Developing a Predictive Model for Monitoring Renal Function in ICU Patients Undergoing Gabapentin Therapy: A Focus on Essential Lab Tests

The systematic and discerning application of laboratory tests is critical in the intensive care unit (ICU), especially for patients on gabapentin, due to its significant implications on healthcare expenses and patient outcomes. This research leverages the Medical Information Mart for Intensive Care IV (MIMIC-IV) dataset to develop predictive models using advanced machine learning algorithms with an objective to forecast critical lab test results for ICU patients. This includes levels of sodium, potassium, blood urea nitrogen (BUN), creatinine, and calcium, particularly focusing on those with chronic kidney disease. By integrating diverse data inputs such as patient demographics, vital signs, medication histories, and previous lab test outcomes, our models aim to predict future lab test results within predefined observational windows, thereby enhancing proactive clinical decisionmaking. Furthermore, this study delves into analyzing trends of abnormal lab test results, focusing on identifying whether patterns signify stability, improvement, or deterioration in patient health, and correlating these trends with the dosage and frequency of gabapentin administration. This methodology not only aids in early diagnosis and timely intervention, but also provides insights into the impact of the drug on patient health. This research aims to facilitate more informed clinical decision-making and enhance patient care in ICU settings.