Estimation of Aminophylline effectiveness at Neonatal Intensive Care Unit (NICU) using Artificial Intelligence

ABSTRACT

@#Introduction: The estimation of drug competence using Artificial Intelligence is presented in various literature for the adult population, but it is still new for drug dosage optimization in neonates. Aminophylline, a methylxanthine is administered as central nervous system stimulant for reducing Apnea episodes in neonates. Methods: The paper describes comparative evaluation of Support Vector Machine (SVM), K Nearest Neighbour (KNN), Decision Tree (DT) and Artificial Neural Network (ANN) for predicting drug effectiveness of Aminophylline. The models were evaluated using 100 Aminophylline cases based on various metrics such as sensitivity, specificity, and accuracy. The data used for the analysis was collected from the population pharmacokinetic study conducted at Kasturba Medical College, Neonatal Intensive Care Unit (NICU). Results: The evaluation result seemed to favour Multi-Layer Perceptron (MLP) with accuracy of 0.92 Area Under the Curve (AUC) followed by 0.85 (AUC) for Support Vector Machine (SVM). The input parameters in particular maternal, pharmacokinetics, demographic and physiological that were identified in literature as predictor variable played an important role in estimating effectiveness of Aminophylline regimens. Conclusion: Artificial Intelligence approach was potentially helpful in analysing drug dosage of Aminophylline and its effectiveness in diagnosing neonatal Apnea.