View Proposal


Proposer
Soo Huei Ching
Title
Predicting Inpatient Admissions: A Comparative Study of Poisson Regression and Machine Learning Approaches
Goal
To analyse how insurance status and socioeconomic characteristics influence non-critical inpatient admissions, and to compare the predictive performance of Poisson regression with machine learning models.
Description
This project aims to statistically and predictively examine the daily rate of non-critical hospital inpatient admissions. The study will focus on how socioeconomic characteristics and insurance status affect admission patterns, and will compare the effectiveness of Poisson regression with selected machine learning models for predictive accuracy. The analysis will use data from the Medical Expenditure Panel Survey (MEPS), specifically the Hospital Inpatient Stay Data and individual-level socioeconomic and insurance data. By merging these sources, the study will create a dataset that captures both demographic and contextual information. The outcomes of this project will provide insights into disparities in hospital utilisation across socioeconomic and insurance groups. Predictive insights generated from the models could help healthcare planners, governments, and hospital administrators in forecasting inpatient demand, reducing unnecessary admissions, and improving resource allocation.
Resources
1. Hunter, A. E. L., Spatz, E. S., Bernstein, S. L., & Rosenthal, M. S. (2015). Factors influencing hospital admission of non-critically ill patients presenting to the emergency department: A cross-sectional study. Journal of General Internal Medicine, 31(1), 37–44. https://doi.org/10.1007/s11606-015-3438-8 2. Kansagara, D., Englander, H., Salanitro, A., Kagen, D., Theobald, C., Freeman, M., & Kripalani, S. (2011). Risk prediction models for hospital readmission: A systematic review. JAMA, 306(15), 1688–1698. https://doi.org/10.1001/jama.2011.1515 Data Source: Medical Expenditure Panel Survey (MEPS) files: o Hospital Inpatient Stay Data: https://meps.ahrq.gov/mepsweb/data_stats/download_data_files_detail.jsp?cboPufNumber=HC-229D o Individual-level Data: https://meps.ahrq.gov/mepsweb/data_stats/download_data_files_detail.jsp?cboPufNumber=HC-233
Background
Url
Difficulty Level
Moderate
Ethical Approval
None
Number Of Students
1
Supervisor
Soo Huei Ching
Keywords
poisson regression, machine learning, healthcare analytics, socioeconomic factors
Degrees
Bachelor of Science in Statistical Data Science