Wendy Webber Chapman
Biomedical Informaticist
Email:
Address:
Melbourne, Victoria, Australia
Current CV here
Who is Wendy - the Failure Version
People always look better on paper - here is the version of my career from inside my head
I stumbled onto biomedical informatics after studying Linguistics and Chinese and failing to secure a paid fellowship for a PhD program at the University of Wisconsin, Madison in Chinese Literature. My husband entered the field and introduced me to natural language processing, which applies linguistics to healthcare data. It's been 30 years since I started a PhD program at the University or Utah with Peter Haug and under the leadership of Homer Warner--in hindsight, perhaps I should have studied clinical decision support, because that is what really excited me in graduate school, and NLP ended up being so far away from impact on patients. Through a winding path that took me to Pittsburgh, San Diego, Utah again to be the chair of the department I graduated from, and now Melbourne, Australia, I am still seeking a way to have some impact in healthcare through analysis of clinical data and implementation of technology.
​
I led the Biomedical Language Understanding Lab (BLULab), and worked with many wonderful people to develop algorithms, annotation schemas and information models, and datasets. In hindsight, there are many things that contributed to our NLP methods and tools not having the impact I eventually craved, including not applying FAIR data principles, lack of rigorous software development, and not paying attention to questions like
-
Do users even need what we are building? What problem are we solving?
-
How would the NLP application fit into a larger ecosystem of tools?
-
How would the application support human workflows?
-
What is a feasible financial model for implementing and scaling the application?
-
What difference will the NLP tool make to health or healthcare delivery?
This crisis of relevance caused me to pivot to leadership roles with the aim of supporting other researchers and innovators in thinking about these questions from the start, to design and develop digital interventions that are more likely to be successfully implemented in healthcare.​ I can't believe I'm in the last phase of my career. I can look back and feel proud of what I've accomplished while I also look forward and hope to achieve more impact.
​
Career Summary
I annotated the outputs I've listed in this summary with the failures or caveats that went along with them.
2000 - 2010
University of Pittsburgh
​
-
Assistant Professor
-
Associate Director of the Training Program
-
Member of the DBMI Steering Committee
Will Bridewell developed NegEx. I joined the project later and asked to be first author on the 2001 paper due to implementing a baseline algorithm and writing the paper for publication.
I wish I would have released a software version of NegEx and ConText and built a community around it. Many others, including my husband Brian, created executable versions later.
-
Co-developed the NegEx algorithm for determining whether a clinical condition is negated
-
Extended the algorithm to other attributes like how much certainty is ascribed, whether it occurred in the past, and who experienced it (ConText algorithm). ConText has been translated into many other languages and applied around the world.
-
Hosted a dataset of UPMC clinical notes and ICD codes for NLP research that was used by 600 researchers in the year it was available.
-
Applied NLP to the field of disease surveillance with a focus on syndromic surveillance from chief complaint and ED notes.
-
Developed and taught courses on NLP and biomedical informatics.
-
Chaired the AMIA NLP Working Group
When I left Pittsburgh, I did not properly organize a handover of the repository and it became a political mess without a local champion, so it eventually closed down.
I struggled to teach clinical NLP because I lacked experience in what I was teaching (coding a syntactic parser, for example). I slowly quit programming, and that left a deficit in teaching and ability to interact with my own tools.
2010 - 2013
University of California,
San Diego
- Associate Professor
-
Member of the DBMI Executive Committee
2013 - 2019
University of Utah
-
Professor
-
Department Chair
-
Co-hosted ShARe and SemEval shared tasks that allowed researchers to compare performance of their algorithms on reference annotated datasets.
-
Developed a set of NLP ontologies that provide templates for representing clinical events and entities, along with their attributes, to enable faster startup of an NLP project and sharing of NLP knowledge bases.
-
Applied NLP to creating retrospective cohorts for comparative effectiveness studies.
-
Created the Data Science Scholars Program to mentor and guide clinical scientists applying NLP to clinical notes.
-
Led design and launch of a professional master's program in biomedical informatics and was the PI on the NLM training grant.
-
Coordinated the ReimagineEHR program for development and implementation of SMART on FHIR apps with existing EHRs.
-
Started and chaired the inaugural Women in AMIA committee and launched a podcast and a leadership program.
The learning curve to use the ontologies we developed was too high--most researchers in my lab wouldn't use them because it was much easier to just build a local information model for each project. The ontology designer improved the open source version an applied them in a commercial entity, which was a better fit.
The Scholars program would have been so much stronger if I would have had the confidence to pursue it as a department activity rather than running it on my own with help from my research lab. My learnings from that have shaped the LHS Academy :)
The WIA Leadership Program has been quite successful but has relied on a few people, so I have had a hard time keeping it going from Australia. Only two cohorts have gone through since 2019. It will benefit from being run by AMIA, which I will pursue, and from being more inclusive (ie, not women only).
2019 - present
University of Melbourne
-
Professor
-
Director, Centre for Digital Transformation of Health
-
Set up and built a new centre for informatics and digital health.
-
Co-created the Digital Health Validitron, a research platform to support development and validation of digital health innovations for clinical settings.
-
Launched the Learning Health System Academy for project-based training of clinicians selected by their healthcare organisations to become digital health champions, including a short course on Applied Learning Health Systems.
Find Me
