Mission impossible: treating serious mental illness and substance use co-occurring disorder with integrated treatment: a meta-analysis

Abstract

Integrated treatment of co-occurring substance use and mental health disorders is defined by the concurrent delivery of treatment for both disorders by a team of providers. This approach is designed to increase effectiveness by minimizing obstacles experienced by individuals who are seeking care for co-occurring disorders. This meta-analysis analyzed the results of several studies conducted in the United States to ascertain the effectiveness of integrated treatment on the discrete outcomes of drug use, alcohol use and psychiatric symptoms and functioning. This meta-analysis also examined the effect of integrated treatment in different settings (i.e. residential vs. outpatient). The meta-analysis systematically examined 13 studies (n = 2824) with similar methodologies and combined their results mathematically to ascertain a more precise assessment of treatment effect. The results indicated that integrated treatment created modest, non-statistically significant improvements in psychiatric outcomes and alcohol use when compared to treatment as usual. In contrast, the comparison treatment produced a significant effect on the reduction of drug use. Further examination of the effectiveness of integrated treatment in outpatient versus residential treatment settings revealed that the effectiveness of integrated care varies by setting. Limitations of the analysis are discussed along with suggestions for future research. The most significant contribution made by this analysis is the discovery of the variability of impact made by integrated care based on treatment setting (outpatient vs. residential). Because of the differences in treatment costs between residential and outpatient settings, these results have important policy implications.

Keywords:

• integrated treatment
• co-occurring disorder
• mental health
• drug use
• alcohol use
• dual disorder
• serious mental illness

Notes

 1. In adapting the instrument, modest revision was made. Item no. 15 was modified to read, Description of sample pre-treatment drinking [added and drug use] behavior, to facilitate the evaluation of studies of drug use and dependence.

 2. About 81% (n = 2275) of the total sample represented individuals from controlled studies, and 19% (n = 559) from studies without controls.

 3. This method was suggested by Dr. Scott Morris of the Illinois Institute of Technology in a personal email correspondence. It is justified by the properties of the variance, where var (Treatment ES - Overall Estimate of Control ES) = Var (Treatment ES) + Var (Overall Estimate of Control ES).

 4. Hypothesis Testing cannot be done through t-test.