Abstract 4370827: Artificial Intelligence Intervention versus Standard Care in Cardiovascular Disease Outcomes: A Rapid Systematic Review

Abstract

Introduction: Cardiovascular disease (CVD) continues to be one of the leading causes of death worldwide. Early detection is crucial for identifying effective ways to improve CVD outcomes. Artificial Intelligence (AI) tools can support clinicians in more effective management of CVD and improved patient outcomes. Hypothesis: We hypothesize that Al tools can contribute to improving CVD outcomes, specifically reducing CVD events and CVD mortality. Methods: We conducted a systematic review to evaluate the effectiveness of Al-supported interventions in CVD management compared to standard clinical practice. A literature search was performed across multiple databases to identify studies that analyze AI detection in CVD. Studies that met the following criteria were included: (1) adult participants (≥ age 18) with diagnosed CVD, (2) Al-driven interventions for CVD detection, monitoring, or management, (3) control groups receiving standard clinical care, and (4) reported outcomes including blood pressure control, myocardial infarction, stroke, or mortality. Data extraction focused on clinical effectiveness and process improvements, which were synthesized using descriptive techniques. Results: Thirteen studies were included (4 randomized controlled trials (RCTs), 3 cluster-RCTs, and 6 observational studies) including up to 22,641 participants across intensive care, community, and remote settings (Tables 1&2). Overall, 11 out of 13 studies (85% reported improvement in cardiovascular outcomes. Reported mortality reduction ranged from 0.8% to 12%, with an odds ratio of 0.39-0.56 for heart failure and sepsis mortality. Major cardiovascular event reductions ranged from 4% to 12% for myocardial infarction, stroke, heart failure, and cardiac death. For other outcomes, improvements included a decrease in blood pressure ranging from 2.3 to 10.1 mmg across 3 studies, an 86.7-minute reduction in door-to-treatment times, and a 40.5% improvement in medication adherence, and an 8.7% improvement in lipoprotein cholesterol (LDL) control. Conclusion: Al-guided interventions consistently improved cardiovascular outcomes, with the strongest evidence for machine learning algorithms and clinical decision support systems. These findings support integrating Al tools into routine cardiovascular care for risk factor management, mortality reduction, and process optimization. Future research should address long-term effectiveness and implementation, especially in populations who are most impacted by CVD.

Ähnliche Arbeiten