Page 35 - The-5th-MCAIT2021-eProceeding

excluded from the model. Table 2 shows that only the primary team handling the patients remained as major
determinants for duration of stay at the ICU (p < 0.01). Given the same age, LOS varied by different primary
teams.
From the regression analysis in this study, it was found that only the type of primary team handling the
patients is the determining factors for LOS. We only managed to collect six main variables (age, gender, race,
primary team, LOS and diagnosis). In this study, LOS at the ICU was 6.54 days (SD: 7.061 days), which was
acceptable although this was a slightly longer length of ICU stay when compared with most other studies. The
regression model lacks of many important variables. The GICU database lack of specific diagnosis as it is only
stated as diagnosis remark. There were also missing information regarding the reasons of patients requiring the
ICU, such as medical emergency, elective surgical, or emergency surgical cases. If the diagnosis remark is
transferred to diagnosis code and reasons of patients requiring the ICU are available, it would be beneficial for
better regression analysis.

4. Conclusion

This study has described the descriptive statistics of GICU patients’ data who were warded in the HCTM. It
was found that GICU patients were staying in the hospital at an average of 6 days across all age. However,
patients of younger age (less than 12 years) require longer ICU stay. Multivariate regression model shows only
the category of the primary team handling the patients is the determinant for LOS. For further work, the
diagnosis remarks in the database can be transformed into a code forms to allow better LOS prediction model
with high performance rate. The information discovered in this study is important as LOS is commonly used as
the proxy indicator for hospital efficiency.

Acknowledgement

The authors would like to thank Ministry of Higher Education Malaysia for funding the study through the
research grant TRGS/1/2019/UKM/01/4/3.

References

Azari, A., Janeja, V. P., & Mohseni, A. (2012). Predicting Hospital Length of Stay (PHLOS) : AAA multi-
tiered data mining approach. Proceedings - 12th IEEE International Conference on Data Mining Workshops,
ICDMW 2012, 17–24. https://doi.org/10.1109/ICDMW.2012.69
Gruenberg, D. A., Shelton, W., Rose, S. L., Rutter, A. E., Socaris, S., & McGee, G. (2006). Factors
Influencing Length of Stay in the Intensive Care Unit. American Journal of Critical Care, 15(5), 502–509.
doi:10.4037/ajcc2006.15.5.502
Hunter, A., Johnson, L., & Coustasse, A. (2020). Reduction of intensive care unit length of stay: The case of
early mobilization. Health Care Manager, 39(3), 109–116
Luo, L., Lian, S., Feng, C., Huang, D., & Zhang, W. (2017). Data mining-based detection of rapid growth in
length of stay on COPD patients. 2017 IEEE 2nd International Conference on Big Data Analysis, ICBDA
2017, 254–258. https://doi.org/10.1109/ICBDA.2017.8078819
Panchami, V. U., & Radhika, N. (2014). A novel approach for predicting the length of hospital stay with
DBSCAN and supervised classification algorithms. 5th International Conference on the Applications of
Digital Information and Web Technologies, ICADIWT 2014, 207–212.
https://doi.org/10.1109/ICADIWT.2014.6814663
Zhang, A. H., & Liu, X. H. (2011). Clinical pathways: Effects on professional practice, patient outcomes,
length of stay and hospital costs. International Journal of Evidence-Based Healthcare, 9(2), 191–192.
https://doi.org/10.1111/j.1744-1609.2011.00223.

E- Proceedings of The 5th International Multi-Conference on Artificial Intelligence Technology (MCAIT 2021) [23]
Artificial Intelligence in the 4th Industrial Revolution