Young Adult Tobacco and E-cigarette Use Transitions: Examining Stability Using Multistate Modeling

Introduction: The objective of this study was to describe tobacco and nicotine product use state transition probabilities among youth and young adults over time. Methods: A national sample of young adult tobacco product users and nonusers between the ages of 18 and 34 years at baseline was surveyed at 6-month intervals for 3 years. Use and nonuse states were defined as mutually exclusive categories based on self-reported, past 30-day use of the various products. Never use, noncurrent use, and current use of combustible, noncombustible tobacco, and electronic cigarette (e-cigarette) products was assessed at each interval. A multistate model was fit to estimate transition probabilities between states and length of stay within each state. Results: After 6 months, same-state transition probabilities were high for all use states (0.76-0.96), except for dual product use (0.48). After 3 years, transition probabilities were smaller and tended to converge toward combustible product use for baseline e-cigarette (0.42), combustible (0.51), and dual product users (0.52). Age was inversely associated with transition risk from never or noncurrent use to use of combustible or e-cigarette products. Conclusions: Never and noncurrent users, followed by combustible product users, were most likely to remain in those states throughout the 3-year observation interval. Users of any tobacco or e-cigarette product at baseline were most likely to transition to combustible product use or noncurrent use by the final follow-up. Implications: This study describes the probability of transitioning between various states of tobacco product use, including never and no current use, over a span of 3 years in a sample of young adults. This type of longitudinal description, which includes all tobacco product use states, is lacking in most studies that tend to focus on one or only a few products. The results suggest that it is important to assess outcomes over a sufficiently long period to capture true variability in patterns of product use.