such as walking, standing, and taking medicines as Activities of Daily Living (ADL)) and high-risk events such as patients’ falling. At the same time, this talk discusses many technical challenges stemming from real-world uncertainties, privacy concerns, noisy input data, miss-calibration across different cameras, and the lack of sufficient labeled data for training AI algorithms.

Biosketch:

Hamed Tabkhi is the assistant professor of computer engineering at the University of North Carolina Charlotte. He received his Ph.D. in 2014 from Northeastern University and then was a post-doctoral fellow at Northeastern University for two more years. Hamed’s research focuses on transformative computer system solutions to bring recent advances in machine learning and data analytics to enhance our communities’ health, safety, and overall well-being. During the last five years at UNC Charlotte, he has advised 29 graduate students (12 PhDs and 19 Master students) and received $4M research funding.

NCCU

Broadcast Link: Seminar

Abstract:

Gulf War Illness (GWI) is a collection of potentially debilitating symptoms that may be present in many veterans (about 30%) of the 1990’s Gulf War. GWI symptoms have been reported in an unknown number of civilians (like media personnel and government employees) of the first Gulf War. The potential role of exposure to chemical weapons (such as sarin and soman) has been well documented along with supportive evidence for other possible toxicity resulting from the exposure to OPs (Organophosphate Pesticides), PB (Pyridostigmine Bromide), CARC (Chemical Agent Resistant Coating), diesel and petrochemical fumes, decontamination solution 2 (DS-2), pollutants emanating from oil well fire smoke and depleted uranium. However, other suspected but not proven exposome or measured exposure factors include infectious diseases, biological weapons, and numerous vaccinations against various illnesses. Critical analysis of the data published by Dr. Damodaran’s group over several years, as well as other relevant research, brings out several novel findings, which may lead to a personalized medicine approach based on annual biomarker screenings and paradigms of affected individuals. These salient findings deserve research focus: A) Gene expression studies in rat model exposed to sarin showed persistence of altered mRNA levels for genes of various categories and molecular pathways at various post-exposure time points (15 minutes, 2 hours, 3 months, and 6 months); B) Higher degree of Inhibition in cholinesterase levels accompanied by memory deficits and increased sensory-motor dysfunction occurred in rat models with exposure to mixture of chemicals used by soldiers during war; C) Functional mapping of altered gene expression of human homologues indicated the role of evolution in organ system pathology of affected individuals; D) Human disorders and disease conditions connected to altered gene products identified from Dr. Damodaran’s group studies showed abnormal changes in cell membrane function, organelle dysfunction, tissue deformities and organ system failure as the cause of clinical severity. Preliminary data on exposome model for GWI will be presented to define clinical presentation and progression of the disease process.

Biosketch:

Dr. Tirupapuliyur Damodaran is Research Assistant Professor at NC Central University’s department of Biological and Biomedical Sciences. His research background includes molecular toxicology and medical genetics with special emphasis on gene expression studies to develop models for disease progression in exposure-induced disorders and genetic disorders. During his tenure at Duke University Medical Center, Dr. Damodaran was involved in several research projects of global gene expression studies that addressed the variabilities of clinical symptoms for the early and late onset forms of genetic disorders (Glycogen Storage Disorders I and II). He was also involved in genetic and molecular studies of kidney disorders such as FSGS (Follicular Segmental Glomerular Sclerosis). His collaborative research efforts with Duke investigators resulted in several novel findings from these research projects. He has been teaching human anatomy and physiology along with an introduction to research course at NCCU. He is also a participating faculty in GEAR (Global Education, Academics and Research Skills) from NC State University, mentoring international students in research projects. He has been actively involved in various capacities at SOT (Society of Toxicology) for the last several years. As a member of a Diversity Initiative efforts team, he has been involved in mentoring research projects by students from HBCUs that resulted in productive data sets presented at SOT conferences. As the founding adviser of Tri Beta Biology Honor Society at NCCU, he has brought programs on leadership training, public awareness, educational, and professional pathways to NCCU. His current research includes environmental health disparities and data analytics relevant to human disorders.

ECU

Broadcast Link:  Seminar

Abstract:

Dr. Mi Hwa Lee will present how culturally tailored interventions combined with mHealth technology and a community-based participatory research approach impact ethnic minority women’s beliefs and attitudes in promoting cancer screenings. Specifically, Dr. Lee will talk about her two past research experiences: 1) a text message-based intervention to promote HPV vaccine and Pap test, and 2) a multimedia messaging intervention to increase breast cancer screening rates in Korean American immigrant women. She will also discuss her current grant proposal plan using a mHealth technology to promote mammography use among Korean American immigrant women in NC.

Biosketch:

Mi Hwa Lee, PhD, MSW, MA, is an Assistant Professor of the School of Social Work at East Carolina University. She earned her Master and PhD in Social Work from University of Minnesota, Twin Cities. Her research interests connect health behavior changes to a reduction in disparities, particularly focusing on cancer screening disparities among Asian American immigrants. Her current research focuses on understanding the impact of social, cultural, and physical environment factors on breast cancer screening behaviors in Korean American immigrant women, as well as develop an intervention to increase mammography use in this group. The ultimate goal of her research is to inform the development and implementation of relevant public policies and practices for reducing cancer screening disparities in Asian American immigrants.

Wake Forest

Broadcast Link:  Seminar

Abstract:

Dr. Da Ma will be presenting his work during his postdoctoral fellowship in Alzheimer’s Society Research Program. The topic is these studies focused on building machine learning and AI models to achieve differential diagnosis for dementia subtypes, as well as prognosis for predicting the future onset of dementia of Alzheimer’s type. This work utilizes the phenotypes and genotypes that are embedded in the neuroimaging as well as genomic data using methods such as stratified multi-modal feature ensemble and deep survival analysis, along with data harmonization techniques as well as methods to improve model explainability for potential clinical translation.

Biosketch:

Dr. Da Ma received his doctoral degree from University College London, concentrated in “Medical and Biomedical Imaging”. He then conducted his postdoctoral research at Simon Fraser University School of Engineering Science, in British Columbia, Canada, working on building computational neuroanatomy methods for age-related neurodegenerative diseases. He was awarded the Alzheimer’s Society Research Program fellowship in 2018, and in Dec 2021, he joined Wake Forest University School of Medicine as an Assistant Professor in the Center for Biomedical Informatics, and Alzheimer’s Disease Research Center in the Department of Internal Medicine, Section of Gerontology and Geriatric Medicine.

His research interest is to use biomedical informatics methods in clinical applications to achieve precision diagnosing, early intervention, and enhance our understanding of age-related diseases such as neurodegeneration and dementia. His research works involve medical image computing, computational neuroanatomy, and machine learning techniques, with applications ranging from brain imaging, retinal imaging, and recently genomics, and translational applications on both clinical and small animal studies.

Duke

Broadcast Link:  Seminar

Abstract:

Reliance on exam reporting of unexpected imaging findings does not ensure receipt of findings or appropriate follow up. A closed-loop communication system should include provider and patient notifications and be auditable through the electronic health record (EHR). An EHR-integrated program was designed at Duke Radiology to enable identification and communication of unexpected findings and aid in next steps in findings management. Three navigators (with prior training as radiologic technologists and sonographers) facilitated communication and documentation of results to providers and patients. Within the first 12 months of use, the most common unexpected findings were potential neoplasms (11%), and the median time between performance of exam and patient notification was 12 days (0-136 days, IQR=13). A total of 2,127 additional imaging studies were performed and 1,078 patients were referred to primary care providers and specialists. Radiologists (89%) and providers (65%) found the system useful and used it most frequently during regular business hours. Overall, the EHR-integrated, navigator-facilitated closed-loop communication program for unexpected findings in radiology led to near complete success in notification of providers and patients and facilitated next steps in findings management.

Biosketch:

Dr. Fides Schwartz trained as an MD in Heidelberg, Germany and did her residency in Radiology in Basel, Switzerland. After that, she joined Duke Radiology to perform research mainly on CT imaging. She is an associate in research with the Duke Department of Radiology and has been in this position for four years. This project was important to her because one of her major concerns as a resident was what happened to the patients after we finished reporting on them; it was always difficult to find out what happened to them in follow-up and having a system like the one we built here at Duke would have been ideal. The program was awarded with the 2020 Imaging Innovations award and has continued to improve patient care.

UNC-CH

Broadcast Link:  Seminar

Abstract:

Use of virtual reality (VR) training in healthcare is becoming an increasingly feasible and effective method for training; research suggests that VR training in healthcare can improve technical and non-technical skills including but not limited to team communication and teamwork, planning and performing a surgery, and medical diagnosis. The use of VR training for the education on non-technical skills, such as the understanding of the holistic view of the culture of patient safety and high-reliability, as well as the sense making of patient safety events, has been less commonly utilized in healthcare. In addition, most of UNC health system current safety education is utilizing the web-page based learning management system or PowerPoint presentations. There is general dissatisfaction among clinicians about this type of training as it is not engaging, which could lead to important information not making sense to the learner. In this presentation I will talk about the development of a patient safety training program that promotes understanding and sense making towards the holistic view of the culture of patient safety and high-reliability. The program helps healthcare workers better understand the overall extent of patient safety from the ‘blunt’ to ‘sharp’ end of the error and, consequently, emphasize a culture of improvement at all levels of the organization.

Biosketch:

Dr. Khasawneh is a faculty member in the Division of Healthcare Engineering (DHE) at the Department of Radiation Oncology at UNC Chapel Hill. Prior to joining UNC, he was a Research Fellow at the Armstrong Institute at Johns Hopkins University and received his PhD in Industrial Engineering with a Human Factors focus from Clemson University in 2019. He also hold a Master’s degree in Industrial and Systems Engineering from the State University of New York at Binghamton and an undergraduate degree in Industrial Engineering from Jordan University of Science and Technology. His current research focuses on using both qualitative and quantitative research methods to improve patient safety as well as patient engagement and his other research interests are modeling online behavior and decision-making as well as applying human factors design principles to teleoperated human-machine systems. This work has been presented and published in several peer-reviewed publications such as Human Factors and Ergonomics Society Annual Meetings and Journal of Automation in Construction. Outside of work, he likes running and playing soccer and exploring any team based sport.

UNC-C

Broadcast Link:  Seminar

Abstract:

Automatic continuous monitoring (ACM) of patient vital signs (e.g., breath rate and heart rate) and risky behaviors (e.g., falls and unsupervised bed-exits) not only can enhance the patient safety, but also can detect clinical deterioration at a very early stage so that interventions can be swiftly performed to treat the patient’s condition and prevent adverse events, such as cardiac arrest and death. Despite the pressing needs of ACM, it is still not standard of care today because currently used methods for ACM have serious limitations. In particular, because current vital sign monitoring systems depend on invasive contact sensors and mobility-restricted cables, vital signs are continuously monitored mainly in the intensive care units (ICUs), and are only spot-checked every 4 – 8 hours in general care wards.  Continuous risky behavior monitoring, on the other hand, is still facing great feasibility challenges. Cameras and weight-sensitive pressure sensors have been used to detect risky behaviors. However, weight-sensitive pressure sensors, which are attached to bed or chair, cannot provide situation-awareness and are easily triggered by a very slight movement, resulting in a significant number of false alarms. The situation-aware camera systems, however, are not privacy preserving and may experience degraded performance under poor lighting conditions. This seminar will provide an overview of the proposed RadarHealth system, which exploits the deep learning-augmented millimeter wave (mmWave) radar for contactless noninvasive continuous patient monitoring, while guaranteeing human comfort and preserving user privacy. It will introduce how to apply the RadarHeath system for domain-invariant gait recognition and briefly showcase an AI-enabled medication adherence system based on AR glasses.

Biosketch:

Pu Wang received a Bachelor of Engineering degree in Electrical Engineering from Beijing Institute of Technology, China, in 2003, and a Masters of Engineering degree in Electrical and Computer Engineering from Memorial University of Newfoundland , Canada, in 2008. He received his Ph.D. in Electrical and Computer Engineering from the Georgia Institute of Technology, Atlanta, GA, in August 2013. Currently, he is an Associate Professor with the Department of Computer Science at the University of North Carolina at Charlotte. His current research interests focus on AI for networked systems, including deep learning for wireless/radar sensing, reinforcement learning for networking optimization, distributed/federated learning over wireless edge computing, and swarming intelligence for multi-robot systems.

NCCU

Broadcast Link:  Seminar

Abstract:

The story of John Henry, the “steel-drivin’ man”, is well known to Black men in the United States. John Henry is considered a hero because he demonstrated tremendous strength and self-determination although it is considered a double-edged sword. The MANUP diabetes program used the John Henryism, defined as high-effort active coping in the face of adversity, as the basis of a diabetes intervention for Black men. MANUP conducted four community-based focus groups identifying topics of concern to Black men with type 2 diabetes (T2D). This seminar will discuss findings demonstrating that combining mobile health technology and moderate physical activity with culturally targeted discussion topics can improve T2D self-management and reduce A1c in Black men.

Biosketch:

Dr. Dana Carthron is a Clinical Assistant Professor at NC Central University’s Department of Nursing and a board-certified adult-gerontology nurse practitioner. Her clinical background is public health with a focus on historically underrepresented and marginalized populations. She currently teaches community/public health.

She completed two fellowships including the Claire M. Fagan fellowship program funded by the John A. Hartford Building Geriatric Nursing Capacity program at the School of Nursing at Duke University and the Center for Health Equity Research (CHER) at the University of North Carolina at Chapel Hill. She most recently completed the Research in Implementation Science for Equity (RISE) Program funded by the National Institute of Health’s Programs to Increase Diversity Among Individuals Engaged in Health-Related Research (PRIDE).

Her research background includes interventions focusing on social determinants of health (SDOH) to improve chronic illness self-management, particularly diabetes, as well as caregiver health.  She is currently involved in several studies including a 3-arm clinical trial for African American men with diabetes in rural NC. She is a co-investigator on a clinical trial to develop an electronic medical record (EMR)-linked, closed-loop referral system as well as a planning grant to establish a practice-based equity research network with a local federally qualified health center (FQHC). She currently teaches Global Health Perspectives, Community/Public Health, Nursing Research and Advanced Health Assessment.

ECU

Broadcast Link:  Seminar

Abstract:

A significant majority (over 90%) of senior adults prefer to age in place in their own homes rather than to move out into assisted living or other long-term care facilities. Over time, most adults will eventually develop health conditions that may make it a challenge to live alone. Determining the appropriate living situation for aging adults with health concerns is often challenging given the security people feel in their own homes while also considering the risks of being in an environment that does not provide ongoing monitoring and care. Compounding the decision making regarding the appropriate living environment for senior adults is the high cost of long term care. Home healthcare is an option for some situations, but is often costly and may not be able to assess health conditions manifesting outside of the time that a care provider is in the home.

In this talk, Dr. Castles will present progress in an ongoing research project designed to collect sensor data from senior adults in a home environment. Data is collected via non-invasive sensors and transmitted over a network to allow for remote analysis. The goal of this project is to use sensor data to train digital learning systems in order to develop heuristics for alerting healthcare providers to the need for further evaluation while allowing senior adults to live alone in the privacy of their own homes while sensors monitor their behaviors and physiology and detect any indications of decline in health or adverse events before they become emergencies.

Biosketch:

Dr. Ricky Castles is an Associate Professor in the Department of Engineering at East Carolina University. His research centers around wireless sensor networks for healthcare applications and engineering education. He is particularly interested in the use of mobile sensor systems to allow for non-invasive physiological monitoring and the analysis of data predicting changes in health. He collaborates with researchers in several disciplines including computer science, nursing, physical therapy and medicine.

Duke

Broadcast Link:  Seminar

Abstract:

The quality of literature used as the foundation to any research or scholarly project is critical. In this research, we analyzed the extent to which predatory nursing journals were included in credible databases, MEDLINE, CINAHL, and Scopus, commonly used by nurse scholars when searching for information. Findings indicated that no predatory nursing journals were currently indexed in MEDLINE or CINAHL, and only one journal was in Scopus. Citations to articles published in predatory nursing journals are not likely found in a search using these curated databases but rather through Google or Google Scholar search engines.

Biosketch:

Marilyn Oermann is the Thelma M. Ingles Professor of Nursing at Duke University School of Nursing. She is the Editor-in-Chief of Nurse Educator. Dr. Oermann is the author/co-author of 26 books, more than 200 articles in peer-reviewed journals, and many editorials, chapters, and other types of publications. Her research is on predatory publishing in nursing and other studies of the nursing literature. She has completed multiple studies on the extent, quality, and citations of articles published in predatory nursing journals and research on predatory conferences, reference accuracy, and citations of nursing articles. Current studies include a scoping review of predatory publishing and a study on retractions of nursing articles.