Application of Machine-Learning to Assess Heterogeneity of Continuous Positive Airway Pressure (CPAP) Effect on Major Adverse Cardiovascular Events in the ISAACC Trial (Impact of Sleep Apnea in the Evolution of Acute Coronary Syndrome)

2021 Strategic Research Grant

Neomi Shah, MD, MS
Icahn School of Medicine at Mount Sinai

Key Project Outcomes

Our AASM Foundation-funded research investigated the impact of continuous positive airway pressure (CPAP) on cardiovascular outcomes in obstructive sleep apnea (OSA) patients. Utilizing data from the ISAACC trial and the SAVE cohort, we employed machine learning and a novel causal survival forest model. Our findings identified subgroups with varied responses to CPAP, showcasing the importance of personalized treatment. The causal survival forest model revealed significant variability in individualized treatment effects, with high ITE scores associated with an 84% decrease in recurrent cardiovascular events risk, while low ITE scores indicated a six-fold increase in risk. Our work highlights the need for personalized approaches in OSA management and contributes to advancing the understanding of CPAP’s impact on cardiovascular health. We thank the AASM Foundation for their support in conducting this impactful research.