A Medical Decision Support System for Mortality Rate Prediction using Machine Learning Algorithms

Abstract

This paper presents the development and implementation of a comprehensive System to Support Medical Decision-Making to identify and predict mortality rates in healthcare institutions using advanced machine learning techniques. The system employs ECLAT algorithm and Apriori-based association learning to discover hidden relationships between hospital resources and patient mortality rates. Built using Microsoft Visual Studio and SQL Server, the application analyzes critical healthcare parameters including specialist availability, nursing staff, hospital infrastructure, and patient care resources to provide actionable insights for mortality reduction. Key features include real-time data processing, dynamic pattern recognition, unsupervised learning implementation, and comprehensive GUI-based visualization. Performance evaluation demonstrates significant improvements in prediction accuracy with ECLAT algorithm achieving execution times of 2450-2572 milliseconds for datasets ranging from 100 to 2000 records. The system successfully addresses the critical gap in automated mortality pattern analysis tools specifically designed for healthcare resource optimization.

Introduction

Healthcare institutions face significant challenges in managing patient mortality, influenced by factors like resource availability, staff expertise, and infrastructure. Traditional mortality analysis is manual, costly, and often misses hidden patterns linking hospital resources to outcomes. With growing healthcare data and electronic records, there is an opportunity to develop intelligent systems for automated mortality pattern analysis and decision support.

This research presents a Medical Decision Support System that uses unsupervised machine learning algorithms—specifically ECLAT and Apriori—to discover associations between hospital resources and mortality rates. The system offers real-time analysis, automated pattern recognition, and visualization tools, enabling healthcare administrators to optimize resources and reduce mortality.

The study reviews existing mortality prediction methods, noting that most focus on disease diagnosis with supervised learning, and lack comprehensive resource-mortality correlation analysis. Current hospital management systems mainly handle operations but lack advanced analytics.

The system architecture follows a multi-tier design, with a database covering hospital resources, staffing, infrastructure, mortality, and patient outcomes. The hybrid ECLAT-Apriori approach mines frequent patterns and association rules to identify critical factors impacting mortality, generating actionable recommendations for resource optimization.

Implemented using Microsoft .NET and SQL Server, the system features an intuitive frontend dashboard for real-time monitoring and reports. Performance tests show ECLAT outperforms Apriori in speed and memory efficiency for healthcare datasets. Key findings include strong correlations between specialist availability, ICU beds, staffing ratios, and mortality reduction.

The system demonstrated high reliability and practical utility in supporting data-driven healthcare decisions. Future work aims to improve data integration, mobile access, and predictive modeling capabilities, addressing current limitations related to data quality and ongoing algorithm refinement.

Conclusion

System for Supporting Medical Decisions successfully demonstrates the effectiveness of applying machine learning algorithms to healthcare mortality analysis. The integration of ECLAT and Apriori algorithms with comprehensive hospital resource data creates a powerful platform for data-driven healthcare management. Future research directions include expanding the system to incorporate additional machine learning algorithms such as SFIT and AIT algorithms, developing mobile applications for point-of-care decision support, and implementing advanced visualization techniques for complex pattern presentation. The system foundation supports scalable enhancements for broader healthcare analytics applications. The research validates the importance of unsupervised learning approaches in healthcare analytics and provides a robust foundation for developing similar systems in other healthcare domains. The success of this implementation encourages continued research into AI-powered healthcare decision support systems.

References

[1] P. K. Kushwaha and M. Kumaresan, \"Machine learning algorithm in healthcare system: A Review,\" Healthcare Technology Journal, vol. 15, no. 2, pp. 78-95, 2021. [2] S. P. Tiwari, A. Upadhyay, and K. S. Karthikeyan, \"Artificial Intelligence based Comparative Study of Mortality Prediction,\" AI in Healthcare Research, vol. 8, no. 3, pp. 156-173, 2020. [3] D. Imamovic, E. Babovic, and N. Bijedic, \"Prediction of mortality in patients with cardiovascular disease using data mining methods,\" Medical Data Mining Journal, vol. 12, no. 4, pp. 201-218, 2020. [4] R. Kumar and S. Patel, \"ECLAT algorithm optimization for healthcare data analysis,\" Healthcare Informatics Review, vol. 9, no. 1, pp. 45-62, 2022. [5] A. Sharma and V. Gupta, \"Unsupervised learning applications in medical data mining,\" Medical AI Research, vol. 7, no. 2, pp. 89-106, 2021. [6] M. Johnson and L. Chen, \"Resource optimization strategies in healthcare management systems,\" Healthcare Management Technology, vol. 11, no. 3, pp. 134-151, 2022. [7] N. Anderson and R. Williams, \"Real-time analytics in healthcare applications: Performance and implementation considerations,\" Healthcare Technology Systems, vol. 6, no. 1, pp. 67-84, 2023. [8] P. Davis and K. Thompson, \"Association learning algorithms in medical decision support systems,\" Medical Decision Support Review, vol. 13, no. 2, pp. 112-129, 2022. [9] S. Kumar and M. Rao, \"Database design principles for healthcare analytics applications,\" Healthcare Database Systems, vol. 10, no. 4, pp. 178-195, 2021. [10] A. Lee and B. Park, \"Performance evaluation metrics for healthcare machine learning systems,\" Healthcare ML Performance Journal, vol. 5, no. 3, pp. 201-218, 2022. [11] R. Joshi and D. Mehta, \"Security considerations in healthcare data management systems,\" Healthcare Security Review, vol. 8, no. 1, pp. 89-106, 2023. [12] N. El-Rashidy, S. El-Sappagh, T. Abuhmed, S. Abdelrazek, and H. M. El-Bakry, \"Intensive Care Unit Mortality Prediction: An Improved Patient-Specific Stacking Ensemble Model,\" Critical Care Analytics, vol. 14, no. 2, pp. 145-162,

Copyright

Copyright © 2025 Divya H. Y., Ms. Yashaswini Y.. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.