I am a trauma and acute care Nurse Practitioner, board-certified in both Family Nurse Practice (FNP) and Adult-Gerontology Acute Care (AGACNP).
I practice at Oregon Health & Science University (OHSU) on the Level 1 Trauma Service, where I care for critically injured patients across the continuum of trauma care — from early resuscitation and procedural management to postoperative and intermediate care.
My clinical work sits at the intersection of high-acuity decision-making, teamwork, and systems-based care, and it strongly influences my research interests.
I completed my PhD in Biomedical Informatics at the OHSU School of Medicine in the Health and Clinical Informatics (HCIN) track within the Department of Medical Informatics & Clinical Epidemiology (DMICE). The department has recently been renamed the Department of Informatics and Clinical Epidemiology (DICE). I suspect someone simply wanted to be able to say “the dice are cast.” I also have a strong suspicion that someone might be my mentor.
My doctoral research focused on applying artificial intelligence and machine learning to clinical decision support in trauma care.
Feasibility of Using Machine Learning for Clinical Decision Support to Optimize Transfusion Practices in Trauma Care
My dissertation explored three main themes:
Understanding clinical decision-making.
The first part focused on identifying and evaluating what information
trauma surgeons use when deciding that a patient requires blood
transfusion. This involved developing a framework for analyzing the
clinical signals and contextual factors that influence transfusion
decisions.
Predictive modeling for massive
transfusion.
The second part involved building a series of machine learning models
designed to identify patients likely to require massive transfusion. One
of the more interesting components was a feature-level fusion model
inspired by methods used in particle physics research. I also applied
several data science techniques to address the challenges of highly
imbalanced datasets, including synthetic minority oversampling and X-ray
image augmentation.
System-level forecasting of blood demand.
The final part examined transfusion needs from the perspective of
hospital systems and blood bank operations. Using time-series
approaches, I explored whether variables such as trauma volume, day of
the week, weather patterns, duration of daylight, and even lunar phases
could help predict blood demand. Because trauma mechanisms are
heterogeneous and difficult to anticipate, these predictions were only
moderately successful for transfusion needs, though patient volume
forecasting performed considerably better.
My broader research interests include:
Outside of clinical practice and research, I am drawn to things that combine curiosity, creativity, and storytelling.
Thanks for stopping by.
You can explore my Publications, Lectures, and Projects for a deeper look at my work.