cpedrasa.github.io


Predicting 30-day All Cause Readmission from Hospital Discharge Summary

In Module 5, I worked on different binary classifier algorithms to predict 30-day readmission for diabetic patients utilizing structured information from an electronic health record. Building on the knowledge gained from the previous project, I plan to explore natural language processing for the capstone to expand that work to optimally predict 30-day, all-cause hospital readmissions, which occur when a patient is readmitted to an inpatient hospital for any reason within 30 days of a prior inpatient discharge.


Predicting 30-day Hospital Readmisssions

Hospital readmissions are associated with unfavorable patient outcomes and high financial costs. Diabetes is one of the most frequently treated condition in US Hospitals with high readmission rate. Healthcare Regulatory Agencies are focused on 30-day readmission rates as a measure of healthcare quality and emphasize its reduction as a strategy to reduce healthcare costs while also maintaining quality. In this Module 5 Project, we tested different binary classifier algorithms to predict 30-day hospital readmissions of patients with diabetes.


Module 3 Project

Project Background

For the Data Science Module 3 project requirements, we worked with the Northwind database–a free, open-source dataset created by Microsoft containing data from a fictional company. This project involves gathering information from a real-world database and use of statistical analysis and hypothesis testing to generate analytical insights to answer the main question, “Does discount amount have a statistically significant effect on the quantity of a product in an order? If so, at what level(s) of discount?”


Racial Bias in Machine Learning Algorithm

For this blog post requirement, I have chosen to discuss the research article,


Exploratory Data Analysis

I may have already mentioned in my introduction that my background is in nursing informatics. I have worked on compliance/quality and health information exchange, so the only technical tool I was equipped with before joining this course was my SQL knowledge. It was very tempting for me to just leverage that skill set in the exploratory data analysis (EDA) phase of the project as I am not as comfortable extracting the information I need for the data analysis using pandas. I feel I really need to demonstrate understanding of the different tools I have learned in Module 1, so I thought I’d share how I could have done the EDA using the SQL tools I was familiar with.