Unveiling Insights From Complexity: Advanced Computational Techniques for High-Dimensional Medical Data

Healthcare generates vast amounts of data daily, from genetic profiles to hospital records, but much of it remains untapped due to its complexity. This dissertation develops new computational tools to unlock this data’s potential, aiming to improve patient care and medical research. Five projects tackle different challenges: Project 1 creates Deep MAGIC, a method to fill in missing genetic and image data accurately, vital for understanding diseases like cancer. Project 2 analyzes how the COVID-19 pandemic disrupted surgeries, finding a 27% drop and temporary complication rises in 2020, guiding future crisis planning. Projects 3 and 4 study kidney disease trials, confirming reliable shortcuts to test treatments faster and cheaper. Project 5 introduces SIREN, predicting health outcomes with limited data, useful for early disease detection. Together, these tools turn raw data into practical solutions, paving the way for personalized medicine and better healthcare decisions.