Home

By using a data set from Norway on 110,204 patient admissions of people with the medical condition Sepsis from 2011 to 2012, I sought to understand the probability of survival based upon the factors of numbers of previous episodes of Sepsis, age and gender. The data was sourced from Kaggle, containing de-identified data on hospitalized patients. I used a logistic regression model with the variables of age (numeric variable), sex (categorical variable) and episode number (numeric) set to the outcome of the probability of survival for the i-th individual. From modeling the variables as a logistic function to the outcome, we discovered that being female can increases the odds of survival, higher age decreases the probability of survival for every year, while higher episode numbers have slightly lower chance of survival. A person’s survival probability will lay between 72.7% to 99.7%, depending on their individual variables, while the confidence interval will vary.

(Episode number refers to the number of times a patient has had Sepsis)