Depression and Substance Use: A Temporal-Ordered Model.

Background and Objectives: This study sought empirical evidence for the self-medication hypothesis. It asked whether diagnosis of depression, together with the use of mental-health care, leads to substance use. Methods: Data came from the 5-wave, longitudinal Fragile Families and Child Wellbeing (FFCW) Study, an investigation of a cohort of nearly 5,000 births, which oversampled non-marital births. FFCW examined newborns' biological mothers and fathers, all of whom lived in the United States. The adults were initially interviewed between 1998 and 2000 (the period of the cohort's births); they were then re-interviewed for four times over 10years. FFCW measured mental health-related variables, level of drug use, and social structural and demographic factors, all of which are also measured by the present study. Respondents in our study's two final subsamples had responded to every FFCW interview item we would use in constructing our final model's variables. Our subsample of mothers totaled 3,477 women, from whom 5,987 person-waves were derived. Our subsample of fathers totaled 2,096 men, yielding 3,543 person-waves. Results: We used STATA generalized estimating equations for panel data, and found some evidence to support the self-medication hypothesis. In addition, our results indicated that the relationship of substance use to the use of mental-health care was gender-specific. Conclusions/Importance: Empirical evidence from this study only partially supported the self-medication hypothesis; the co-occurrence of depression and substance abuse in our sample was also reflective of social structural and demographic variables and of prior substance use variables.