Abstract:
|
A Markov model to predict transitions between tobacco product use categories was developed from a cohort of 1210 adult tobacco users in urban Franklin County Ohio and six rural Appalachian Ohio counties. These participants completed a baseline in-person interview as well as follow-up interviews at 6, 12, and 18 months. At each interview the types of tobacco products used was documented and participants were classified into one of 5 tobacco use categories: exclusive combustible, exclusive smokeless, exclusive e-cigarette, dual (users of at least 2 of the previous 3 categories), and occasional or no use. Separately within the urban and rural cohorts, the data from the baseline, 6, and 12 month interviews was used to estimate the 6 month transition probabilities between the 5 tobacco use categories. At 18 months the predicted frequencies in each of the categories are close to those observed. Further, significant differences were found between the transition probability matrices associated with the urban and rural cohorts. Within each cohort the long-run distributions are found which can be used to inform tobacco policy.
|