ALOS Project Overview
| Background |
Length of stay (LOS) in hospital for inpatient treatment is a measure of crucial recovery time and is often used as a measure of hospital performance and a proxy of hospital resource consumption.
LOS is an indicator that will help JHS predict the LOS of patient based on history or diagnosis for previous emergency or hospital admissions. JHS has identified factors affecting LOS. Based on these factors it is crucial for JHS to plan their management as consumption of hospital resources such as bed occupancy rates might not be sufficient or given to patients who need it more. By predicting the LOS, they will be able to react earlier when faced with such a problem, which will improve the health care policies and health services. Another crucial factor for JHS in analysing LOS is for JHS to perform capacity planning for the new Ng Teng Fong General Hospital slated to be in operations in 2015.
On the other hand, from the patient’s point of view, LOS might be a variable to determine the quality of life. This is particularly an issue that affects the family members; if care is not available this will result in longer LOS and in some instances patient might be referred to intermediate and long-term care center (ILTS). Besides, this gives an opportunity for cooperation between JHS and family members to resources in the community to better provide care for patient not only in terms of medical support but also support after they have been discharged through the integrated care pathway (ICP) programme.
| Objective |
- Business objective: To identify the contributing and confounding factors that affects LOS and give recommendations to better help in hospital manage its resources and improve its treatment for patient.
- Technical objective: To use data analytics techniques such like exploratory data analysis (EDA), and statistical methods to study and gain insights from the data to identify patterns that aid business objective. We will explore the use of data mining if there is time.
| Scope |
- Perform data cleaning on the data set received to consolidate the important fields that are required for analysis.
- Perform EDA to identify patterns that will help in the study of LOS.
- Explore the various methods (mixed model, decision trees, simple regression analysis) to improve the prediction for LOS.
