2005 HSR&D National Meeting Abstract
3009 — Dropouts from VA SUD Treatment: Predictors of 5-Year Outcomes
Harris AH (Center for Health Care Evaluation, VAPAHCS)
McKellar JD (Center for Health Care Evaluation, VAPAHCS)
Moos RH (Center for Health Care Evaluation, VAPAHCS)
Substance use disorders (SUD) are chronic, relapsing conditions with profound medical, social, and personal consequences. Although it is generally assumed that patients who drop out have worse outcomes than patients who complete treatment, very little is known about the natural history of dropouts. This study uses multilevel modeling of individual trajectories of substance-related problems (SRP) over 5 years to increase our knowledge about both the outcomes and change in problems of SUD patients who drop out of VA treatment.
This longitudinal study tracked 388 male dropouts of VA residential substance abuse programs at baseline, 1-year, 2-year, and 5-years, with over 75% providing information on at least three occasions. Patients completed questionnaires regarding substance use and related problems, social and family characteristics, health care utilization, psychiatric co-morbidities, substance use expectancies, and coping strategies. Using multilevel modeling, baseline characteristics were used to predict individual changes in SRPs as well as SRP 5-year status.
Dropouts had somewhat higher SRP than a sample of 3047 patients who completed treatment. However, there was significant variability among dropouts in both changes in SRP from baseline and 5-year SRP. Significant baseline predictors of individual improvement in SRP over time were more depression, anxiety, substance use, and number of inpatient admissions in the previous year. Significant predictors of more 5-year SRP included demographic and social functioning indicators (less income, poor housing, more arrests, unemployed), more alcohol consumption, and substance dependency, positive substance use expectancies, negative expectancies of sobriety, not having a 12-step sponsor, and risk-taking as a coping strategy.
This study adds to what is known about the natural history VA patients who dropout of SUD treatment, as well as about baseline predictors of their individual trajectories (change and end-status). Some patients who drop out of SUD treatment do well and others do poorly. Those with greater challenges and substance use typical improve more than those with less severe profiles, but still have more SRP at 5-years.
Knowing the baseline characteristics that predict individual trajectories can help clinicians and program directors target those dropouts with poor prognoses for reengagement efforts.