Events Add an event Speakers Talks Collections
 
Duration 03:16
16+
Video

C25 The use of natural language processing to improve identification of patients...

Hope Weissler
Vascular surgery resident at Duke University
+ 5 speakers
  • Video
  • Table of contents
  • Video
Machine Learning for Healthcare
August 8, 2020, Online, Los Angeles, CA, USA
Machine Learning for Healthcare
Request Q&A
Machine Learning for Healthcare
From the conference
Machine Learning for Healthcare
Request Q&A
Video
C25 The use of natural language processing to improve identification of patients...
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Add to favorites
81
I like 0
I dislike 0
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
  • Description
  • Transcript
  • Discussion

About speakers

Hope Weissler
Vascular surgery resident at Duke University
Jikai Zhang
PHD Graduate Student at Duke University
Steven Lippmann
Senior Analyst Programmer at Duke Department of Population Health Sciences
Shelley Rusincovitch
Associate Director of Informatics at Duke AI Health and Duke Forge
Ricardo Henao
Assistant Professor in Biostatistics and Bioinformatics at Duke University
Schuyler Jones
Associate Professor of Medicine at Duke University

I'm interested in limb salvage, aortic disease, and vascular trauma. Academically, I'm interested in improving diagnosis and care of peripheral artery disease, specifically in regards to disparities in care and outcomes.

View the profile

Jikai is a PhD student in the Electrical and Computer Engineering Department. Prior to joining Dr. Mazurowski’s team, he completed his master’s degree in Biostatistics at Duke University.

View the profile

Shelley Rusincovitch is an informaticist and technical leader with extensive background in healthcare data, clinical trials, and outcomes registries. In her role with the Forge, she co-leads the Demonstration Program with a portfolio of project illustrating the “art of the possible” in the vision of actionable health data science, and, in collaboration with the Forge Principal Data Scientist, provides direction and operational leadership for the transdisciplinary teams. She also manages the Data Science Core and oversees the Health Data Science Internship Program in partnership with the Duke Clinical Research Institute. Before joining the health data science initiative in 2017, Ms. Rusincovitch led the development of the PCORnet Common Data Model and managed the PCORnet Data Committee, a governance entity overseeing the data network, informatics, and research data innovations. Her previous technical experiences as a database developer include registry systems, economics/quality-of-life research studies, clinical trials data management, and health system enterprise data warehousing. Her specialties include solutions design and architecture, data modeling, distributed research networks, convening and facilitating stakeholder groups, and machine learning applications with health data.

View the profile

Ricardo is a respected expert in machine learning. He serves as Assistant Professor in Biostatistics and Bioinformatics at Duke University and is a member of the Duke Clinical Research Institute. He also regularly serves as a reviewer for journals such as the Journal of Machine Learning Research, Journal of Biomedical Informatics, and many others.He has coauthored more than 50 academic papers in fields such as machine learning, biostatistics, neuroscience, cancer and infectious disease.

View the profile

Dr Schuyler Jones is an interventional cardiologist and Associate Professor of Medicine in the Division of Cardiology at Duke University Medical Center, and his clinical interests include peripheral artery disease, coronary artery disease, and percutaneous coronary intervention. He currently serves as the Medical Director of the Duke Adult Cardiac Catheterization Laboratory and the medical director of the Duke Heart Center Clinical Research Unit (that oversees site-based research within the Heart Center). Since joining the Duke Faculty in 2010, Dr Jones has worked as an investigator at the Duke Clinical Research Institute where he has been part of the leadership team for several clinical trials across a broad spectrum of cardiovascular care, including CAD, PAD, PCI, and CABG. He serves as the Co-Medical Director of the Clinical Endpoints Committee (CEC) at the DCRI, and has been an active participant in the angiographic core laboratory at DCRI. Recently, he has joined the Duke CTSA Recruitment Innovation Center as a co-Director. Currently, he is a co-principal investigator of the PCORI-funded demonstration project called Aspirin Dosing: A Patient-Centric Trial Assessing Benefits and Long-Term Effectiveness (ADAPTABLE) study which is enrolling 15,000 patients with established cardiovascular disease at 40 clinical sites within PCORnet.

View the profile
Share

Hello everyone. My name is Hope eissler and I'm excited to be sharing our work with you today. We have no relevant disclosures. The aim of this collaboration was to come up with a better way to identify patients, with peripheral artery disease and the electronic health record. Kiddie affects about 10 million Americans and leaves the heart attacks, education's, other surgeries and death. The care of PID is complex and expensive. And furthermore is subject to several race gender and geography related disparity, in order to improve the care of p80. We need to be able to reliably and

efficiently find station to have it. We also need to be able to construct cohorts of patients, who has PID for further investigation. Currently p. I d identification. In electronic health record generally uses administrative data which often takes the form of diagnosis and procedure code as well as other structured data that has been created for billing rather than, for clinical or research purposes. This approach either necessitates selecting a t, a t subgroup, such as only patients with a specific type of theater or only patients who have undergone certain procedures

or ends of yielding. A cohort of patients of whom many don't actually have PID. And we know this because we've tried it and we actually ended up building an algorithm that uses administrative data through a lasso approach, which outperform diagnosis codes alone. We decided to make use of patients, unstructured data using a natural language processing basis. This data came from Clinical notes, which contain unstructured and semi-structured narrative, such as history, physical exam, and list of medication procedures and comorbidities To do this. We took

clinical note from a cohort of patients whose Pat status had previously been adjudicated with the help from the lasso model I mentioned earlier. We fed those notes into a hierarchical modification of a label and Benning attentive model, which we called a demo. As you can see here, pademelon blue significantly, outperformed the administrative database approach using a Delong test. Paramount also outperformed the administrative database approach on a Precision recall curve. By the way, the words identified by pademelons, most strongly associated with PID. Make clinical sense

as erecta me and sending our endovascular intervention, use to improve blood flow in patients, who have Katie and many of the rest of these terms are an atomic location that are frequently affected by p a t. In conclusion we show that natural language processing outperform and administrative database approach for p. A d identification. We felt comfortable with the words identified by Pat amount of Highly relevant as a make sense from a medical standpoint. We're now working on next test, which are to work with other institutions to deploy by the mail. And there is a charge for

validation and cohort, construction. Perfect, thank you so much for your time and attention.

Cackle comments for the website

Buy this talk

Access to the talk “C25 The use of natural language processing to improve identification of patients...”
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free

Ticket

Get access to all videos “Machine Learning for Healthcare”
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Ticket

Interested in topic “Artificial Intelligence and Machine Learning”?

You might be interested in videos from this event

February 4 - 5, 2021
Online
26
104
ai, application, bot, chatbot, conversation, data, design, healthcare, ml

Similar talks

Divya Gopinath
Research Engineer at TruEra
+ 5 speakers
Monica Agrawal
Software Engineering Intern at Flatiron Health
+ 5 speakers
Luke Murray
Research Intern at Microsoft
+ 5 speakers
Steven Horng
Associate Director, Multi-disciplinary Fellowship in Clinical Informatics at Beth Israel Deaconess Medical Center
+ 5 speakers
David Karger
Advisory Board Member at Instabase
+ 5 speakers
David Sontag
PhD, Associate Professor of Electrical Engineering and Computer Science at MIT
+ 5 speakers
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Ziad Obermeyer
MD, MPhil, Acting Associate Professor of Health Policy and Management at School of Public Health, UC Berkeley
+ 1 speaker
Byron Wallace
Associate Professor at Northeastern University
+ 1 speaker
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Ariane Marelli
Professor of Medicine at McGill University
+ 10 speakers
Chao Chao Li
E. B. Eddy Professor and Canada Research Chair at McGill University
+ 10 speakers
Aihua Liu
Psychologist at McGill University
+ 10 speakers
Hanh Nguyen
Рsychologist at McGill University
+ 10 speakers
James Brophy
Professor at McGill University
+ 10 speakers
Liming Guo
Research at McGill University
+ 10 speakers
David Buckeridge
Professor at McGill University
+ 10 speakers
Jian Tang
Assistant Professor at HEC Montreal
+ 10 speakers
Joelle Pineau
Associate Professor at McGill University
+ 10 speakers
Yi Yang
Associate Professor of Statistics at McGill University
+ 10 speakers
Yue Li
Assistant Professor at McGill University
+ 10 speakers
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free

Buy this video

Video
Access to the talk “C25 The use of natural language processing to improve identification of patients...”
Available
In cart
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free

Conference Cast

With ConferenceCast.tv, you get access to our library of the world's best conference talks.

Conference Cast
949 conferences
37757 speakers
14408 hours of content
Hope Weissler
Jikai Zhang
Steven Lippmann
Shelley Rusincovitch
Ricardo Henao
Schuyler Jones