In this thesis research, my aims are: (1) to investigate the associations linking depression with ever-heroin use (EHU), (2) to estimate depression’s association with past-year heroin use (PYHU); and (3) to identify an association-optimizing cut-off score for a positive depression screen based on the Patient Health Questionnaire (PHQ-9).The population under study consists of United States adult community residents, as sampled for the National Health and Nutrition Examination Survey (NHANES), 2005-2014, according to institutional review board-approved protocols. The analysis sample included 16,326 20-to-59-year-olds and estimated odds ratios (OR) and receiver operating characteristic (ROC) curves, with SAS Version 9.4 software. As compared with never-users, EHU and PYHU are more likely to screen depression-positive (OR=3.30, 95% CI=2.34, 4.65; OR=4.21, 95% CI=1.89, 9.35, respectively). ROC estimates an association-optimizing cut-off PHQ-9 score of 12 versus the standard cut-off of 10.Subject to limitations such as the self-report interview, the main estimates lead to rejection of a null hypothesis about the heroin-depression association, and help substantiate a need for more research, such as a prospective/longitudinal approach. The standard depression score cut-off might need adaptation in future studies of heroin-depression associations.One interpretation of this this study suggests that heroin use may influence the emergence of depression in US adults 20-59 years old. Results of this research might stimulate a more nuanced approach to understanding depression in the context of the heroin-using population.
Read
