Research Project Title:
Improving In-Hospital Mortality Prediction Algorithms: Study and Evaluation on HiRID Data
abstract:Intensive care medicine requires managing critical care patients who are often at the highest risk of deterioration and death. Current severity of illness scoring systems are often old, based upon U.S. patients, and may be limited by poor generalizability in other countries. The goal of the project is to leverage the HiRID dataset from patients admitted to the Bern University Hospital in Switzerland to evaluate and improve the Global Open Source Severity of Illness Score (GOSSIS). Particular areas of research I will study and explore to refine the GOSSIS algorithm include strategies for harmonizing data from various sources and countries, exploring various machine learning and statistical models, imputing missing data, and improving the interpretability and explainability of the model.
I am very passionate about bridging computation, healthcare, and statistics to leverage the prolific amount of available big data that models processes in the world all around us, especially within our healthcare system. I am participating in SuperUROP to further my knowledge of statistical models as well as machine learning as it applies to healthcare big data.