University of Minnesota
Transportability Methods for Estimating the Effects of a Low Nicotine Standard on Smoking Behavior
Abstract
dc:description.abstractIn January of 2025, the FDA announced a proposed rule which would cap the amount of nicotine within cigarettes to approximately 2.5% of the nicotine content of cigarettes currently sold on the market. Randomized trials have consistently shown that participants randomized to receive very low nicotine content (VLNC) cigarettes (0.4 mg/g) have significantly reduced cigarette consumption, nicotine dependence, and biomarkers of nicotine and toxicant exposure. However, these trials may not be representative of the general population of people who smoke (PWS) in the U.S. or other populations of PWS that may be of interest. Methods for generalizability and transportability allow researchers to use baseline covariate data to transfer inferences from a source sample to a target sample from a population of interest. However, since nicotine is highly addictive and NNC cigarettes are widely commercially available, not all trial participants assigned to the VLNC arm adhered to their assigned treatment. Standard methods for generalizability and transportability target the ITT effect in the target population; however, we are instead interested in the effect under full adherence in the target population, as this more closely reflects the amount of nicotine in cigarettes being set by regulation. Although standard transportability methods can easily be extended to account for non-adherence if adherence behavior is known, self-reported adherence data within the RCTs has been found to be unreliable, and so we must use biomarkers of nicotine metabolization to estimate adherence in the VLNC arm. Current methods within transportability that address adherence do not account for adherence being unknown and target the effect among those likely to adhere in the target. Current causal methods targeting the effect under full adherence when adherence is unknown and must be estimated have not yet been extended to a transportability setting. We propose transportability methods that account for unknown non-adherence within the source trials, estimate adherence behavior using biomarkers of nicotine exposure, and estimate the effect if all participants in the target population were compelled by regulation to adhere. First, we develop an inverse odds weighting approach which we use to assess the impacts of a nicotine standard on the average number of cigarettes smoked per day (CPD) in U.S. PWS. Then, we propose a weighted least squares approach for transporting effects when individual participant data (IPD) are not available in the target population and estimate effects on CPD in PWS in eight European countries. Finally, we apply our inverse weighting method to additional measures of smoking behavior among U.S. PWS.
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Silva, Milena
Subjects
dc:subject × 5Rights
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Dc Identifier Citation
- Silva, Milena. (2026). Transportability Methods for Estimating the Effects of a Low Nicotine Standard on Smoking Behavior. Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/283453.