Sepsis is a leading cause of death and excessive cost to patients, hospitals, and insurers. Although research studies have been conducted, and ideal protocol changes to inpatient treatment have been published, the readmission rate and costs associated with sepsis are still increasing. While sepsis can form outside of the treatment center or hospital, there are still a considerable number of cases, especially involving individuals older than 65, documenting the development of sepsis for inpatients. By working with our medical sponsor and adviser Philip A Verhoef, we were able to obtain a de-classified clinical dataset that contains a wealth of variables related to patients who have:
- Become septic while inpatient
- A long-term, non-chronic condition (asthma, allergies, inflammatory disease)
- Been admitted to the Chicago hospital system throughout 2006 to present day
Given this dataset and the cost related to sepsis, we turned our focus to how we can predict readmission of a patient who has experienced sepsis, and what these key predictors for readmission are. If we are successful in answering either of these questions using the dataset, we can then have a wider discussion with the medical community about how to prevent readmission after sepsis.