AI-Based Hybrid Decision Support Model for Patient-Centric Personalized IVF Treatment Recommendation and Outcome Optimization

Abstract

In Vitro Fertilization (IVF) is one of the assisted reproductive techniques applied for couples with the inability to conceive naturally. As IVF has become rather complex, it has also become necessary for healthcare providers to provide patients with tailored recommendations with regard to every health profile. This study seeks to develop a hybrid Machine Learning (ML) based, patient-oriented Decision Support System (DSS) to enhance the treatment of patients undergoing IVF. Taking extensive data regarding patients’ history, clinical factors, and lifestyle, this study employs advanced algorithms such as Long Short-Term Memory (LSTM), Convolutional neural Network (CNN), Light Gradient Boosting Machine (LightGBM), Multi-Layer Perceptron (MLP) and hybrid CNN-LightGBM with remarkable prediction accuracies. Among the evaluated models, the hybrid CNN-LightGBM demonstrates the highest accuracy of 90.34%, precision of 89.45%, and recall of 88.23%, thereby indicating the reliability of the model in recommending IVF treatment. The proposed model improves decision-making with economically varying models of IVF treatments and advocates for an individualized, evidentiary understanding of each IVF patient’s requirements since it potentially enhances treatment outcomes. Future work includes increasing model robustness across heterogeneous populations.