Value-Driven Healthcare: Cost-Benefit ML Approach to AKI Management in Cardiac Surgery

Acute Kidney Injury (AKI) following cardiac surgery is a significant complication that impacts patient outcomes and healthcare costs. This study introduces a cost-sensitive predictive model integrating Random Forest (RF) and eXtreme Gradient Boosting (XGB) algorithms to enhance the identification of AKI risk. Our model achieved a recall of 96%, demonstrating its high sensitivity in identifying at-risk patients, which is critical for minimizing missed diagnoses and improving early intervention strategies. By incorporating cost considerations into the machine learning framework, the model ensures a balance between clinical and economic outcomes, leading to an estimated net savings of $8,101,676 through optimized resource allocation and reduced complications. This work highlights the potential of integrating cost-sensitive methodologies with predictive modeling to promote value-driven healthcare and improve decision-making in clinical settings.