Hospital-Acquired Infections: An Analytical Approach with the Integration of Statistical Machine Learning Methods

Background: Hospital-acquired infections pose a significant challenge to patient safety within the healthcare sector. These clinical conditions manifest within 48 hours of patient admission, presenting a substantial clinical burden. Mortality and morbidity indicators exhibit an alarming increase, and the associated financial implications are staggering. Consequently, healthcare organizations and insurance companies turn to machine learning and data utilization techniques to effectively manage this issue. Method: This study constitutes a systematic literature review that examines the contribution of statistical machine learning to healthcare risk management in the field of health. Specifically, it aims to explore strategies for effectively managing hospital-acquired infections using statistical science. A literature review was conducted, initially identifying a total of 59 studies. Subsequently, 27 studies were excluded, leaving 32 selected for analysis, with 23 of them providing the most relevant statistical information. The methodology used to source the literature employed the PICOS methodology (population, intervention, comparison, outcome, study design). Results: The findings reveal a multitude of machine learning methods employed by researchers. The most prominent among these include logistic regression, artificial neural networks (ANNs), k-nearest neighbors (KNNs), support vector machines (SVMs), decision trees, Bayesian networks, principal component analysis (PCA), and the random forest algorithm. Conclusion: Statistical science serves as a dependable scientific tool for ensuring the delivery of quality and safe healthcare services. From a managerial perspective, it is of paramount importance to facilitate informed decision-making processes through data-driven insights.