RESUMO
OBJECTIVE: The prevention research context includes current epidemic levels of hazards associated with extra-medical use of OxyContin® (to get high or otherwise outside prescribed boundaries) among teenagers and young adults, and a recent OxyContin® re-formulation with an intent to reduce these hazards, plus hope for possibly beneficial primary prevention impact. The aim is to create a benchmark of risk estimates for the years just prior to OxyContin® re-formulation in anticipation of potential public health benefit in future years, with a focus on teens and the youngest adults in the United States, and to compare two methods for estimating peak risk. METHOD: The data are from nationally representative probability sample surveys of 12-21 year olds, yielding estimates for incidence of extra-medical OxyContin® use. Samples are of the non-institutionalized United States population, recruited and assessed in National Surveys on Drug Use and Health (NSDUH), each year from 2004 through 2008. In aggregate, the sample includes 135,552 young people who had not used OxyContin® extra-medically prior to their year of survey assessment. RESULTS: The main outcome was the estimated population-level age-specific incidence of extra-medical OxyContin® use, 2004-2008. We found that during the 2004-2008 interval the estimated risk accelerated from age 12 years, reached a peak value in mid-adolescence at roughly five newly incident users per 1000 persons per year (95% confidence intervals, 0.3%, 0.7%), and then declined. A meta-analysis approach to year-by-year data differentiated age patterns more clearly than a pooled estimation approach. CONCLUSION: Studying young people in the United States, we have discovered that the risk of starting to use OxyContin® extra-medically rises to a peak by mid-adolescence and then declines. From a methods standpoint, the meta-analysis serves well in this context; there is no advantage to pooling survey data across years. We also discovered that during any given year a pediatrician might rarely see even one patient who has just started to use OxyContin® to get high or for other extra-medical purposes. Implications for screening are discussed.