This repo consists of a project examining the mortality effect of various cardiac drugs on septic ICU patients. All the health data was obtained from MIMIC-III, an open source database that contains ~46,000 de-identified patients health record. This project examined in particular the usage of 5 different cardiac drugs including Beta-Blocker, ACE Inhibitor, Angiotensin II Receptor Blockers, Diuretucs and Calcium Channel Blocker. The med-taken info was parsed from the past medical history section in the discharge summary. The analysis is done in two parts - the jupyter notebooks contain data extraction including med-usage, demographic info, mortality rate etc., whereas the R script performs statistical analysis such as EDA, regression and propensity score matching. Note that the propensity score matching code used the psa-workshop from MIT-HST 953 as reference. See the critical data website for more info.