As is well known, data analysis is frequently applied across a variety of fields. Both hospital management and patients benefit from the use of data analysis in the healthcare industry. The article discusses how data analysis may be used in hospitals and how it can help them predict demand and patient needs to find estimated costs. Here is just a small example of it.
Length of Stay is one of the factors that influence the hospital performance as well as it helps hospital organization for better resource utilization with anticipating the demand as well as insurance companies can know about the patients stay if they claim for it. Also predicting total expenses of patient stay will benefit patients to plan beforehand as well.
The overall objective of the experiment is to conduct an analysis and develop a more accurate model for multiple diseases and with multiple hospitals to predict length of stay and also total expenses for the stay of the patient. The use of regression models was done to compare different machine learning models. There are four steps in the methodology: Data collection and preparation, model building, implementation and evaluation. Below is the figure of the methodology followed during this research.
While performing data analysis, few fields were chosen to check the relationship with length of stay. Below we can see one of them. It is the relationship with a diagnosis related group.
After analysis, the model development phase comes where different regression models are used to compare the data after performing other data processing. We needed to perform some hyperparameter tuning to yield the best result when comparing different machine learning models. Out of all the models compared, XGBoost Regression yielded the best result for Length of Stay which was the primary output variable and used the same model for predicting total charges for patients as well. For comparison of models the metric of MSE and RMSE was chosen. Out of all the variables the most impactful predictor for both Length of Stay and Total Charges was CCS Procedure according to our study. Implementation of the system was done through Flask and deployed on the web. The trained model is used for implementation purposes. The system evaluation is done with giving value to test the model whether it gives the desired output or not for predicting length of stay and the total expenses both. Main limitation of study is the dataset does not contain date of discharge or date of admission which would have made the research more insightful with greater analysis.
Al Taleb, A. R., Hoque, M., Hasanat, A., & Khan, M. B. (2017). Application of data mining techniques to predict length of stay of stroke patients. 2017 International Conference on Informatics, Health and Technology, ICIHT 2017.
Andreu-Perez, J., Poon, C. C. Y., Merrifield, R. D., Wong, S. T. C., & Yang, G. Z. (2015). Big Data for Health. IEEE Journal of Biomedical and Health Informatics, 19(4), 1193–1208.
Anitha, S., & Sridevi, N. (2019). HEART DISEASE PREDICTION USING DATA MINING TECHNIQUES. Journal of Analysis and Computation.
Baek, H., Cho, M., Kim, S., Hwang, H., Song, M., & Yoo, S. (2018). Analysis of length of hospital stay using electronic health records: A statistical and data mining approach. PLOS ONE, 13(4), e0195901.
Carter, E. M., & Potts, H. W. (2014). Predicting length of stay from an electronic patient record system: A primary total knee replacement example. BMC Medical Informatics and Decision Making, 14(1), 1–13.
Chen, Y., Patel, M. B., McNaughton, C. D., & Malin, B. A. (2018). Interaction patterns of trauma providers are associated with length of stay. Journal of the American Medical Informatics Association, 25(7), 790–799.
Zolbanin, H. M., Davazdahemami, B., Delen, D., & Zadeh, A. H. (2020). Data analytics for the sustainable use of resources in hospitals: Predicting the length of stay for patients with chronic diseases. Information & Management, 103282.
Mr Aryal holds a Masters of Engineering from Asian Institute of Technology, Bangkok, Thailand, and is currently a faculty member at KIST College of Information Technology, Kamalpokhari, Kathmandu, Nepal.