Leveraging Causal Machine Learning Methods to Enhance Tobacco Control Interventions

/ABSTRACT During the past decade, the tobacco product landscape has evolved rapidly with a remarkable decrease in cigarette smoking prevalence, a dramatic increase in the popularity of electronic cigarettes (e-cigarettes), and the emergence of other novel tobacco products such as heated tobacco products. E-cigarettes have been the most used tobacco products among US middle and high school students for several years. In response to this trend, the Surgeon General declared an epidemic of e-cigarette use among youth in 2018. The escalating use of e-cigarettes, particularly among adolescents and young adults, raises significant concerns about their nicotine exposure and potential health harm. Therefore, it is essential to understand the complex interplay of individual modifiable risk factors and the initiation of e-cigarette use to craft effective preventive strategies for reducing e-cigarette use among these young demographics. Survey data has traditionally been the cornerstone of tobacco regulatory research, offering valuable insights into behavior patterns, monitoring changes in tobacco usage, and evaluating the impact of regulatory measures. Yet, as more comprehensive tobacco-related datasets become available, innovative methods are needed to analyze this wealth of information and to extract deeper behavioral insights, forecast trends, and investigate the causes of these trends. Despite the proven value of causal machine learning in various fields, their use in tobacco control has been limited. Therefore, I plan to integrate this advanced approach with survey data to enhance the understanding and prediction of tobacco use behaviors and aid in designing optimal tobacco interventions. I will focus on understanding the uptake of e-cigarette use and studying causal pathways leading to this behavior among tobacco-naïve adolescents and tobacco-naïve young adults. To achieve this, I will need further training in machine learning, statistics, and youth e-cigarette use. As such, I will take a series of formal courses to fill in my knowledge gaps. In addition, I will work closely with my mentors and collaborators whose areas of expertise will help me to realize my training and research goals. This K01 proposal will provide me with the protected time to acquire the skills and training necessary to become a leading researcher specializing in the application of causal machine learning to address tobacco-related issues and develop policy assessment tools. Project Number: 1K01CA308939-01 | Fiscal Year: 2026 | NIH Institute/Center: National Cancer Institute (NCI) | Principal Investigator: Thi Thien Thuy Le | Institution: UNIVERSITY OF MICHIGAN AT ANN ARBOR, ANN ARBOR, MI | Award Amount: $158,895 | Activity Code: K01 | Study Section: Special Emphasis Panel[ZRG1 CDPT-P (56)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11282631