Psychosocial Interventions for People With Both Severe Mental Illness and Substance Misuse

Abstract

Over 50% of people with a severe mental illness also use illicit drugs and/or alcohol at hazardous levels. This review is based on the findings of 32 randomized controlled trials which assessed the effectiveness of psychosocial interventions, offered either as one-off treatments or as an integrated or nonintegrated program, to reduce substance use by people with a severe mental illness. The findings showed that there was no consistent evidence to support any one psychosocial treatment over another. Differences across trials with regard to outcome measures, sample characteristics, type of mental illness and substance used, settings, levels of adherence to treatment guidelines, and standard care all made pooling results difficult. More quality trials are required that adhere to proper randomization methods; use clinically valuable, reliable, and validated measurement scales; and clearly report data, including retention in treatment, relapse, and abstinence rates. Future trials of this quality will allow a more thorough assessment of the efficacy of psychosocial interventions for reducing substance use in this challenging population.

Background

Even low levels of substance misuse by people with a severe mental illness can have detrimental effects.

Objectives

To assess the effects of psychosocial interventions for reduction in substance use in people with a serious mental illness compared with standard care.

Search Methods

For this update (2013), the Trials Search Coordinator of the Cochrane Schizophrenia Group (CSG) searched the CSG Trials Register (July 2012), which is based on regular searches of major medical and scientific databases. The authors conducted two further searches (October 8, 2012 and January 15, 2013) of the Cochrane Database of Systematic Reviews, MEDLINE, and PsycINFO. A separate search for trials of contingency management was completed because this was an additional intervention category for this update.

Selection Criteria

We included all randomized controlled trials (RCTs) comparing psychosocial interventions for substance misuse with standard care in people with serious mental illness.

Data Collection and Analysis

We independently selected studies, extracted data and appraised study quality. For binary outcomes, we calculated standard estimates of relative risk (RR) and their 95% confidence intervals (CI) on an intention-to-treat basis. For continuous outcomes, we calculated the mean difference (MD) between groups. For all meta-analyses, we pooled data using a random-effects model. Using the grading of recommendations assessment, development, and evaluation approach, we identified 7 patient-centered outcomes and assessed the quality of evidence for these within each comparison.

Results

We included 32 trials with a total of 3165 participants. Evaluation of long-term integrated care included 4 RCTs (n = 735). We found (Table 1) no significant differences on loss to treatment (n = 603, 3 RCTs, RR 1.09, CI 0.82–1.45, low quality of evidence), death by 3 years (n = 421, 2 RCTs, RR 1.18, CI 0.39–3.57, low quality of evidence), alcohol use (not in remission at 36 months; n = 143, 1 RCT, RR 1.15, CI 0.84–1.56, low quality of evidence), substance use (n = 85, 1 RCT, RR 0.89, CI 0.63–1.25, low quality of evidence), global assessment of functioning (n = 171, 1 RCT, MD 0.7, CI 2.07 to 3.47, low quality of evidence), or general life satisfaction (n = 372, 2 RCTs, MD 0.02, higher CI 0.28–0.32, moderate quality of evidence).

Table 1.

Integrated Models of Care Compared With Treatment as Usual for Both Severe Mental Illness and Substance Misuse

Patient or population: People with both severe mental illness and substance misuse
Settings: Outpatient
Intervention: Integrated models of care
Comparison: Treatment as usual
Outcomes	Illustrative Comparative Risksa (95% CI)		Relative Effect (95% CI)	No. of Participants (Studies)	Quality of the Evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Treatment as usual	Integrated models of care
Lost to treatment Follow-up: mean 36 months	212 per 1000	231 per 1000 (174−308)	RR 1.09 (0.82−1.45)	603 (3 studies)	⊕⊕⃝⃝ low b,c	Data for 36 months only
Death Follow-up: mean 36 months	28 per 1000	33 per 1000 (11−101)	RR 1.18 (0.39−3.57)	421 (2 studies)	⊕⊕⃝⃝ low b,d
Alcohol use: Not in remission Follow-up: mean 36 months	500 per 1000	575 per 1000 (420−780)	RR 1.15 (0.84−1.56)	143 (1 study)	⊕⊕⃝⃝ low b,d
Drug (non-alcohol) use: Not in remissionFollow-up: mean 36 months	650 per 1000	578 per 1000 (409−812)	RR 0.89 (0.63−1.25)	85 (1 study)	⊕⊕⃝⃝ low b,d
Mental state Days of hospitalization	See comment	See comment	Not estimable	0 (3 studies)	Data skewed. In all trials hospitalization less in treatment group—approximately 3 days—but SD large and overlapped
Global Assessment of Functioning GAF scale of 1–100 follow-up: mean 12 months		Mean GAF in intervention group was 0.7 higher (2.07 lower to 3.47 higher)		171 (1 study)	⊕⊕⃝⃝ low b,e	Difference detected not of clinical importance
General life satisfaction Quality of Life Interview section, scale of 1−7 Follow-up: mean 12 months		Mean general satisfaction in intervention groups was 0.02 higher (0.28 lower to 0.32 higher)		372 (2 studies)	⊕⊕⊕⃝ moderate b	The scale is from 1 to 7 and the very small difference was not statistically significant and is not of clinical importance

Note: CI, confidence interval; RR, risk ratio.

^aThe basis for assumed risk is provided in the following footnotes. The corresponding risk is based on the assumed risk in the comparison group and the relative effect of the intervention.

^bRisk of bias: Rated as “serious”: Blinding of participants and personnel not possible and performance bias rated as “unclear” risk of bias. Similarly all trials were at an unclear risk of detection bias.

^cImprecision: Rated as “serious”: number of events <300 and overall sample size small.

^dImprecision: Rated as “serious”: event rate very low and 95% CI wide.

^eImprecision: Rated as “serious”: confidence interval very wide and sample size small.

For evaluation of nonintegrated intensive case management with usual treatment (4 RCTs, n = 163), we found no statistically significant difference for loss to treatment at 12 months (n = 134, 3 RCTs, RR 1.21, CI 0.73–1.99, very low quality of evidence).

Motivational interviewing plus cognitive behavioral therapy compared with usual treatment (7 RCTs, total n = 878) did not reveal any advantage for retaining participants at 12 months (n = 327, 1 RCT, RR 0.99, CI 0.62–1.59, low quality of evidence) or for death (n = 493, 3 RCTs, RR 0.72, CI 0.22–2.41, low quality of evidence), and no benefit for reducing substance use (n = 119, 1 RCT, MD 0.19, CI −0.22 to 0.6, low quality of evidence), relapse (n = 36, 1 RCT, RR 0.5, CI 0.24–1.04, very low quality of evidence), or global functioning (n = 445, 4 RCTs, MD 1.24, CI 1.86–4.34, very low quality of evidence).

Cognitive behavioral therapy alone compared with usual treatment (2 RCTs, n = 152) showed no significant difference for losses from treatment at 3 months (n = 152, 2 RCTs, RR 1.12, CI 0.44–2.86, low quality of evidence). No benefits were observed on measures of lessening cannabis use at 6 months (n = 47, 1 RCT, RR 1.30, CI 0.79–2.15, very low quality of evidence) or mental state (n = 105, 1 RCT, Brief Psychiatric Rating Scale MD 0.52, CI −0.78 to 1.82, low quality of evidence).

We found no advantage for motivational interviewing alone compared with usual treatment (8 RCTs, n = 509) in reducing losses to treatment at 6 months (n = 62, 1 RCT, RR 1.71, CI 0.63–4.64, very low quality of evidence) although significantly more participants in the motivational interviewing group reported for their first aftercare appointment (n = 93, 1 RCT, RR 0.69, CI 0.53–0.9). Some differences, favoring treatment, were observed in abstaining from alcohol (n = 28, 1 RCT, RR 0.36, CI 0.17–0.75, very low quality of evidence) but not other substances (n = 89, 1 RCT, RR −0.07, CI −0.56 to 0.42, very low quality of evidence), and no differences were observed in mental state (n = 30, 1 RCT, MD 0.19, CI −0.59 to 0.21, very low quality of evidence).

We found no significant differences for skills training in the numbers lost to treatment by 12 months (n = 94, 2 RCTs, RR 0.70, CI 0.44–1.1, very low quality of evidence).

We found no differences for contingency management compared with usual treatment (2 RCTs, n = 206) in numbers lost to treatment at 3 months (n = 176, 1 RCT, RR 1.65, CI 1.18–2.31, low quality of evidence), number of stimulant positive urine tests at 6 months (n = 176, 1 RCT, RR 0.83, CI 0.65–1.06, low quality of evidence), or hospitalizations (n = 176, 1 RCT, RR 0.21, CI 0.05–0.93, low quality of evidence).

We were unable to summarize all findings due to skewed data or because trials did not measure the outcome of interest. In general, evidence was rated as low or very low due to high or unclear risks of bias because of poor trial methods, or poorly reported methods, and imprecision due to small sample sizes, low event rates, and wide CI.

Authors’ Conclusions

We included 32 RCTs and found no compelling evidence to support any one psychosocial treatment over another for people to remain in treatment or to reduce substance use or improve mental state in people with serious mental illnesses. Furthermore, methodological difficulties, which hinder pooling and interpreting results, exist. Further high-quality trials, which address these concerns and improve the evidence in this important area, are required (see review¹ for full details).