Longitudinal predictors of depressive symptoms among low income injection drug users

A. R. Knowlton, C. A. Latkin, J. R. Schroeder, D. R. Hoover, M. Ensminger, D. D. Celentano

Research output: Contribution to journalArticlepeer-review

35 Scopus citations


Effects of changes in physical health status and drug use, and prior social support on depressive symptoms were assessed in low income injection drug users. Data are from participants (n=503) enrolled at baseline (1994-1995) who remained at one-year follow-up (79%), of whom 37% were HIV-positive and 36% female. Physical health was measured by HIV symptoms, AIDS, CD4 count and functional limitation (IADLs). One-third scored high on depressive symptoms (CES-D≥16) at one-year follow-up, representing no statistically significant change from baseline (38%). In multiple logistic regression, after controlling for baseline depression scores (OR=6.11, p<0.001) and drug use (OR=1.20, p=0.192), baseline functional limitation (OR=3.28, p<0.001) and declining functioning (OR=3.60, p<0.001) were positively, and quitting drug use was negatively, associated with depressive symptoms at follow-up. Low social support at baseline (OR=0.58, p<0.10) was marginally predictive of depressive symptoms. Depressive symptoms did not differ by gender. For HIV-positive respondents, functional limitation was predictive of depressive symptoms, but HIV illness and drug use were not. Facilitating drug treatment and preventive medical care may aid in reducing depression in this population. For HIV-positive drug users, drug treatment prior to AIDS may help reduce depressive symptoms, with potential implications for HIV service utilization and medical adherence.

Original languageEnglish (US)
Pages (from-to)549-559
Number of pages11
JournalAIDS Care - Psychological and Socio-Medical Aspects of AIDS/HIV
Issue number5
StatePublished - 2001

ASJC Scopus subject areas

  • Social Psychology
  • Health(social science)
  • Public Health, Environmental and Occupational Health


Dive into the research topics of 'Longitudinal predictors of depressive symptoms among low income injection drug users'. Together they form a unique fingerprint.

Cite this