Study of Healthcare Data Related to Hospital Readmission of Patients with Cardio Vascular Diseases Using Machine Learning Tools

Abstract

Hospital readmissions are indicators of the quality
of service offered by hospitals and give an insight into
the performance measures on the cost at the hospital The
Nationwide Readmissions Database (NRD) is a database
developed for the Healthcare Cost and Utilization Project
(HCUP). For this research, the data for the year 2016 and
2017 from the National Readmission Database (NRD) will
be studied and machine learning models built to model
the relationship between readmission and various factors
related to the patient. The observations for Ischemic heart
diseases and diseases of pulmonary circulation (diagnosis
codes I20 to I28) will be used for this study. Analysis
and goodness of model indexes such as the confusion
matrix, Area Under the Curve(AUC) index, Mean Square
Error(MSE), and R squared scores and findings from
the study will also be evaluated and reported taking into
account the model parameters.