This R01 grant application asks for a 20 month grant award for new analyses of survey data gathered in the latest National Household Surveys on Drug Abuse (NHSDA), i.e., those completed since 1988. These new analyses will build from prior NIDA research in order to shed light on suspected determinants of non-medical drug use, including not only the characteristics of local areas, but also separately, the characteristics of individuals that might influence local area clustering of non-medical drug use. The NHSDA survey data span a range of suspected determinants of non- medical drug use from the individual level (e.g., age, sex, educational attainment) to the community level (e.g., U.S. Census data on census tracts or other local area segments from which the survey respondents were sampled, such as percent of occupied housing units that are owner- occupied). There are possibilities also for integrating other community- level characteristics (e.g., local area crime rates and indices of drug availability), as discussed in this proposal. Whereas prior investigations have studied neighborhood, family, and personal characteristics that might influence prevalence or risk of non- medical drug use, this research project will differ in its use of a new statistical method called """"""""alternating logistic regression"""""""" and other innovative strategies that have demonstrable value in the study of the multi-level determinants of drug-taking behavior. The analyses also will attend to the complex survey design of the NHSDA, with its multi-stage area probability sampling, and differential probabilities of selection across and within household and local area segments, which might affect inferences, if ignored. The public health significance of this relatively low-cost data analytic project will be seen not only in new findings that concern the epidemiology of non-medical drug use, based on new analyses of the already gathered NHSDA data, but also in the application of recently introduced statistical procedures that can be tailored by other NIDA investigators for use in their own work. Due to this study's focus on both substantive and methodologic implications of clustering of non-medical drug use within epidemiologic samples, this new research should be helpful to prevention and intervention researchers as well as those who conduct epidemiologic surveillance activities. The proposal includes a plan for dissemination of information about the substantive and methodologic aspects of this research, via scientific papers in the peer-reviewed literature, and also via presentations at scientific meetings (e.g., Society for Prevention Research).
Showing the most recent 10 out of 22 publications
