International Research Journal of Education and Technology

In the healthcare industry, predicting patient outcomes is essential for early intervention, improving treatment plans, and optimizing hospital resource management. This project focuses on leveraging Snowflake’s built-in capabilities to perform end-to-end predictive modeling for patient outcomes using historical medical data, lab results, and treatment histories. The solution utilizes Snowflake SQL functions for data preprocessing, including handling missing values, normalizing patient data, and aggregating time-series health records. Machine learning models are implemented directly within Snowflake using Snowflake ML functions and Python UDFs to predict patient risks based on past diagnoses, medication history, and vital sign trends. The model results are stored and queried efficiently using Snowflake’s optimized storage and indexing techniques, ensuring real-time insights for healthcare professionals. This approach enables hospitals and research institutions to identify high-risk patients, optimize care pathways, and improve overall healthcare outcomes—all within the Snowflake ecosystem.