Degenhardt · Drug and alcohol dependence 2010 · cross-sectional survey analysis · n=?

Evaluating the drug use "gateway" theory using cross-national data: consistency and associations of the order of initiation of drug use among participants in the WHO World Mental Health Surveys.

Cited 274 times in the scientific literature.

Level 4 - case-series / case-control

Cross-national cross-sectional epidemiological survey analysis

PubMed 20060657 · doi:10.1016/j.drugalcdep.2009.12.001 · record verified 2026-08-26

What was done

Researchers evaluated the gateway hypothesis using data from the World Health Organization (WHO) World Mental Health Surveys conducted across 17 countries with diverse drug use prevalence. The study examined the sequence and patterns of initiation for alcohol, tobacco, cannabis, and other illicit drugs to determine whether progression to later drug use, conditional on earlier use, varied across differing background prevalence environments.

What was found

The abstract reports no numerical values, odds ratios, or effect sizes. It reports that the association between initiating gateway substances (alcohol, tobacco, cannabis) and subsequent onset of other illicit drug use varied based on background prevalence of gateway substance use. Differences in national prevalence also corresponded to differences in the likelihood of individuals reporting non-normative drug initiation sequences.

Why it matters

These findings suggest that drug initiation sequences are substantially driven by shared background causes and drug availability rather than direct pharmacological or psychological causation by gateway substances. Consequently, interventions focused solely on preventing initial gateway drug use may have limited direct impact on preventing later illicit drug use.

Limits

The abstract does not provide sample size numbers, statistical estimates, or confidence intervals. The cross-sectional survey design relies on retrospective self-reports of initiation age and sequence, which are susceptible to recall and reporting biases, and cross-national observational comparisons cannot eliminate unmeasured confounding.

Cited by