Causal Machine Learning and Trial Emulation for Precision Cardiovascular Risk Stratification in Obstructive Sleep Apnea (CLARITY-OSA)

About This Grant

ABSTRACT This proposed training grant is structured to support Dr. Oren Cohen, MD, in becoming an independent investigator by leveraging his strong foundation in clinical sleep medicine and providing rigorous, mentored training in machine learning and artificial intelligence (ML/AI)-based clinical research. Through this training, Dr. Cohen will acquire the skills necessary to lead innovative research aimed at developing clinically actionable models to guide patient care and advance future investigations into the cardiovascular effects of obstructive sleep apnea (OSA) treatment. Through this proposal, Dr. Cohen will receive multidisciplinary training to investigate treatment response heterogeneity within the OSA population and identify both (1) individualized treatment effect scores and (2) eligibility criteria to inform future randomized controlled trials (RCTs), using causal inference-based ML methods. The research plan builds on Dr. Cohen’s pilot work, funded by the Stony Wold- Herbert Fund and American Thoracic Society ASPIRE fellowship, and aims to determine whether the null results observed in prior RCTs of continuous positive airway pressure (CPAP) therapy reflect a uniform treatment response or mask important heterogeneity that can be revealed through causal ML. Two parallel approaches will be employed: (1) causal survival forests to generate individualized treatment effect scores that prioritize patients on a spectrum from most to least likely to benefit from CPAP for reduction of composite cardiovascular disease (CVD) outcomes and (2) in silico trial emulation to identify inclusion and exclusion criteria that define patients most likely to benefit from CPAP, as well as the outcomes most responsive to treatment, to inform and streamline future RCT design. Upon successful completion, this work will provide clinical and research tools to support precision medicine in OSA and lay the foundation for future R01-level proposals. These tools will directly inform patient-level CPAP decisions and enable more efficient trial design. Dr. Cohen’s career development plan will take place in the rich research environment of the Icahn School of Medicine at Mount Sinai (ISMMS), supported by a mentoring team with internationally recognized expertise: Dr. Neomi Shah (OSA and CVD), Dr. Mayte Suárez-Fariñas (biostatistics and causal inference), and Dr. Girish Nadkarni (ML and AI in healthcare). His career development plan includes targeted coursework and hands-on experience in causal ML model development, real-world data curation and large language models, RCT design and trial emulation, and scientific leadership and grant writing. With strong mentorship and a solid foundation, Dr. Cohen will become an independent investigator and leader in the use of causal ML/AI to advance precision medicine in OSA and cardiovascular health.