The Pot Calling the Kettle Black? A Comparison of Measures of Current Tobacco Use

The "Real" Number of Washington State Adolescents Using Marijuana, and Why: A Misclassification Analysis

BACKGROUND

Although causality is difficult to establish, the regular use of marijuana has been associated with many adverse physiological and sociological consequences. There is considerable concern regarding marijuana use among adolescents, as the likelihood of adverse consequences increases significantly for this age group. The most comprehensive data for identifying risk factors for adolescent marijuana use is typically self-report, which may be over- or under-reported for a variety of reasons, including stigmatization, peer-pressure, or fear of being discovered.

OBJECTIVES

To identify the prevalence of adolescent marijuana use in Washington State, and the associated risk and protective factors, while controlling for and estimating the extent of misreporting, and its determinants.

METHOD

Data came from the 2014 Washington State Healthy Youth Survey. We accounted for missingness using chained multivariate imputation equations, resulting in 33,320 complete observations. Our model was estimated using a maximum likelihood multiple regression designed to control for systematic misclassification in binary dependent variables.

RESULTS

Approximately 12% of Washington adolescents claimed to have used marijuana in the past 30 days. Our estimates indicate this figure is likely closer to 18%. Determinants of use included use of other substances, gender, age, and measures of deviant social influences, personality/attitude, school and family bonds, bullying, and acquisition ease. Determinants of misreporting included use of other substances, gender, parental education, and family bonds.

CONCLUSIONS

Failing to control for misreporting considerably underestimates the prevalence of marijuana use among adolescents. Our model allows us to better identify at-risk adolescents and inform focused prevention efforts.

Assessing 30-day quantity-frequency of U.S. adolescent cigarette smoking as a predictor of adult smoking 14 years later

BACKGROUND

To improve measures of monthly tobacco cigarette smoking among non-daily smokers, predictive of future non-daily monthly and daily smoking.

METHODS

Data from United States National Longitudinal Study of Adolescent to Adult Health, tracking adolescents, ages 12-21, over 14 years were analyzed. At baseline, 6501 adolescents were assessed; 5114 individuals provided data at waves 1 and 4. Baseline past 30-day non-daily smokers were classified using quantity-frequency measures: cigarettes smoked/day by number of days smoked in the past 30 days.

RESULTS

Three categories of past 30-day non-daily smokers emerged using cigarettes/month (low:1-5, moderate: 6-60, high: 61+) and predicted past 30-day smoking at follow-up (low: 44.5%, moderate: 60.0%, high: 77.0%, versus 74.2% daily smokers; rτ=-0.2319, p<0.001). Two categories of non-smokers plus low, moderate and high categories of non-daily smokers made up a five-category non-daily smoking index (NDSI). High NDSI (61+ cigs/mo.) and daily smokers were equally likely to be smoking 14 years later (High NDSI OR=0.97, 95% CI=0.53-1.80 [daily as reference]). Low (1-5 cigs/mo.) and moderate (6-60 cigs/mo.) NDSI were distinctly different from high NDSI, but similar to one another (OR=0.21, 95% CI=0.15-0.29 and OR=0.22, 95% CI=0.14-0.34, respectively) when estimating future monthly smoking. Among those smoking at both waves, wave 1 non-daily smokers, overall, were less likely than wave 1 daily smokers to be smoking daily 14 years later.

CONCLUSIONS

Non-daily smokers smoking over three packs/month were as likely as daily smokers to be smoking 14-years later. Lower levels of non-daily smoking (at ages 12-21) predicted lower likelihood of future monthly smoking. In terms of surveillance and cessation interventions, high NDSI smokers might be treated similar to daily smokers.