Assessing Readmission Rates in a Sharjah Healthcare Facility

Abstract

The healthcare industry is one of the most sensitive industries as it deals with patients' health. Machine Learning techniques have been implemented to assess the performance of such industries and further improve the allocation of their resources. Many measures of performance exist to infer how a healthcare facility uses its resources. Readmission rate is a very popular rate in analyzing the performance of a healthcare facility. In this paper, we assess the readmission rate of a Sharjah healthcare facility in the first ten months of 2021. We have used classification techniques such as Logistic Regression, Random Forests, Neural Networks, and Gradient Boosting to find the best prediction model. We then used logistic regression to infer the relationships between the most important variables and the readmission rate. Results showed that the readmission rate was most influenced by the hospital departments, insurance type, marital status, age, and diastolic blood pressure. Relationships of such variables are outlined in the paper and can be further investigated to reduce readmission rates for cost reduction.