Maxim Topaz
Elizabeth Standish Gill Associate Professor
Columbia University School of Nursing & Data Science Institute
Senior Research Scientist, VNS Health
Overview
Dr. Maxim Topaz PhD, RN, MA, FAAN, IAHSI, FACMI is an Associate Professor of Nursing at Columbia University. His research uses artificial intelligence and innovative technologies, including natural language processing, speech recognition and machine learning to improve healthcare. He is a pioneer in applying natural language processing to nursing data and is currently focused on developing AI solutions for clinical decision-making. Dr. Topaz has experience in health policy, leadership, and entrepreneurship. Dr. Topaz has an extensive history of research funding in the USA (e,g., AHRQ, NIA, NINR) and internationally. He has clinical experience in internal medicine and holds a Ph.D. from the University of Pennsylvania (USA) and a Master’s and Bachelor’s degrees from the University of Haifa (Israel). He completed a postdoctoral fellowship at Harvard Medical School and has published over one hundred articles on health informatics, earning numerous awards for his work.
Education & Training
•
BA, 2007 Nursing , Technion-Israel Institute of Technology (Israel)
•
PhD, 2014 Nursing Informatics , University of Pennsylvania, Philadelphia
•
2016 Health Informatics (Postdoctoral fellowship), Harvard Medical School
PROFESSIONAL EXPERIENCE
•
2016-2018 Assistant Porfessor, Unviersty of Haifa, Isael
•
2018-now Associate professor, Columbia University, NYC, USA
Selected Publications
•
Zolnoori, M., Sridharan, S., Zolnour, A., Vergez, S., McDonald, M. V., Kostic, Z., Bowles, K. H., (2023). Utilizing patient-nurse verbal communication in building risk identification models: the missing critical data stream in home healthcare. Journal of the American Medical Informatics Association : JAMIA, ocad195. Advance online publication.
•
Hobensack, M., Song, J., Oh, S., Evans, L., Davoudi, A., Bowles, K. H., McDonald, M. V., Barrón, Y., Sridharan, S., Wallace, A. S., & Topaz, M. (2023). Social Risk Factors are Associated with Risk for Hospitalization in Home Health Care: A Natural Language Processing Study. Journal of the American Medical Directors Association, S1525-8610(23)00621-7. Advance online publication.
•
Chae, S., Davoudi, A., Song, J., Evans, L., Hobensack, M., Bowles, K. H., McDonald, M. V., Barrón, Y., Rossetti, S. C., Cato, K., Sridharan, S., & Topaz, M. (2023). Predicting emergency department visits and hospitalizations for patients with heart failure in home healthcare using a time series risk model. Journal of the American Medical Informatics Association : JAMIA, 30(10), 1622–1633.
Harnessing AI to Identify and Correct Biases in Healthcare: A Research Trajectory
Abstract
The urgent need to tackle biases within healthcare is a burgeoning domain of investigation and intervention. This presentation will embark on a research trajectory focused on examining and rectifying healthcare biases utilizing Artificial Intelligence (AI). Initially, a broad overview of the topic will set the stage for delving into a series of studies centered on AI and biases. The first study to be discussed will highlight the potential impact of language within hospital care for pregnant and birthing individuals, exploring how stigmatizing language can reinforce social hierarchies and biases, potentially detrimentally affecting patient care. The subsequent study will extend this narrative by identifying stigmatizing language in electronic health record (EHR) notes of pregnant individuals during their birth admission, illuminating the pervasive nature of implicit biases among clinicians.
Transitioning to the realm of home health care, a third study will utilize Natural Language Processing (NLP) to uncover racial disparities in judgment language used by clinicians during home visits, establishing a linkage between judgment language and reduced visit durations, which could reflect on care quality. The fourth study will shift the spotlight towards machine learning, evaluating the fairness of predictive models for hospitalization and emergency department visits among heart failure patients receiving home health care. The disparities in model performance across demographic subgroups will be revealed, underscoring the necessity for continuous monitoring and enhancement in fairness metrics to mitigate biases.
Lastly, a fifth study, termed the ENGAGE study, will introduce an NLP-driven system aiming to identify and reduce stigmatizing language within home healthcare settings. This venture will not only accentuate the prevalence of biases in clinical documentation but also showcase a promising technological pathway to counter such biases. Through these diverse yet interconnected studies, this presentation will strive to engender a comprehensive understanding of healthcare biases and the potential of AI as a pivotal tool for uncovering and addressing these biases, thereby fostering enhanced healthcare equity and quality.
Clinical Care Quality Enhancement through Digital Health Advancement: Architectures and Experiences
Abstract
The imminent presentation explores the compelling capacity of Artificial Intelligence (AI) to harness extensive healthcare data to enhance clinical outcomes, primarily in the sphere of home healthcare. As we approach a significant demographic shift towards an aging global population, the demand for robust home-based healthcare solutions becomes paramount. AI, with its ability to emulate human cognition, emerges as a quintessential tool for advancing data processing, generating insights, and fostering automation across healthcare domains. The discourse will unfold the remarkable potential of AI in augmenting healthcare delivery tailored to individual and community needs, thereby enriching the lives of older adults. Through a series of exemplified initiatives and state-of-the-art AI applications, the presentation will unveil automated tools and natural language processing techniques aimed at identifying high-risk patients and developing unbiased risk predictions during routine home healthcare services. Additionally, innovative explorations into speech recognition and machine learning will be highlighted, showcasing their potential in enhancing risk prediction accuracy and uncovering under-documented patient problems. The talk will further delve into various ongoing large clinical trials, proposing a futuristic perspective on integrating these AI-driven tools into routine clinical practice, addressing biases, and redefining healthcare delivery. Through a blend of theoretical insights and practical examples, the presentation aims to foster an engaging narrative on leveraging AI to transcend traditional healthcare paradigms, thereby propelling the sector into a new era of informed, timely, and personalized care.