Prediction and Prevention of Postoperative Mortality and Morbidity

This study will cover the following two specific aims: Aim 1. A pragmatic, non-randomized study to assess the effectiveness of preoperative personalized prehabilitation with proactive cognitive and behavioral interventions versus standard of care on reducing postoperative cognitive outcomes (including postoperative delirium within 30 days, postoperative cognitive decline, and dementia), MACCE, and mortality in high-risk surgical elderly patients (≥65 years). Our Electronic Health Record (EHR)-based automated machine-learning risk prediction algorithm for postoperative mortality and MACCE has been developed using \>1.25 million surgical patients' data and implemented with superior performance to comparators. This EHR algorithm will identify approximately 1,000 patients at high risk for 30-day mortality and MACCE who proceed to surgery for Aim 1. Participants will receive either standard of care (n=500) or CPC-guided personalized preoperative prehabilitation with proactive cognitive or behavioral interventions (physical exercise, cognitive training, enhanced social support, and depression support) (n=500), based on real-world clinical referral practices rather than randomization. Cognitive assessments will be performed at baseline, discharge, and at 1, 3, 6, and 12 months. Aim 2 Hypothesis: Proactive bundled intraoperative interventions are superior to standard intraoperative care in reducing postoperative cognitive outcomes, MACCE, and mortality. Expected Outcome: The refined EHR-based risk prediction algorithm will demonstrate improved accuracy for predicting MACCE, mortality, and postoperative cognitive outcomes.