Secondary Schizophrenia (160 page)

compromising their commitment to best outcomes.

lenges with substance users who have psychotic symp-Even
with

effective

engagement

processes,

toms are similar to those faced by opportunistic inter-assertive outreach is often needed. Frequently,
ventions for others with negative sequelae to substance
patients are only tenuously engaged in treatment or
use. It is now well established that confrontational
have difficulty keeping appointments, especially dur-methods in those contexts are ineffective
[17].
The sening symptom exacerbations
[28].
Like other mental
sitivity of people with psychosis to interpersonal stress
health services, integrated programs for people with
[18]
presents a further reason to avoid such approaches
substance use and psychotic symptoms typically trace
in this population. Instead, motivational interview-and contact people in community settings when they
ing
[19]
is typically used. This approach empathi-miss appointments
[29].
Rapid follow-up helps to
cally elicits the patient’s own ambivalence and con-retain these people in treatment and avert relapse.

cern about their substance use and potential behavior
In common with other community mental health

change. It also addresses low self-efficacy about their
services, most integrated treatments for substance use
ability to carry out change and helps them develop
and psychosis address a wide range of needs, includ-their own goals and plans. Whereas cognitive deficits
ing housing, work, education, social skills, recreation,
in people with psychosis present significant challenges
and clinical problems. Dealing with these needs can
for this technique, a version with several short ses-consolidate motivations for abstinence, address some
sions, simplified questions, visual aids, more-frequent
triggers for consumption (e.g., dysphoria, social pres-summaries and greater rehearsal can be used
[20,

sure), and develop activities and social networks that
407

21].
This work can even occur during an acute psy-are inconsistent with substance use
[30].

Treatment – Section 5

Randomized controlled trials on substance

controlling for therapist contact, failed to find a signif-misuse and psychosis
icant effect from the intervention
[42].

Based on published data, we awarded a point for
Research into the management of substance use and
each of 10 methodological criteria (> 50% of the eligi-psychosis has typically focused on co-occurring disor-ble sample entering the study, confirmation of diagno-ders rather than on psychosis that is clearly secondary
sis by standard interview, appropriate randomization
to substance use, although (partly because of diag-procedure, baseline equivalence or statistical control,
nostic uncertainties at initial episodes) some studies
equivalence of contact time,
≤
33% loss from attri-do include people who have received a diagnosis of
tion, independent checks on protocol adherence, cor-substance-induced psychosis
[20].
However, data on
roboration of substance use reports, blind ratings, and
people with separable disorders that are often linked
intention to treat analyses). Scores rose from 2.0 in
by ties of mutual influence
[10,
31]
should have direct
1993, to an average of 7.1 in 2006. Four studies had
bearing on strategies that would be effective in sec-a score
≥
8
[37, 38, 41, 43]
, three of which were pub-ondary psychosis.

lished in 2006 or 2007.

A systematic search for randomized controlled tri-Eleven of the studies (65%)
[20, 22, 29,
32, 33,

als on psychosis and substance misuse using research
35, 36, 37, 39, 40, 44]
had at least some substance-databases, reference lists of published papers, and
use outcomes in an experimental group that were bet-personal communications with researchers resulted
ter than controls, although in almost all cases, these
in identification of 17 randomized controlled trials
results were inconsistent across substance-use targets,
(Table 33.1).
Studies that focused solely on program
were not obtained at all assessment points, or were not
engagement or forensic outcomes were excluded, as
obtained on intention-to-treat analyses.

were quasi-experimental and within-subject designs,
One problem with detection of treatment appears
and those having only a small proportion with psy-to be that improvements may only be seen on some
chosis. Nine studies (53%) were conducted in the
substances, symptoms, or functional domains, and
United States, six in Australia, and one
[32]
was from
are often unstable over time. Setbacks may occur,
the United Kingdom. Most studies (11, 65%) recruited
even when the overall trajectory is positive. Existing
outpatient participants (including two on homeless
research may often underestimate the true impact of
people
[33, 34]),
and another two studies included
treatment when it focuses on being completely sub-both outpatients and inpatients
[35, 36].
Fifteen stud-stance free, on days to first substance use, or similar
ies (88%) had mixed samples with serious mental ill-indices of ultimate success. More sensitive indices of
ness or with schizophrenia spectrum disorders: of the
transition toward better control of substance use (e.g.,
remaining two, one was on bipolar disorder
[37]
and
percentage of assessments that were substance free,
the other was on schizophrenia
[35].
Participants var-or showed reductions in substance use
[22])
may be
ied from ones experiencing their first episode
[38],

required, especially at early stages. Similarly, continu-to people with chronic and disabling disorders. Most
ous measures of symptoms or of functioning may be
studies had a majority of men (48%–97%, median
=

required to detect the full impact of interventions on
74%). Sample sizes ranged from 25 to 485, and the
these outcomes.

median sample size of 120 gives confidence that most
A further challenge for existing studies was that
studies have power exceeding 0.80 of detecting a post-over half (N
=
9, 53%)
[20, 29,
34, 37, 38, 39, 41, 45,

treatment difference between means of moderate size
46]
reported at least some substance-related improve-

(d
=
.50).

ments across the whole sample. For example, this was
As
Table 33.1
shows, a wide range of interventions
true for five of the six studies that did not find sigis represented. Contact time ranged from one session
nificantly superior substance-related results for exper-of 30–45 minutes
[39, 40]
to intensive case manage-imental groups over those found in controls
[34, 38,

ment over 3 years
[29,
41].
Overall durations of the
41, 45, 46].
These apparent improvements may reflect
studies varied from 3 months to 5 years post-baseline,
regression to the mean, since people are more likely to
with a satisfactory median duration of 12 months. No
enter treatment when their problems are worse than
standardized intervention has yet been examined in
usual, and later assessments on average will tend to
more than one published study. Furthermore, an as-yet
be better. However, it is important to recognize that
408

unpublished test of the intervention by the first author,
the control interventions in most cases were active
Chapter 33 – Nonpharmacological interventions in secondary schizophrenia

Table 33.1
Studies included in a review of randomized controlled trials for treatment of substance misuse and psychosis
Study

Design

Quality Score/10

Lehman
et al.
(1993) [99]

TAU vs. ICM
+
Gp

2

Burnam
et al.
(1995) [33]

S-I Control vs. NRes vs. Res (NResRes: P-Ed
+
S-H
+
Gp
+

4

CM
+
Rec)

Hellerstein
et al.
(1995, 2001) [45, 100]; Miner
TAU (Par) vs. Int (Gp
+
P-Ed
+
S-H)1

6

et al.
(1997) [28] [45, 100, 28]

Herman
et al.
(1997, 2000) [44, 101]

TAU vs. Int (P-Ed
+
R-Ed
+
S-H
+
Gp)
4

Drake
et al.
(1998) [29]

CM vs. ACT

7.5

Barrowclough
et al.
(2001); Haddock
et al.

TAU vs. Int (MI
+
CBT-S
+
FI)
6

(2003) [32, 58]

Baker
et al.
(2002) [39, 102]

Ad vs. MI

4.5

Hulse and Tait (2002, 2003) [40, 103]

Inf vs. MI

5

Graeber
et al.
(2003) [35]

P-Ed vs. MI

4

James
et al.
(2004) [36]

P-Ed (SUD) vs. Gp (P-Ed, MI, CBT)

4.5

Kavanagh
et al.
(2004) [20]

TAU vs. MI

6

Calsin
et al.
(2005); Morse
et al.
(2006) [34, 57]

TAU vs. ACT vs. Int ACT

6

Baker
et al.
(2006) [46]

TAU vs. MI
+
Int CBT

7

Bellack
et al.
(2006) [22]

Support
+
P-Ed vs. MI
+
CBT for SUD

5.5

Edwards
et al.
(2006) [38]

TAU2
+
P-Ed vs. MI
+
P-Ed
+
Int. CBT

8

Essock
et al.
(2006) [41]

Int CM vs. Int ACT

8

Weiss
et al.
(2007) [37]

Int Gp vs. SUD Gp

9

Notes:

TAU: Treatment as usual or routine care

Ad: Advice (may include referral to parallel service)
S-I Control: Service information only

Par: Parallel treatment for psychosis, alcohol/drug problems
Int: Integrated treatment for comorbidity

Ed: Education (P-Ed: Patient; R-Ed: Relatives/carers)
Inf: Written information

Res: Residential treatment (NRes: Nonresidential)
Wk: Work program

Gp: Group intervention

S-H: AA or other self-help groups

CM: Standard case management (ICM: Intensive)
ACT: Assertive community treatment

MI: Motivational interviewing

Inc: Incentives

CBT: Cognitive-behavior therapy (CBT-S: for psychotic symptoms)
Rec: Development of recreational activities

RI: Relatives/carers intervention

FI: Family intervention (patient and relative(s))
SUD: Substance use disorder

1. Gp was a manualized support group, with issues and skill foci modified according to individual needs. Housing, medical, prevocational,
family interventions were also offered as needed.

2. Case management, mobile assessment and treatment, family intervention, group programs, and a recovery clinic for early psychosis
was received by all participants.

treatments (e.g., parallel or sequential interventions,
sistent with observations of brief interventions for
or standard case management). Furthermore, where
substance misuse having substantial impact in the
the control treatment was a brief intervention, results
general population
[17]
, and with many people with
may also be due to some participants only requir-psychosis and substance misuse making temporary
409

ing an intervention of minimal intensity. This is con-modifications to their consumption after having a
Treatment – Section 5

single negative experience with the substance
[47].

substance assessment
[42]
. The one positive trial of
The challenge for researchers is to discover treat-brief intervention
[35]
had a relatively small sample
ments that are clearly better than these control group
size (N
=
30).

responses, which in some studies were substantial
The current data are even equivocal over the impact
[20].

of motivational interviewing on sustained engage-Effects on symptoms, relapses, or rehospitaliza-ment in treatment, with at least one study find-tions are more difficult to demonstrate, given that
ing that brief advice gave similar effects
[39].
How-each may have other contributing factors, especially
ever, motivational components may have contributed
in a mixed psychotic sample. Accordingly, only six
to the impact of combination treatments in Table