Citation:
Casstevens, W.
J.,
Cohen D., Newman F.L., & Dumaine, M.. (2006) Evaluation of a
Mentored Self-Help Intervention
for the Management of Psychotic Symptoms. International Journal
of Psychosocial
Rehabilitation. 11 (1), 37-49
Acknowledgement
The authors would like to acknowledge and thank research assistants
Constanza Bade, Ann J. Galinanes and Danielle Vandenbent for
contributing to data entry and data collection, and the Florida
International University Graduate Student Association for contributing
funding.
Contact:
W. J. Casstevens
Department of Social Work
Campus Box 7639
North Carolina State University
Raleigh, NC 27695-7639, USA
Abstract
This pilot study employs a quasi-experimental pre-post design (n = 27)
to evaluate the impact of a mentored self-help workbook (Coleman &
Smith, 1997) intervention. Participants are diagnosed with severe
and persistent mental disorders and experience medication-resistant
psychotic symptoms. The cognitive-behaviorally based workbook is
used to target improved self-management of affective and psychotic
symptoms. The intervention can be implemented in community mental
health settings by staff with less training than specialized or
licensed clinicians. Results show statistically significant
improvement on the Brief Psychiatric Rating Scale factor for Anxious
Depression. This is particularly relevant given the high levels
of depression found among individuals diagnosed with schizophrenia.
Keywords: cognitive-behavioral,
hallucinations, schizophrenia, self-help
Introduction
This pilot study explores a low-cost, mentored self-help intervention
for auditory hallucinations in a community mental health setting.
The intervention involves a mentor’s supportive assistance with written
assignments exploring individual voice-hearers’ experiences of auditory
hallucinations (Coleman & Smith, 1997; Hustig & Hafner,
1990). It focuses on assignments operationalized in a published
workbook, co-authored by a former psychiatric patient and a psychiatric
nurse (Coleman & Smith, 1997). Study participants completed
the workbook, rather than attending a specific number and duration of
treatment sessions. Participants were adults diagnosed with a
severe and persistent mood or psychotic disorder experiencing
medication-resistant psychotic symptoms. Results show
statistically significant improvement in the Brief Psychiatric Rating
Scale (BPRS) factor Anxious Depression for Intervention group
participants, over those in the Comparison group.
The “hearing voices movement” that began in Europe following a 1987
conference in Utrecht (Romme & Escher, 1996) inspired Coleman and
Smith (1997) to challenge the view of “voices” as symptoms of an
illness unrelated to a person’s history. In their workbook,
Coleman and Smith also challenged the corollary that “voice
hearers” are powerless with regards to the voices. These authors
noted that hearing voices is “not the exclusive prerogative of saints
and psychotics” (Coleman & Smith, p. 8), but rather part of the
human condition. Coleman and Smith provided no specific goals for
their workbook, but rather suggested that any action plan a voice
hearer develops should “be focused around your experiences and how you
understand them, and should work to your goals and nobody else’s [sic]”
(emphases in the original, p. 9).
Despite departures from mainstream mental health/psychiatric thinking,
seventeen of the twenty workbook exercises share similarities with
various techniques found in the cognitive behavioral treatment (CBT) of
psychosis. An established adjunctive treatment approach in Europe
and particularly the United Kingdom, CBT of psychosis is increasingly
recognized in North America (Dickerson, 2000; Kinderman & Cooke,
2000). Overall, both Coleman and Smith’s (1997) workbook and CBT
of psychosis: (1) attempt to normalize the voice-hearing experience,
and (2) view hallucinations as on a continuum with normalcy.
Following Romme and Escher’s (1989, 1996) normalizing rationale,
workbook exercises begin with an exploration of the onset of voice
hearing, and responses to the voice hearing experience, then continue
with the voice hearer’s written “Life History” (Coleman & Smith,
1997, pp. 15-18).
Four of the 20 workbook exercises closely resemble “focusing” (Bentall,
Haddock, & Slade 1994; Haddock, Bentall, & Slade, 1996).
Focusing is a CBT of psychosis strategy that examines the features,
contents, related thoughts, and attributed meaning(s) of
“voices.” The workbook, for example, includes an “I’ve just heard
voices” checklist (Coleman & Smith, 1997, pp. 19-20) to photocopy
and use daily to describe the voices, voice content and surroundings,
and the voice hearer’s associated feelings and thoughts. This
checklist also asks for the voice hearer’s explanation of the
voices. Subsequent workbook exercises have the voice hearer
address attributed meaning(s) of the voices through identification of
personal frames of reference, beliefs about the voices, and an
exploration of alternative belief systems. A difference between
the workbook’s orientation and CBT of psychosis is that the latter
views telepathy as a delusional or maladaptive explanation of voices
that should be challenged (e.g., Chadwick, Lowe, Horne, & Higson,
1994; Nelson, 1997). Workbook exercises, in contrast, neutrally
explore alternative belief systems (i.e., the illness model, the
psychological model, and telepathy).
Workbook exercises move on to emphasize coping with voices, by using
focusing and coping strategy enhancement (CSE) techniques, strategies
developed in CBT of psychosis (Tarrier et al., 1993; Yusupoff &
Tarrier, 1996). Coping strategy exercises begin with an
explanation of what “coping strategy” means and examples of different
types of coping strategies. Exercises are meant to examine
current strategies, changes desired (if any), supports, and attributed
meanings. Further exercises are meant to explore alternative
strategies and trials of new strategies. A total of thirteen of
twenty workbook exercises have similarities to CSE. The workbook
frames hearing voices as a potential adaptation to, or even survival
strategy for, life events. Guidance is provided to voice hearers
for working with professionals in this context. The net result
for voice hearers interested in working with their voices is an unusual
combination of client driven, non-judgmental exercises roughly similar
in many respects to CBT intervention strategies with psychosis.
At the time of this study’s inception, empirical results from case
study, single subject design, and small group comparison study
methodologies cautiously favored using CBT of psychosis with clients
experiencing various psychotic symptoms. Subsequent randomized
clinical trials (RCTs) have provided further support for individual CBT
of psychosis (Durham et al., 2003; Kuipers et al., 1998; NHS Centre for
Reviews and Dissemination, 2000; Rector & Beck, 2001; Sensky et
al., 2000; Tarrier, et al. 1999; Turkington, Kingdon & Turner,
2002). RCTs have utilized multiple outcome measures, including
measures of overall psychotic symptoms, negative psychotic symptoms,
depression, anxiety, self-esteem, self-concept and other
constructs. Generally, CBT of psychosis in the United Kingdom is
largely based on the work of Beck and colleagues, with overall
agreement that the principal aim “for medication-resistant psychosis is
to reduce the distress and interference with functioning caused by the
psychotic symptoms” (Garety, Fowler & Kuipers, 2000, p. 73).
It is noteworthy that the outcome measures used in CBT of psychosis
intervention research do not directly measure this intent, but rather
measure symptom reduction, in addition to a plethora of other
variables.
Summarizing outcome results with psychotic symptoms and adjunctive CBT
of psychosis, Boyle (2002) stated that “although some therapeutic
results may be modest, most statistical comparisons with other
treatments, across a range of outcome measures, favour CBT” (p.
296). Gaudiano (2005) tabulated 19 publications on 16 RCTs of CBT
for psychosis that included individual and group modalities, inpatient
and outpatient samples, first episode psychosis, recurrent psychosis,
and older patients with schizophrenia. Gaudiano cautiously
concluded that “whether commonly used therapies such as CBT are
specifically efficacious in treating psychotic symptoms” needs to be
further researched, although evidence is clear that psychosocial
interventions generally can contribute “significantly to the well-being
of individuals suffering from psychosis beyond the effects of routine
care” (p. 46). Mueser and Noordsy (2005), following Gaudiano’s
review, nevertheless concluded that although “the mechanisms and
specificity of CBT for psychosis remain unknown, the evidence amassed
supports its effectiveness” (p. 68).
Coleman and Smith’s (1997) workbook is intended for use with the
support of a trusted other (e.g., friend, significant other, family
member, and/or mental health professional), referred to in this study
as a “mentor.” With the workbook, power and authority are to
reside solely with the voice hearer, who sets the pace and may
disengage with the supportive other and/or discontinue the process
without negative consequences at any point. In addition, the
voice hearer “owns” the workbook in a tangible, concrete way seldom
applicable to traditional therapy. This locus of control reflects
the self-help philosophy of the workbook.
It occurred to the present authors that Coleman and Smith’s (1997)
written guidelines might constitute an additional repertoire of
teachable coping strategies for psychotic symptoms in typical mental
health settings. Despite workbook similarities to strategies
found in formal individual CBT of psychosis, mentoring the workbook
requires less training and experience than is needed for CBT.
Could an intervention built around the workbook produce measurable
positive changes in self-esteem, social functioning, depressive and
psychotic symptomatology, the very areas that formal CBT of psychosis
has explored? The present study addressed this question.
Methods
This study utilized a non-equivalent comparison group design with
non-random assignment to evaluate pre-post intervention change in
scores on standardized measures (see below). The study was
conducted in a south Florida community mental health agency that serves
adults diagnosed with severe and persistent mental disorders. At
the start of the study, the host agency served 480 “members.”
Approximately 61% of agency members were male and 39% female, 51% Anglo
and 49% Minority, with an age range from 18 to over 65. Agency
members had to report experiencing and/or display psychotic symptoms to
meet inclusion criteria (see below).
The sample was one of convenience, based on staff referrals from the
host agency. The Comparison group (n = 13) consisted of
participants not scheduled to intervention protocol (n = 17) after
signing informed consent paperwork and completing pre-test
packets. Randomization was not possible, as participants were
assigned to intervention protocol based on their ability to meet
initial scheduling windows available. Over the course of the
study, one participant suffered a stroke and left the Intervention
group and two Comparison group participants discontinued services at
the agency. Pre-test data from these three non-completers are not
included in the data set, since no corresponding post-test data was
obtained. Thus, data from the total of 27 participants who
completed the study are included in the analysis (Comparison group n
=11, Intervention group n = 16).
The sponsoring university’s Internal Review Board and the host agency’s
administration approved the study. The primary researcher met
individually with each person referred to the study in order to explain
the project and answer any questions about informed consent
documentation. Once informed consent was given, participants
could withdraw from the study at any time without penalty or reduction
of agency services.
Measures
The current study elected to measure change in overall symptomatology,
symptoms of depression and anxiety, and self-esteem, using standardized
instruments. These instruments have been used in CBT of psychosis
outcome studies in the United Kingdom (e.g., Haddock, Bentall, &
Slade, 1996; Kuipers et al., 1998; Tarrier et al., 1999), or in
community mental health research in the United States of America
(Newman, DeLiberty, McGrew, & Tejeda, 2005). Measures are
discussed below.
The Rosenberg Self-Esteem Scale (RSE, Fischer & Corcoran, 1994;
Rosenberg, 1989) is a self-report instrument used to measure
self-esteem. The RSE specifies a four-point scale used to
self-rate ten statements and was scored as a Likert scale.
Possible scores range from a low of 0 to a high of 30 (highest
self-esteem). The instrument’s short length and ease of
administration made it suitable for inclusion in a pre and
post-intervention packet of multiple measures.
The Hoosier Assurance Plan Inventory – Adult (HAPI-A) is a
standardized, clinician-rated psychosocial assessment instrument
utilized by the state of Indiana with adults diagnosed with severe and
persistent mental disorders. The HAPI-A includes the factor
Symptoms of Distress and Mood (Factor 1, based upon three items, A –
Consumer’s Rating of Symptom Distress, B – Anxiety-Worrying, and C –
Depression-Sad, Blue, or Suicidal Thoughts/Actions), and the item
Thought Disorder, Item H. Item H reads: “Have you had any
unusual experiences (e.g., are there times you hear, see, or smell
things other would claim are not there)?” Two other HAPI-A
factors can be used to measure disruption in life (Factor 3 – Community
Functioning, and Factor 4 – Social Support-Skills & Housing).
In the state of Indiana, the HAPI-A demonstrated sensitivity to change
for clients with psychiatric diagnoses, or psychiatric diagnoses and
chronic addiction, over a 90-day period HAPI-A (Newman, DeLiberty,
McGrew, & Tejeda, 2005).
For the HAPI-A, a lower score indicates a more severe problem or
symptom. Score range per item on the HAPI-A is from seven to
one. Since Factor 1 – Symptoms of Distress and Mood is made up of
three items, its score ranges from 21 to three. Factor 3 –
Community Functioning, and Factor 4 – Social Support-Skills &
Housing, are each comprised of four items, hence these scores range
from 28 to four. Item H (Thought Disorder) ranges from seven to
one and is reported separately.
The BPRS is a clinician-rated global instrument for measuring symptoms
of psychopathology that includes items related to depression and
anxiety (Faustman & Overall, 1999; Overall & Gorham,
1962). Overall and Klett (1972) identified four general factors
within the BPRS, including the factor Anxious Depression. This
factor is based upon three subscales, i.e., Anxiety, Guilt Feelings,
and Depressed Mood. The BPRS measures clinical symptoms across a
range of mental diagnoses, such as those present in the clinical sample
studied. Possible BPRS global scores for overall symptomatology
range from 18 to 126, where 18 indicates no symptoms and 126 indicates
all symptoms rate as extremely severe. The higher the BPRS global
or factor score, the more severe is the symptom rating.
Inclusion Criteria
Study data collection began in the fall of 2001 and extended into the
winter of 2004. Given a sample of convenience, non-random
assignment to group, and non-equivalent groups, results should not be
generalized to the larger population of Americans diagnosed with major
mental disorders and experiencing psychotic symptoms who reside in the
community. Study inclusion criteria included agency membership;
this excluded developmentally delayed or mentally retarded individuals,
as well as individuals using alcohol or illicit substances.
Additional inclusion criteria were as follows: (1) 21 to 65 years
of age; (2) no legal guardian; (3) English literacy/fluency; (4) agency
record of a DSM-IV (American Psychiatric Association, 1994) diagnosis
of Schizophrenia, Schizoaffective Disorder, or Mood Disorder with
psychotic features; if the disorder on record was none of these and
criterion six (below) was met, intervention protocol was deemed
potentially appropriate and the agency member was permitted to enroll
in the study (n = 1); (5) agency record of a DSM-IV Global Assessment
of Functioning (Axis V) score in the range of 35 to 60; (6) reports by
agency staff of observable symptoms such as delusional verbalization or
aberrant behavior, and/or verbalization of a problem related to
auditory hallucinations, “voices,” intrusive thoughts, or excessive
doubts or worries, despite reported adherence to prescribed
psychotropic medication; (7) community residence (defined to include
residential treatment facilities, assisted living facilities, and group
homes); and (8) either no alcohol or illegal substance use diagnoses,
or no current alcohol or illegal substance use per staff and
self-report, if such diagnoses were on record at the agency.
The first author extracted data on participant demographic and clinical
characteristics from agency case files. MSW-level social workers
not affiliated with the host agency and blind to
intervention/comparison group status administered pre and post-test
instruments. Weekly intervention sessions with each participant
began after completion of the pre-test packet, with the host agency
requirement that the first author act as mentor. Broad variation
in both number of sessions and duration of the intervention was
expected and, indeed, occurred (range from 12 to 42 weeks, with one
outlier at 57 weeks; details below).
A mentored session with an Intervention group participant began with
the Topography of Voices Rating Scale (TVRS). The TVRS is a
one-page, unstandardized self-report instrument (Chadwick, Birchwood
& Trower, 1996; Hustig & Hafner, 1990) used as an “ice-breaker”
to support open discussion of symptoms. Sessions were one-on-one
and followed a semi-structured format of 15 to 45 minute
duration. (The mentor documented details of each session on the
Mentor Report Form.) After completing the TVRS and responding to
initial inquiries on medication and current status, the participant was
handed the workbook. Each workbook exercise was read aloud and
any confusion clarified prior to the participant completing it.
Verbal responses were redirected with a gentle “write it down.”
Once the workbook was completed, the mentor facilitated contact between
participant and assessor to schedule the post-test. As a
participant completed intervention protocol, the mentor referred both
that participant and a Comparison group participant to an assessor for
post-testing (this was not possible for five Intervention group
participants, because of differing group sizes).
Research Questions and Data Analysis
This study aims to determine if a low cost mentored intervention that
uses Coleman and Smith’s (1997) workbook provides benefits to clients
experiencing psychotic symptoms beyond what treatment as usual
provides. This aim is reflected in the following four research
questions: (1) does self-esteem increase post-intervention,
relative to self-esteem of a comparison group not receiving the
intervention?; (2) does depression-anxiety decrease post-intervention,
relative to a comparison group?; (3) does overall psychotic
symptomatology decrease post-intervention, relative to a comparison
group?; and (4) does disruption in life lessen post-intervention,
relative to a comparison group?.
Repeated measures analysis of variance (Repeated measures ANOVA, or
RMANOVA) assessed whether positive change occurred over: (1)
self-esteem, (2) depression-anxiety, (3) overall psychotic
symptomatology, and (4) disruption in life. Two measures were
used for the latter three of the four constructs, and Bonferroni
corrections set Type I (p) error at .025 (i.e., .05/2) for these
analyses. The general linear model (GLM) approach within the
Statistical Package for Social Sciences (SPSS) Version 11.0 was used to
conduct the analysis.
ANOVA is relatively robust regarding failures to meet assumptions of
homogeneity and normality (Garson, 2005). The F-test (or F-ratio)
is the key statistic for ANOVA and its formula reflects whether the
variance among the group means (for given sample size and within group
variances) is significantly larger than the error variance within the
groups. With smaller variances and smaller samples, F is
conservative, i.e., it is more difficult to detect statistical
significance (Garson, 2005). When using the GLM model within
SPSS, the F-test is unaffected by unequal group sizes. In studies
evaluating cognitive-behavioral or other individual psychotherapy
approaches, large samples of individuals diagnosed with serious mental
disorders are seldom available (Gottdiener & Haslam, 2002).
Researchers have used any one of several effect size statistics to
describe the magnitude of any significant differences observed between
intervention conditions. In this study, effect size was estimated
using the “partial eta square” statistic, which reflects the proportion
of variance associated with a given variable.
Results
Specific demographic and clinical characteristics of Intervention and
Comparison group participants are shown in Table 1. Overall,
Intervention group participants tended to be: (1) an average of seven
years older, (2) more likely to have more debilitating psychiatric
disorder diagnoses, and (3) more likely to be of non-Hispanic White
ethnicity/race, than Comparison group members.
<>Table 1
Intervention and Comparison Group Demographics (n =
27)
|
|
Intervention
Group
(n
= 16)
|
Comparison
Group
(n
= 11)
|
|
Characteristic
|
n
(%)
|
n
(%)
|
|
Gender
|
|
|
|
Female
|
4
(25)
|
3
(27)
|
|
|
12
(75)
|
8
(73)
|
|
Race/Ethnicity
|
|
|
|
White
|
13
(81)
|
7
(64)
|
|
Minority
|
3
(19)
|
4
(36)
|
|
Marital Status
|
|
|
|
Single/Divorced
|
15
(94)
|
10
(91)
|
|
Married
|
1
(6)
|
1
(9)
|
|
Education
|
|
|
|
High
School or below
|
7
(44)
|
6
(55)
|
|
Post
High School/GED
|
9
(56)
|
5
(45)
|
|
Residence
|
|
|
|
Independent
|
11
(69)
|
7
(64)
|
|
Group
home
Diagnostic Category
|
5
(31)
|
4
(36)
|
|
|
15(94)
|
7(64)
|
|
Mood
Disorder
|
1(6)
|
3(27)
|
|
|
0(0)
|
1(9)
|
Table 2 shows that at pre-test, the Intervention group had
significantly higher BPRS global scores. However, the difference
on the BPRS Anxious Depression factor was not statistically significant
and other differences appeared trivial and were not statistically
significant.
This pilot study
employs a
quasi-experimental pre-post design (n = 27) to evaluate the impact of a
mentored self-help workbook (Coleman & Smith, 1997) intervention.
<>Table 2
Between Group Differences at
Pre-test
|
Measure
|
Group
(I = 16,
C = 11)
|
Mean +
SD
|
Range
|
F
(dF: 1,
25)
|
p
|
Partial
Eta
Squared
|
|
RSE
|
Inter.
|
19.31 +
6.26
|
8-30
|
0.613
|
|
.024
|
|
|
Comp.
|
21.09 +
5.03
|
14-30
|
|
BPRS Anx.Dep.
|
Inter.
|
8.06 +
3.13
|
3-13
|
3.428
(1,24)
|
.076
|
.125
|
|
|
Comp.
(10)
|
5.80 +
2.86
|
2-13
|
|
HAPI-A Mood
|
Inter.
|
15.94 +
4.19
|
9-21
|
.146
|
.706
|
.006
|
|
|
Comp.
|
16.55 +
3.82
|
6-21
|
|
BPRS Global
|
Inter.
|
37.56 +
7.79
|
25-50
|
5.729
(1,24)
|
.025
|
.193
|
|
|
Comp.
(10)
|
29.60 +
8.97
|
19-43
|
|
HAPI-A Tht.Dis.
|
Inter.
|
4.88 +
2.03
|
1-7
|
1.552
|
.224
|
.058
|
|
|
Comp.
|
5.82 +
1.78
|
1-7
|
|
HAPI-A Comm.
|
Inter.
|
20.56 +
5.14
|
11-28
|
.439
|
.514
|
.017
|
|
|
Comp.
|
21.82 +
4.35
|
14-28
|
|
HAPI-A Soc.Supp.
|
Inter.
|
23.69 +
4.01
|
11-28
|
.001
|
.982
|
<.001
|
|
|
Comp.
|
23.73 +
5.02
|
14-28
|
The number of mentored self-help sessions for 15 Intervention group
participants ranged from eight to 18 (mean = 12.6) with a duration
range of 12 to 42 weeks (mean = 28.3), excluding the one participant
with 25 sessions over 57 weeks. In sum, there was a pronounced
central tendency of 13 sessions over a 29-week period, despite
differing levels of education/literacy and scheduling
idiosyncracies. As noted above, the intervention aimed only for
workbook completion (a behavioral criteria) rather than for a specific
“dosage” of intervention, thus neither the duration of the intervention
nor the number of sessions should be taken as an indicator of quantity
or intensity: Longer duration indicated slower progress in
completing workbook exercises.
Pre-post differences analyzed using RMANOVA are shown in Table 3.
Using RMANOVA and appropriate Bonferroni corrections, no significant
differences were found for self-esteem, overall psychotic symptoms, or
disruption in life. For depression and anxiety, the two measures
used were the BPRS Anxious Depression and HAPI-A Distress-Mood
factors. Because the statistical test required two separate
analyses, a Bonferroni correction set Type I error (p) at .025 (i.e.,
.05/2). No significant difference was found between groups for
the HAPI-A factor, however, a significant difference was found between
groups for the BPRS factor, with a strong effect size, η2 = .218.
Further, the correlation between the change in scores for the BPRS
factor Anxious Depression and the change in scores for the HAPI-A
Factor Symptoms of Distress and Mood was significant at the .05 level
(r = -.341). In sum, one of two measures of depression-anxiety
symptoms suggested that these symptoms showed significantly greater
reduction for the Intervention group and the two measures were
significantly correlated with each other.
<>Table 3
Pre-post Test Differences: ANOVA
Results
|
Measure
|
Group
(I
= 16, C = 11)
|
Mean
+ SD
|
Range
|
F
(dF:
1, 25)
|
p
|
Partial
Eta Squared
|
|
RSE
|
Inter.
|
18.81
+ 4.17
|
10-26
|
0.006
|
|
<.001
|
|
|
Comp.
|
20.73
+ 4.27
|
16-30
|
|
BPRS Anx.Dep.
|
Inter.
|
6.69
+ 3.24
|
3-13
|
6.985
(1,
24)
|
.014
|
.218
|
|
|
Comp.
(10)
|
7.70
+ 3.89
|
1-15
|
|
HAPI-A Mood
|
Inter.
|
15.31
+ 4.50
|
8-21
|
.193
|
.664
|
.008
|
|
|
Comp.
|
16.73
+ 3.04
|
11-21
|
|
BPRS Global
|
Inter.
|
37.94
+ 12.0
|
18-65
|
.577
(1,
24)
|
.455
|
.023
|
|
|
Comp.
(10)
|
34.70
+ 10.8
|
23-51
|
|
HAPI-A Tht.Dis.
|
Inter.
|
5.25
+ 1.77
|
2-7
|
.016
|
.902
|
.001
|
|
|
Comp.
|
6.09
+ 1.14
|
3-7
|
|
HAPI-A Comm.
|
Inter.
|
22.19
+ 3.60
|
14-27
|
.067
|
.797
|
.003
|
|
|
Comp.
|
23.82
+ 3.09
|
16-28
|
|
HAPI-A Soc.Supp.
|
Inter.
|
23.13
+ 4.84
24.64
+ 3.07
|
14-28
17-28
|
.707
|
.408
|
.028
|
|
|
Comp.
|
Discussion
The study used three standardized instruments with an Intervention and
a treatment-as-usual Comparison group, to examine pre-post differences
over four constructs. Research questions related to self-esteem,
depression-anxiety, overall psychotic symptomatology, and
disruption-in-life, respectively. Interestingly, five of seven
outcomes showed non-significant change scores that favored the
Comparison group over the Intervention group. These outcomes do
not reach statistical significance, and whether they would sustain with
a more powerful study is a question for future research.
The study is an exploratory pilot and findings are limited.
Results cannot be generalized, given the small sample size, non-random
assignment to group, non-equivalent groups, and other study limitations
(see recommendations below). Overall, a conservative
interpretation of results is that the one significant difference in
pre-post instrument scores occurred due to chance. An alternative
interpretation is that the mentored self-help intervention made an
actual improvement in the level of depression-anxiety experienced by
Intervention group participants. This is particularly relevant,
if so, given the high levels of depression and depression-associated
suicide among individuals diagnosed with schizophrenia. The
alternative interpretation supports further research on the
intervention.
Recommendations for a future study address limitations of the current
one and include: (1) random assignment to group; (2) use of
different mentors simultaneously, each mentor working one-on-one with
individual participants, to avoid the confound of a single individual
administering the intervention; (3) advance stipulation of the number
of sessions to standardize “dosage;” (4) use of specific standardized
instruments to assess symptoms of depression and anxiety; (5) use of a
standardized measure of self-efficacy rather than self-esteem; and
possibly (6) use of qualitative measures, rather than the HAPI-A, to
assess disruption in life.
The statistically significant result obtained was over the construct
depression-anxiety and one might cautiously conclude that the single
pre-post difference was an artifact of multiple measures over multiple
questions, i.e., that one of seven measures (14%) showed improvement
purely due to chance, despite a priori formulation of research
questions and an appropriate use of Bonferroni corrections in the
analysis. The change in the more disturbed Intervention group
might also indicate regression towards a mean; if so, however, one
might have expected the global score of the same scale to show a
similar shift and this did not occur. The BPRS Anxious Depression
factor pre-post difference showed a strong effect size, and although
the difference was only seen with one of two measures, the difference
scores on both the BPRS and the HAPI-A factors correlated significantly
at the .05 level. A more positive interpretation of observed
results is that the intervention served to assist individuals with
severe and persistent psychotic symptoms in reducing distress
associated with at least some of those symptoms.
Few CBT studies have measured secondary aspects of psychosis such as
depressive symptoms or anxiety, although Sensky et al. (2000) and
Turkington, Kingdon and Turner (2002) reported improvements in
depressive symptoms. Rector and Beck (2001) noted “upwards of
two-thirds of patients receiving a diagnosis of schizophrenia will also
experience a major depressive episode” (p. 285). Kaplan and
Sadock (1998) reported up to ten percent of people diagnosed with
schizophrenia die from suicide and that an estimated 4,000 people
diagnosed with schizophrenia in the United States die annually by
suicide (see also World Health Report, 2001). Further, only a
small percentage of these people commit suicide because of psychotic
symptoms: depressive symptoms are associated with over two-thirds of
these suicides. Beck (in Kingdon & Turkington, 1994)
correctly highlighted depressive symptoms as natural sequelae of a
schizophrenic diagnosis. If the mentored self-help workbook
intervention is shown in future studies to decrease levels of
depression in this population, the intervention format lends itself to
implementation in community mental health agency settings through
supportive staff.
An excursion into speculative thought is in order, given (1) the
observed reduction in BPRS Anxious Depression scores for the
Intervention group, (2) the positive impact of all major forms of
individual therapy for those diagnosed with schizophrenia (Gottdiener
& Haslam, 2002), and (3) the prevalence of depressive symptoms and
suicide in this population. Given the demonstrated positive
impact of all forms of individual therapy, depressive symptoms and
associated suicides might be reduced by use of interpersonal,
relational interventions. The implication for practice with this
population is that developing interpersonal practitioner-client
relationships based on mutual respect, trust, and client choice is a
critical foundation of treatment. Use of a self-help workbook
co-authored by a mental health consumer might assist the development of
such practitioner-client relationships, whether the workbook was used
one-on-one or within a group format. Study results indicate that
the workbook – despite its departures from mainstream mental health
thinking – when used one-on-one with a supportive mentor, appeared to
do no harm and that it may in fact be a positive process for clients
experiencing medication resistant psychotic symptoms.
