A comparison of tobacco product prevalence by different frequency of use thresholds across three US surveys

Background: With the increasing changes in tobacco use patterns, "current use" definition and the survey used may have important implications for monitoring population use trends.

Methods: Using three US surveys (2014/15 TUS-CPS, NHIS and PATH), we compared the adult (age 18+) prevalence of four product groups (cigarettes, other combustibles, smokeless tobacco, and e-cigarettes) based on three past 30-day frequency of use thresholds: 1+, 10+, and 25+ days. We also examined mutually exclusive single, dual, and polytobacco users as a percentage of total users for each product group.

Results: Regardless of threshold or product, the prevalence was higher in PATH followed by NHIS and TUS-CPS, in some cases by large percentages. The differences in cigarette and smokeless tobacco use prevalence in going from the 1+ to 10+ days and to the 25+ days threshold were minimal. Applying different frequency thresholds had the largest impact on other combustibles prevalence, with a 60% reduction with the 10+ days threshold and a 80% reduction with the 25+ days threshold, compared to the 1+ days threshold, followed by e-cigarettes with 40 and 60% reductions, respectively. The proportion of dual and polytobacco users decreased considerably when using the 10+ vs. the 1+ days threshold and polytobacco use was almost non-existent with the 25+ days threshold.

Conclusion: The estimated prevalence of each tobacco product use depends largely on the survey and frequency of use threshold adopted. The choice of survey and frequency threshold merits serious consideration when monitoring patterns of tobacco use.