University of Southern California Dissertations and Theses

Client characteristics, service characteristics, and psychosocial outcomes in the community treatment of schizophrenia

(USC Thesis Other)

Client characteristics, service characteristics, and psychosocial outcomes in the community treatment of schizophrenia

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zeeb Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600

CLIENT CHARACTERISTICS, SERVICE CHARACTERISTICS, AND
PSYCHOSOCIAL OUTCOMES IN THE COMMUNITY TREATMENT OF
SCHIZOPHRENIA
By
Mark G. Ansel
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment o f the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Social Work)
May 1996
Copyright 1996 Mark G. Ansel
UMI Number: 9636319
Copyright 1996 by
Ansel, Mark Gerard
All rights reserved.
UMI Microform 9636319
Copyright 1996, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES. CALIFORNIA 90007
This dissertation, written by
Mark G. Ansel
under the direction of h...VA Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillment of re
quirements for the degree of
DOCTOR OF PHILOSOPHY
.....
Dean of Graduate Studies
D a te 3/ 27/96
DISSERTATION COMMITTEE
Chairperson
'JJXQUeAAAczi.. Qli
O aH . .
DEDICATION
This dissertation is dedicated to my wife and partner in life Miki Grimes, and our
children Christopher, J.D, and Emma. Without your love, faith, and encouragement this
never could have occurred. This dissertation and degree are every bit as much yours as
they are mine. May our lives be as exciting and fun-filled in the future, as they were
challenging during this process.
ACKNOWLEDGEMENTS
When you spend a number o f years completing a degree such as this, the number
of close relationships you develop during that process is truly amazing. To begin, this
dissertation and degree would have been impossible to complete without the support of
the Halpem-Brekke Home for Wayward Doctoral Students (whose hospitality can not be
surpassed by any 5 star hotel), and particularly that o f Dr. John S. Brekke. You have not
only been my mentor and sponsor in life throughout this endeavor, but more than that you
have been my good friend. Thanks for the constant encouragement and helping to keep
me focused.
I am also deeply indebted to my mother-in-law, Ms. Ginger Everhart, who
supported us not only through this degree, but also through the completion of my M.S.W.;
and my sister-in-law, Ms. June Grimes, who was there for us literally from the time we
landed in Los Angeles with a built in network of new friends that I will never forget.
I would also like to acknowledge the support and encouragement I received from
Dr. Paul Carlo, Dr. Robert Nishimoto, Mrs. Ruth Britton, the faculty and staff of the USC
School o f Social Work, my peers, and especially the soon to be Drs. Dennis Durby,
Richard Renz-Beaulaurier, and Judith DeBonis. It's been one heck of a ride folks, but if
you were all up for it I would definitely do it again.
Back home in Hawaii I would also like to acknowledge the support and
encouragement I have received over the past 10 years of my life from my good friends Mr.
Bruce Kennard, Mr. Hugh Okuda, the soon to be Dr. Langley Frissell, and their families.
I am eternally grateful for everyone's friendship and assistance. Mahalo and Aloha.
Table o f Contents
Introduction.......................................................................................................................................1
Review of the Literature and Preliminary Studies.......................................................................... 4
Study Hypotheses.......................................................................................................................... 35
Method........................................................................................................................................... 39
Results.............................................................................................................................................54
Discussion.......................................................................................................................................81
List of Figures and Tables............................................................................................................. 97
References.....................................................................................................................................115
Figure I.
Table 1
Table 2.
Table 3.
Table 4.
Table 5.
Table 6.
Table 7.
Table 8.
Table 9.
Table 10.
Table 11.
Table 12.
Table 13.
Table 14.
List o f Tables and Figures
Definition o f Major Variables............................................................................. 96
Demographic Characteristics.............................................................................. 97
Levels o f Participant Baseline Functioning for 6 Month and 12 Month
Samples................................................................................................................ 98
Service Process Variable Descriptive................................................................. 98
Intercorrelations Between Service Process Variables at 6 and 12 Months.... 99
Mean Contact Duration and Frequency by Type of Service at 6 and 12
M onths............................................................................................................. 100
Correlations Between Service Process Variables and Participant Baseline
Functioning at 6 and 12 M onths...................................................................... 101
Paired T-Test's Using 6 and 12 Month Samples On All Participant Outcome
Indicators...........................................................................................................102
Correlations Between Participant Characteristics and 1 Year Change Score
Outcome Variables............................................................................................ 103
Correlations Between Service Process Variables and 6 and 12 Month Change
Score Outcome Variables................................................................................ 104
Correlations Between Participant Characteristics and 6 Month Service
Process Variables...............................................................................................105
Gender, Age at Study Entry, and RFS Independent Living 12 Month
Outcomes........................................................................................................... 105
Gaps in Service, Comprehensiveness, and 6 Month SCOS Use of Institutions
Outcomes........................................................................................................... 105
Total Duration, Comprehensiveness, and 12 Month BPRS Outcomes 106
Total Duration, Comprehensiveness, and 12 Month SCOS Use of Institution
Outcomes........................................................................................................... 106
Table 15. Total Duration, Comprehensiveness, and 6 Month GAS Outcomes 106
vi
Table 16. Total Duration, Comprehensiveness, Gaps in Service, and 6 Month RFS Total
Outcomes............................................................................................................107
Table 17. Total Duration, Comprehensiveness, Gaps in Service, and 6 Month RFS
Independent Living Outcomes...........................................................................107
Table 18. Total Duration, Comprehensiveness, Gaps in Service, and 12 Month SCOS
Total Outcomes.................................................................................................. 108
Table 19. Total Duration, Comprehensiveness, Gaps in Service, and 12 Month GAS
Outcomes............................................................................................................109
Table 20. Total Duration, Comprehensiveness, Gaps in Service, and 12 Month RFS
Total Outcomes................................................................................................. 109
Table 21. Gaps in Service, Comprehensiveness, and 12 Month RFS Work
Outcomes............................................................................................................110
Table 22. Total Duration, Age at Study Entry, and 12 Month SCOS Use of Institutions
Outcomes...........................................................................................................110
Table 23. Total Duration, Race, and 6 Month RFS Independent Living Outcomes 110
Table 24. Total Duration, Age at Study Entry, and 12 Month GAS Outcomes 110
Table 25. Total Duration, Age at Study Entry, and 12 Month RFS Total Score
Outcomes............................................................................................................I ll
Table 26. Correlations Between Specificity o f Treatment and 6 Month Change Score
Outcomes............................................................................................................112
Table 27. Correlations Between Specificity of Treatment and 12 Month Change Score
Outcomes............................................................................................................113
ABSTRACT
This dissertation examined hypothesized relationships between participant
characteristics, service characteristics, specificity of treatment, and psychosocial outcomes
in the community treatment of schizophrenia at 6 month (n=41) and 12 month (n=30)
outcomes. The participant characteristics included gender, ethnicity, age at study entry,
and the Strauss and Carpenter Prognostic Scale. The Daily Contact Log measured the
type of service, as well as variables that represented the concepts o f Continuity o f Care,
Treatment Intensity, and Service Comprehensiveness. Participant functional outcomes
were measured across the clinical, psychosocial, and subjective experience domains. The
outcome measures were inclusive of the BPRS, QLS, SCOS, GAS, RFS, SLS, and the
ISE. Change scores were calculated from index to the respective outcome period.
Findings indicated that better continuity of care was related to, and explained better
clinical outcomes. More treatment intensity was associated with, and explained better
psychosocial outcomes. Support was also generated for specificity of treatment.
INTRODUCTION
1
Specific Aims
This dissertation describes a 1 year prospective investigation that evaluated
the impact o f community based care and rehabilitation of individuals diagnosed
with schizophrenia or schizophrenia related disorders. The data utilized came from
a single "model" Community Support Program (CSP) that was an integral part of a
multi-site longitudinal evaluation o f two model CSP's and traditional board and
care/case management. The aims o f this dissertation were to test and explore the
relationships between participant characteristics, service characteristics, and
participant outcomes over a 1 year period. The following questions were
addressed:
1. Given participant characteristics, which participant variables have the most
significant impact upon various participant outcomes at 6 and 12 months?
2. Are service characteristics (continuity o f care, comprehensiveness of
service, and intensity o f service) predictive of improvement in various
participant outcomes at 6 and 12 months?
3. Do groups of participant and service variables exist that predict various
participant outcomes in this sample at 6 and 12 months?
4. Given the various service domains (i.e., living skills, vocational
rehabilitation) provided in this model CSP, are specific service variables
associated with corresponding participant outcomes in community
treatment at 6 and 12 months?
Significance
Since the National Institute o f Mental Health (NIMH) instituted the
Community Support Program (NIMH, 1982) treatment ideology there have been
several investigations concerning the efficacy of this treatment method. Findings
from reviews o f the literature (e.g., Bond, McGrew, and Fekete, in press; Test,
1992, 1984; Solomon, 1992; Olfson, 1990; Hargreaves and Shumway, 1989;
Tessler and Goldman, 1982; Braun, et al., 1981; Test and Stein, 1978) indicate
that persons with severe and persistent mental illness can survive in the community
when supportive community intervention is available. Yet, other than the fact that
we now know that CSP treatment can be beneficial to clients when executed
without alteration (Stein, 1990); little else is known about what makes these
models o f care operate successfully (Test and Brekke, 1992; Brekke, 1987; Test,
1984; Test and Stein, 1978; Bachrach, 1982). Therefore, what is currently needed
is better program implementation information to aid in the study of specific
intervention ingredients and to facilitate the replication o f successful programs.
Similarly, while reviews have indicated that CSP treatment is efficacious,
the many studies that investigated the efficacy of CSP intervention concern (a) the
application o f community based treatment in one form or another to a hospital
based control (Solomon, 1992; Hargreaves and Shumway, 1989; Test 1984; Test
and Stein, 1978; Braun et al., 1981) and/or (b) offer little or no program
implementation information (Brekke and Test, 1992; Brekke, 1988; Brekke, 1987;
Test, 1984). In regard to these matters, several investigators have indicated there
is a need for more inquiry into the identification of beneficial service elements and
exploration of service functions (Solomon, 1992; Holloway, McClean, and
Robertson 1991; Olfson, 1990; Taube, Morlock, Bums, and Santos, 1990; Test,
1984), as well as, aiding in comparisons across studies (Kluiter, Giel, Nienhuis, et
al., 1992). Additionally, a review of the literature also revealed that while client
characteristics had been investigated in schizophrenia research, the impact o f these
variables had not been evaluated in community based care.
The present study will be the first empirical investigation in the community
treatment o f schizophrenia to examine the relationships between participant
characteristics, service characteristics, and participant functional outcomes. By
measuring service characteristics the present investigation will examine the
relationships and explanatory impact of service process variables on participant
outcomes across three distinct domains o f functioning. Additionally, this
investigation will examine the impact of both specific and non-specific treatment
upon participant functional status. In so doing ,this study will be the first of its
type and will attempt to answer the question of, "What treatment elements, effect
what outcomes, in the community treatment of schizophrenia."
REVIEW OF THE LITERATURE AND PRELIMINARY STUDIES
Community Treatment
One o f the earliest critical reviews of community based treatment was
conducted by Test and Stein (1978). In this review community treatment was
defined by those authors as, "any treatment that takes place in the community
either in lieu o f hospitalization (alternatives to the mental hospital), following early
discharge ('premature release'), or after hospitalization (aftercare)" (p. 351).
Although the treatment modality was in its developmental stage at the time o f the
review, Test and Stein concluded that community treatment o f serious mental
illness was at least as effective as hospital based treatment.
In a similar undertaking Braun, Kochansky, and Shapiro, et al. (1981)
conducted a review of deinstitutionalization. These authors reviewed 30 studies
that investigated (a) alternatives to hospitalization, (b) modifications of
traditionalized hospitalization, (c) alternatives to continued long-term
hospitalization, and (d) alternatives to hospital admission. Numerous design
weaknesses were noted in several o f the studies. However, these authors
concluded that, "The most satisfactory studies allow the qualified conclusion that
selected patients managed outside the hospital in experimental programs do no
worse and by some criteria have psychiatric outcomes superior to those o f
hospitalized control patients" (Braun, et al., 1981, p.747). On the other hand,
these authors also suggested that deinstitutionalization and the study of community
programs may prove to be unsuccessful if there is a lack of continuity of care.
Tessler and Goldman (1982) have also conducted an empirically based
exploratory investigation o f CSP's. In their study a sample of 1471 clients were
randomly selected from 248 case managers representing 4288 clients from multiple
sites across the nation. Tessler and Goldman observed that (a) a greater number
of services were received by clients with significant behavioral problems and
deficits in basic living skills; (b) work status was affected by predictors o f basic
living skills, and behavioral and somatic problems; (c) basic living skills were the
best predictor o f work status and social activity; (d) the best predictor of
hospitalization and emergency contact were the number and severity o f behavioral
problems; and lastly (e) work performance and type o f job were also affected by
living skills. A concomitant finding was that somatic problems were also
important in number of hours worked and earned income. The overall conclusion
offered by these authors was that CSP interventions provided a wide range of
services to meet clients needs in community living.
In a later review o f CSP's Test (1984) also indicated that studies of
community alternatives, although not originally conceptualized as CSP's, can be
considered as such because they meet a cluster of Community Support System
(CSS) functions. Test identified eleven studies as alternatives to hospitalization,
only one o f which evaluated community treatment per se (Stein & Test, Test &
Stein, cited in Test 1984). Regarding these studies Test concluded:
...it appears that CSP's can be a useful and effective alternative to
mental hospital treatment for patients on the brink of admission.
They reduce time spent in hospitals and lower symptomatology as
much as hospitalization. A few CSP models show advantages in
the psychosocial area. Also, when assessed, patients indicated
greater satisfaction when treated in this option relative to
hospitalization. (1984, p.357)
Test (1984) also identified three studies that investigated the effectiveness
o f CSP's in relation to continued hospitalization. The findings of this group of
studies was consistent with the findings o f CSP's as alternatives to hospitalization.
A fourth group of studies reviewed by Test involved comparison o f two or more
programs o f treatment in the community, with varying degrees of
comprehensiveness. These community care studies were categorized by type of
treatment compared and included: (a) out-patient contacts, (b) family plus
intervention, (c) day treatment, (d) residential milieu, (e) residential and work
milieu, and (f) psychosocial rehabilitation programs. Although there was
significant diversity across categories Test offered in summary that in the studies
that compared some form o f organized community treatment, relapse was
significantly reduced and that in some programs clients social functioning
improved. Second, in programs that were compared on the level of
comprehensiveness, the more comprehensive program had fewer patient relapses
and, in some cases, patients had better social adjustment.
In another review concerned with the efficacy of research findings in
various methods o f community treatment, Hargreaves and Shumway (1989)
reviewed forty-five empirical studies inclusive o f those reviewed by both Test
(1984) and Braun et al. (1981). These authors concluded:
...caring for severely ill psychiatric patients in ways that avoid or
shorten traditional hospital treatment is, on average, at least equally
effective and may be more effective than standard hospital care.
Well organized services using alternatives to hospitalization can
cost less, sometimes much less, without incurring offsetting social
or private costs, and may provide greater improvement in
symptoms or social functioning, (p.267).
Concerning program components, Rosenfield and Neese-Todd (1993) have
also investigated which elements o f a clubhouse program (based on the Fountain
House model) were associated with a satisfying quality of life. Findings indicated
areas that can be enhanced or replaced by psychosocial rehabilitation (i.e., work
status, social relations, and leisure) were related to greater satisfaction of quality of
life. Additionally, the perception o f empowerment was related to satisfaction with
a broad range of quality of life considerations. It should be noted that while this
study focused on psychosocial rehabilitation it did not specify the amount of
service intensity necessary to replicate successful service delivery.
In reference to residential treatment, Wherley and Bisgaard (1987)
observed three levels o f care existed (intensive, transitional, and supportive) in a
countywide study of 19 residential programs. Their findings indicated that
residential programs aided in community assimilation and were cost effective
regardless of intensity. Similarly, Fisher et al. (1992) found that large numbers of
persons with severe and persistent mental illness can be supported in the
community without high recidivism if comprehensive services are available.
However, these same data also indicated there were no significant differences in
patterns o f recidivism when regions in Massachusetts were compared on level of
service comprehensiveness. Additionally, a 10 year follow-up study concerning
the effects of deinstitutionalization in a Massachusetts state facility indicated that
while short-term stays (30 or less days) were decreased, there was a subsequent
increase in stays between 30 days to 1 year (Dorwart, 1988). Dorwart attributed
the decrease in short-term stays to an increase in the use of hospitalization in
community general hospitals.
Equivocal findings have also been noted in vocational rehabilitation. For
example, Bond and Boyer (1988) considered some o f the same studies noted in
the reviews discussed above in an analysis o f the impact of vocational rehabilitation
upon employment. The general conclusion o f Bond and Boyer was that neither
vocational rehabilitation programs, hospital based programs, sheltered work,
assertive case management, psychosocial rehabilitation, supported employment,
rehabilitation counseling, job clubs, remedial education and occupational training,
nor post-employment services, "demonstrated efficacy in helping clients achieve
and maintain employment over any sustained period o f time" (p.252). On the
other hand, these authors also suggested when the outcome criterion was paid
employment, findings indicated that when clients are offered intensive support and
placed in positions that have a low demand, they function higher than anticipated.
Hence, Bond and Boyer proposed that supported work programs may be the most
efficacious way to increase employment rates for this population.
Additional equivocal findings have also been noted in studies that
concerned organic community programs. For example, hospital based day
treatment has been documented to be efficacious regardless o f psychiatric illness
(Kluiter, et al., 1992). Additionally, Muijen et al. (1992) conducted a controlled
study of home based community care and hospital based treatment. While these
authors found that hospital use was decreased by 80% for the home based group,
total expenses for that group were only 1% less at 18 month follow-up than for
hospital based care (Muijen, et al., 1992).
Case Management
As noted by Test (1984) for a CSS to function efficiently all of its
component parts would need to be integrated into one cohesive system. One
method o f service coordination in the community has been the use of a case
management. In a sense, case management is a means of establishing artificial
linkages between the different components of care in a community through a
person or team (case manager), that coordinates and maintains continuity o f care
for mentally ill persons (Hargreaves and Shumway, 1989; Talbott, 1988).
Bachrach (1993 a) reports there is a continuum to case management services. At
one end is the broker o f services, while at the opposite end is the clinical case
manager who assumes clinical duties in addition to brokering services (it appears
an MSW is best suited for clinical case management).
Clark et al. (1993) indicated there are four models that fall into the case
management continuum. First, is the Expanded Broker Model with responsibility
for assessing needs and linking clients to existing services. Second, is the Personal
Strengths approach, where the case manager identifies clients abilities and creates
10
situations where strengths can be exercised to achieve personal goals. This is
generally accomplished through linkage to resources such as vocational training,
social support, etc. Third, the Rehabilitation Model emphasizes client goals and
focuses on identifying skills and teaching clients the skills needed to overcome
deficits. Lastly, is the Clinical Case Manager or Full Support Model, which
provides both direct care and brokerage of services (i.e., money management,
housing linkages). Recent conceptualizations of case management have suggested
it may be efficacious to use a multi-disciplinary team approach to further integrate
services (Munich and Lang, 1993). However, it should be recognized that
regardless of the approach utilized, the correct way to do case management
ultimately depends on a given patients' needs and on the systems desire and
potential for responding to them (Bachrach, 1993a).
While much o f the recent research concerning community support and
treatment has in some way concerned the efficacy o f case management, it should
be noted that such studies look more at models of service delivery, rather than an
integrated and comprehensive treatment program or the psychosocial rehabilitative
services that may compose a CSP. Results of these studies and reviews have been
equivocal at best. For example, Solomon (1992) in a review o f twenty empirical
investigations o f case management found: (a) most were effective in reducing the
number of hospitalizations, and in instances of hospitalization, reducing the length
of stay; (b) case management appeared cost effective and improved clients quality
of life; (c) treatment satisfaction was consistently high; and (d) while case
management appeared to have negligible positive effects in clinical functioning it
did not appear to produce negative effects and was as effective as hospitalization
with subsequent aftercare. Similarly, other investigations have also found that case
management may decrease hospitalization (Homstra, et al., 1993; Santos, et al.,
1993; Bond et al., 1988), and it has been shown that this effect may be related to
intensity o f service, rather than quantity of services (Dietzen, et al., 1993). On the
other hand, in a multivariate model continuity o f service was included rather than
intensity (Tessler, 1987). In contrast, other investigations have shown that case
management had modest effects on quality of life (Bond, et al., 1988; Franklin et
al.,1987).
It has also been indicated that a case management approach may be more
expensive, but this could be due to the element o f ensuring continuity o f care
(Franklin, et al., 1987). In contrast, Bond et al. (1988) found that an assertive
case management approach had a savings of $5500 per client per year compared to
controls. Similarly, Santos et al. (1993) found in a rural implementation o f the
Training in Community Living (TCL) model (Stein and Test, Test and Stein cited
in Test, 1984) which has empirical support and emphasizes aggressive community
treatment and outreach, there was a 52 percent reduction in estimated cost of care.
Franklin et al. (1987) also suggested that while adding case management to a
system with plentiful resources may have overall meager effects, it may also
prevent some persons from "falling through the cracks". Similarly, it has been
12 .
observed that when persons do not receive the needed services immediately after
discharge outcomes may be poorer (Tessler, 1987).
Recently, Olfson (1990) has also reviewed twenty-two empirically based
implementations of what he considered to be replications o f the TCL model. In
this review Olfson concluded that the positive findings reported in TCL treatment
have been limited to the Madison studies (Stein and Test, 1980; Test, Knoedler,
and Allness, 1985, cited in Stein, 1990) and a replication in Australia (Hoult,
Reynolds, Charbonneau-Pwis, et al., cited in Olfson, 1990). However, in a strong
rebuttal Stein (1990) has commented that the findings of Olfson (a) point to the
tendency to "lump a variety of similar interventions under one name and then treat
them as if they were all the same" (p.651), and (b) that there were obvious
differences in the implementation of programs reviewed in comparison to the
Madison TCL model which is the likely explanation of differences in outcome.
Test (1992) has also conducted a review of empirical studies that replicated
the Training in Community Living model of community care, which was inclusive
of studies reviewed by Olfson. In contrast to Olfson (1990), Test (1992)
indicated that the TCL model is (a) effective in reducing time spent in hospitals, (b)
that the positive results have been found in a variety of different geographic
regions in North America and elsewhere, as well as, different staffs. Perhaps most
significantly, Test (1992) noted that the model has also been generalized across
different patient groups. Test noted that the TCL model appears most effective
when services are provided by a "core" team rather than brokered as in other case
management models. Additionally, in offering an alternative explanation to the
findings of Franklin et al. (1987), Test suggested case management models that
demonstrate poor outcome, "suggest that a large proportion of persons with
serious mental illness may require the kind o f assertive outreach and continuity
provided by a TCL core team" (Test, 1992, p. 166).
One possible way of explaining away the equivocal nature regarding
effectiveness o f case management studies and reviews would be to conduct a
review that utilized meta-analytic methods. Recently, Bond, McGrew, and Fekete
(in press) have completed an undertaking using this empirical methodology in a
review of the Thresholds Bridge model and replications. Concomitantly, it should
be noted that the Thresholds model is a derivative of the TCL model (Bond,
Boyer, and McGrew, in press). The findings from this empirical review indicated
that clients had a high level of retention (80%) in assertive programs, and that
heavy users of hospitalization had inpatient days decrease by 50%. On the other
hand, the data concerning quality of life over time were mixed. However, Franklin
et al. (1987) have speculated such findings may be observed later in treatment.
Additionally, it has also been suggested that poor outcomes in quality of life may
be due to measurement error (Lehman and Bums, cited in Bond, et al., in press).
It would seem that case management may be a means of prevention of
illness and enhancing outcome for some clients. On the other hand, it may be
expensive and unnecessary for others depending on the model of care
implemented. As the review above indicates there is need for continued research
14
into this method of coordinating the delivery of client services as well as agreement
on which variables should be measured in defining outcome.
Putative Service Variables Related to CSP Outcome
Essentially, continuity o f care and intensity o f service delivery are the two
primary service characteristics that have been discussed as possibly being related to
positive client outcomes. On one hand, continuity o f care has been widely
discussed in the literature ( e.g., Bachrach, 1993a, 1993b; Brekke and Test, 1992;
Talbott, 1988; Tessler, 1987; Tessler, Willis and Gubman, 1986; Braun, et al.,
1981; Test, 1979; Tessler and Mason, 1979). On the other hand, Tessler, Willis,
and Gubman (1986) have indicated the concept suffers from various conceptual
dilemmas, such as a lack of clear definition and boundaries in interpretation that
lead to confounds. In comparison to continuity of care, service intensity is a
relative newcomer in schizophrenia intervention research, but also suffers from
similar confusion in definition and operationalization. Similarly, the importance of
treatment specificity has been raised (Solomon, 1992; Olfson, 1990; Test, 1984),
but has not been thoroughly discussed nor empirically investigated. This section
will present the various operationalization o f the concepts and the emerging
discussion regarding the importance of treatment specificity. The subsequent
section will be dedicated to empirical investigation o f these variables in treatment
settings.
Continuity of Care
Regarding the importance of continuity of care, Braun et al. (1981) noted
in their review o f community treatment (noted earlier) that deinstitutionalization is
likely to be unsuccessful in the absence of continuity of care in community
treatment efforts. Similarly, Talbott (1988) indicated that continuity of care may
be the single most important factor in the community treatment o f mental illness.
In a recent conceptualization o f continuity of care Bachrach (1993 a) indicated nine
guiding principles. The nine identified characteristics ranged from administrative
mandates that endorse the concept, to cultural appropriateness to ensure clients
remain in the community support system.
Continuity o f care has also been characterized across both cross-sectional
and longitudinal dimensions (Brekke and Test, 1992; Test, 1979). The cross-
sectional dimension concerns the comprehensiveness and level o f integration of a
client at any given point o f time they are involved in the community support
system. Test (1979) has indicated that the primary problems in ensuring cross-
sectional continuity are (a) gaps in the system that cannot meet a clients needs, (b)
poor coordination and communication in the system, as well as undefined
responsibilities, and (c) difficulty in keeping the client engaged in the system of
care. In operationalizing cross-sectional continuity o f care Brekke and Test (1992)
have suggested that comprehensiveness may be represented by "the extent to
which the program itself is comprehensive as well as the amount o f advocacy and
linkage services which it provides" (p. 232). Concerning longitudinal continuity
16
Brekke and Test (1992) indicated that this dimension of continuity refers to care
that is provided continuously in an integrated fashion. Additionally, the system
must be sensitive and able to meet the clients individual changing needs throughout
the course of treatment (Test, 1979). Brekke and Test (1992) suggested
longitudinal continuity may best be operationalized as the extent that clients are
involved in treatment continuously over time without long gaps in services, as well
as, the measurement and explication of treatment drop-out rates.
Similarly, Tessler, Willis, and Gubman (1986) indicated that dimensions of
continuity of care be inclusive of: (a) discharge planning, as this is a crucial
juncture between hospital care and entry into community treatment; (b) successful
and rapid transfer to services that are recommended (i.e., were suggested services
actually received); and (c) implementation of individualized treatment plans to
meet client needs. These authors have also suggested that when continuity o f care
is studied that barriers to continuity such as difficulty in obtaining a suggested
service or client related breaks in service be measured as well.
Treatment Intensity
In contrast to continuity o f care the discussion concerning intensity of
service delivery has been limited. Apparently this is due to the reality that the
majority o f community services research has been focused on the effectiveness of
service delivery with little regard for the explication of specific aspects of
treatment (Brekke and Test, 1992; Snowden and Clancy, 1990; Brekke, 1988,
1987). In fact, it has only been recently that intensity of service delivery has
17
surfaced as a variable worthy of empirical analyses. In the studies that were
identified regarding service intensity there has not been clear agreement between
researchers as to what the operational definition of intensity should be.
Most commonly service intensity has been represented by the number of
service contacts (Dietzen and Bond, 1993; Brekke and Test, 1992; Snowden and
Clancy 1990; Snowden, Storey, and Clancy, 1989). Brekke and Test (1992) have
also suggested that intensity be measured in terms of the actual minutes o f contact
with the client. In contrast, Tessler (1987) has operationalized intensity as the
number o f services received, which Brekke and Test (1992) have defined as
service comprehensiveness. It is also of interest to note that while
comprehensiveness has been hypothesized as being an important clinical
component, only the pilot study of Brekke and Test (1992) has investigated this
variable in service delivery research.
Specificity of Treatment
A related discussion has also been developing regarding the importance of
conducting empirical investigation concerning the various treatment domains of a
CSP and client outcomes. Perhaps not by chance, these suggestions have come
from the various reviews of community treatment and case management. The
initial suggestion was generated by Test (1984) in calling for, "future research that
addresses the important specificity issues o f what kinds o f programs are most
beneficial to what kinds of patients under what conditions" (p. 369). Olfson
(1990) also concluded in his review of assertive community treatment that the
18
challenge o f future research would be in defining the conditions in which
community treatment recipients could be expected to demonstrate superior
functional outcomes. Similarly, Taube et al. (1990) have offered that it is essential
to design research that may be able to distinguish the differential effects o f
intervention (e.g., training in activities of daily living) with associated client
outcomes (e.g., independent living). Most recently, and perhaps most succinctly
from a research perspective, Solomon (1992) has stated that, "future research in
case management should couple process and outcome variables, so that it can be
determined which service elements appear to relate to which outcomes" (p. 177).
More specifically, Solomon calls for measurement of the nature, amounts, and
location o f actual services.
What appears to be most ironic is that to date the only published research
concerning treatment specificity has been the investigation o f Brekke and Test
(1992) that used the model guided method of program implementation (Brekke,
1987). It appears highly relevant as we enter a new age o f managed care and
funding cuts that specificity of treatment be fully explored in future investigations
of community programs to ensure recipients of care the most efficient,
comprehensive, and cost effective interventions possible.
Clearly, there needs to be agreement among researchers as to conceptual
definitions o f service variables. As will be demonstrated in the following section
operational definitions o f service variables are diverse, which leads to confusion
when comparing study outcomes. Until there is clear agreement upon definitions
19
difficulties in replication o f service delivery models and research findings that lead
to generalization to other settings and populations will remain difficult at best.
When these concepts are clearly defined and agreed upon only then will we be able
to determine what levels o f specific service elements are most beneficial to clients,
which in turn will have direct implications for future program development and
funding.
Empirical Analyses of Putative Service Variables Related to CSP Outcome
An exhaustive review o f the literature across both the MEDLINE and
PSYCH-LIT data bases revealed there have been few evaluations concerned with
service characteristics in CSP research. Existing studies included one that
evaluated client referral to three different levels o f day treatment by case managers
using hypothetical case vignettes (Berthaume, et al., 1994), one that investigated
compliance with referrals as an aspect o f continuity o f care; one that included
continuity o f care and service intensity as a secondary variable, two that
considered service intensity at the same location, and three that evaluated the
effect of intensity of case management services on client outcome (two of which
utilized the same sample. In an effort toward parsimony, this review will only
consider those studies which concerned in-vivo intervention.
In an early study o f continuity o f care Tessler and Mason (1979)
interviewed 146 participants in a study concerned with referral compliance
following state hospital discharge. Forty-six participants were referred to other
facilities (i.e., halfway houses, nursing home, courts, etc) and were not followed
20
after discharge, while the remaining 100 participants were referred to community
clinics (persons referred to the halfway house were not included because it was felt
that they were not in the community). Data were collected on 98 o f these 100
persons for 90 days following discharge to the community. While 54% of the 146
participants had a diagnoses of schizophrenia, no breakdown was offered for the
sample that was followed for the study period. Continuity was operationalized on
a continuous scale ranging from 0-3. A score of 0 was assigned if the participant
refused the referral, a score o f 1 if the participant accepted the referral but did not
make contact, a score of 2 if the participant made only one appointment, and a
score of 3 if the participant continued in treatment beyond the initial contact.
Their data revealed that 65% of the participants referred made at least one contact
and the majority kept at least two appointments. Surprisingly, it was noted that
there were no statistically significant differences in hospitalization among
compliant and non-compliant groups. Additionally, it was observed that persons
with a diagnosis of schizophrenia were highly compliant with aftercare.
In a later retrospective study that was primarily concerned with continuity
of care and participant outcomes, Tessler (1987) gathered data on 112 participants
(53 % carrying a diagnosis of schizophrenia) who were followed for 6 months
after discharge from a public mental hospital. Continuity of care was
operationalized as (a) number of services recommended but not received within 90
days o f discharge, (b) average severity of participant related breaks in services, and
(c) fulfillment o f a program-specific treatment plan. Service intensity was defined
as number o f services received. Participant outcome was defined in two ways.
First participants were rated on a five-point scale in terms o f their adjustment to
community living in six areas (civic behavior, social and recreational activities,
home/residential living, work or school, and activities of daily living) and an
aggregate score was computed. Second, outcome was also represented by number
of complaints received concerning a participants' behavior. The data revealed that
when participants did not receive recommended services within 90 days of
discharge outcomes tended to be poorer. Concomitantly, when services were not
received during that period more complaints were generated. The data also
demonstrated that participant related breaks in service did not achieve statistical
significance in multivariate analyses and it was suggested that this aspect of
continuity may be of less importance in comparison to other aspects of the
concept. Additionally, it was observed that continuity of care and service intensity
as defined in this study had essentially the same contribution in a multivariate
predictive model of outcome. On the other hand, a paradoxical effect was noted
concerning service intensity in that persons in this sample who received more
services also generated more complaints. An alternative explanation here may be
that persons who received more services exhibited a higher severity of illness and
required additional services. As such this may be a confound in the findings of
Tessler (1987). It should be reiterated here that the operationalization of intensity
in this study was number of services received, and not number or minutes of
contact.
Snowden, Storey, and Clancy (1989) have evaluated race as a factor for
premature treatment termination at a Community Mental Health Center (CMHC)
designed to serve a predominantly African-American clientele. Their sample
consisted o f varying diagnosis categories, among which 33% were diagnosed with
a schizophrenia related disorder. These authors found that while race was not
related to termination, it was related to service intensity which was defined as the
number and type o f contact. Additionally, they also found that amount of service
consumption (intensity) was related to adjustment and diagnosis. A concomitant
finding was that persons with a schizophrenia related disorder had a higher
tendency to drop out of treatment, which the authors suggested could be related to
the inadequacy of this treatment modality for such persons. In a subsequent
investigation Snowden and Clancy (1990) utilized the same data to evaluate the
impact o f service intensity. In this study intensity was more clearly defined as the
number of service units consumed. A service unit was defined as an individual,
group, crisis intervention, collateral, or medication contact. Findings indicated that
while service intensity had a positive effect upon participant outcome, there was a
non-linear relationship supportive of diminishing returns. It would appear to make
sense that there would be diminishing returns in intensity over time if the
assumption were made that participants improved from initial intense services, then
maintained treatment gains with a lower amount of contacts in the domain in which
they had improved. Additionally, a concomitant finding with this sample was that
when participant related control variables (social background, preparation and
23
motivation for treatment, and diagnosis) were implemented there was little effect
upon the relationship between service intensity and outcome.
In an investigation of service intensity (defined as number of contacts) and
outcome, Dietzen and Bond (1993) evaluated the impact of case manager contact
and outcome for 155 participants who had received a minimum o f 12 months of
assertive case management (ACM) services. Data were derived from four
previous studies. These authors used correlation and cluster analysis to evaluate
seven case management locations. Findings indicated there was wide variation in
the implementation of the ACM model. The data revealed that two locations had a
high reduction in hospital use, while two were moderate, and three had no
significant reduction. Most significantly the four locations with moderate to high
levels o f service intensity demonstrated moderate to high reduction in
rehospitalization. Dietzen and Bond (1993) concluded from these data that
programs must provide a minimum intensity of services to prevent hospital use.
These authors also used cluster analysis to group participants by differing types of
contact (e.g., high contact, office contact, in-vivo contact, indirect contact, low
contact). When groups were compared at outcome, no significant differences
were found on either change in hospital use, or participant satisfaction. In future
research endeavors, these authors suggested that, "a more comprehensive analysis
would have examined levels o f functioning and other variables" (p. 842).
Most recently, in an investigation of two ACM models implemented in the
Pennsylvania, Sands and Canaan (in press) have also operationalized service
24
intensity as the frequency o f contacts. While these authors reported there were
few statistically significant differences between the two ACM models, it is
interesting to note that participants that had higher contact and number o f services
had higher overall levels of functioning, were more compliant with medications,
and demonstrated a non-significant trend o f spending less days in the hospital.
While there has been much discussion as to the importance of continuity of
care few studies have investigated this concept longitudinally. On the other hand,
the CSP investigations that have studied this concept offer support for its
importance. Additionally, while there has been some discussion concerning
comprehensiveness of care only the investigation of Brekke and Test (1992),
which was designed primarily to test a model for measuring program
implementation, has specifically operationalized and included this variable.
Similarly, there has been little investigation o f service intensity, although a few
recent studies point toward the necessity o f assuring at least a minimal level of
intensity to avoid hospitalization. While we currently have emerging data
concerning these service characteristics it is requisite that further studies are
conducted to determine how these variables behave in relation to other indicators
o f client outcome.
Client Variables Related to Outcome in CSP Research
Jonsson and Nyman (1991) have noted that while there are few instances in
which causal relationships may be investigated in psychiatric research, the
confirmation o f personal characteristics that are connected with course o f illness
25
and outcome is requisite to the formulation o f causal hypotheses that may be
empirically tested and contribute to theory development. Additionally, the
identification o f such characteristics is of significance if the question concerning
what treatment, for whom, and at what time is to be answered (Test, 1992, 1984).
In recognition o f this, a review o f the literature was undertaken to identify
prognostic studies in the community treatment o f schizophrenia. The literature
revealed few published studies in community treatment, and surprisingly none in
CSP research, had been concerned with the relationship between client
characteristics and outcomes.
Not surprisingly, the studies identified offered similar finding to those of
Carpenter and Strauss (1991) and Strauss and Carpenter (1977, 1974, 1972). In
this group of studies these authors followed persons diagnosed with schizophrenia
or schizoaffective disorder in a longitudinal study of outcome. Their findings were
in agreement with others (McGlassan, 1988; Harding, 1988; Carpenter and
Kirkpatrick, 1988) that have indicated there is heterogeneity in both course and
outcome of the illness. Additionally, the data consistently exhibited autonomy
between the various outcome indices Strauss and Carpenter developed and lends
support to the concept of open ended and reciprocal systems, particularly in the
domain of outcome. This finding was also consistent throughout the duration of
the study period. In regard to prognostic variables, the data indicated that persons
with more frequent social contacts and more stable heterosexual relationships prior
to baseline admission demonstrated more favorable longitudinal outcome with
26
respect to social contacts, employment status, symptom severity, and total
outcome score. Second, short duration of hospitalization at index was associated
with more frequent social contacts and more stable employment. Lastly, the
presence and severity of thought disorder, delusional thinking, or hallucinations at
baseline were significantly related to symptom severity at 11 year follow-up.
Additional recent studies that considered client prognostic factors and outcomes
are discussed below.
A two year prospective design has also been utilized to examine the utility
of historical (demographic), genealogical, course of illness, and clinical dimensions
to predict outcome in a sample o f 37 persons with acute schizophrenia (Kay and
Lindenmeyer, 1987). Outcome measures were the Positive and Negative
Symptoms Scale (PANSS), Span o f Attention Test (SOA), SCOS, and length of
inpatient hospitalization for 18 months after baseline. Findings indicated that
demographic factors (i.e., age, gender, ethnicity, education) were not significantly
correlated with any o f the outcome measures. In contrast, the negative syndrome,
depression, psychomotor retardation, premorbid school functioning, sudden onset,
disposition since index, nonparanoid subdiagnosis, and family history o f alcoholism
account for between 51-87% o f the variance on 14 outcome variables. The
conclusions of these authors were congruent with the studies above in that they
suggest that prognostication in schizophrenia outcome is enhanced with a
multidimensional approach.
Prudo and Blum (1987) used data from the London site of the World
Health Organization (WHO) International Pilot Study o f Schizophrenia (IPSS). A
sample o f 100 participants was drawn and 5 year outcome data were analyzed.
Findings indicated there was heterogeneity in outcome for the sample. For
example, one-fifth experienced no further major mental disorder after index, while
one-tenth never recovered from the index episode. The data indicated further that
approximately half o f the participants had a good outcome. Regarding the
predictive utility o f the prognostic measures employed: clinical outcome was
explained by illness history, occupational functioning, social functioning, and
negative and non-specific symptoms; while social outcome was predicted by
occupational and social functioning, housing status, and negative and non-specific
symptoms. Similar to the findings of Strauss and Carpenter, these data indicated
that previous hospitalization, previous social contact, and previous employment
were the most efficacious means o f predicting 5 year outcome in this sample.
These findings were also similar to those o f Gaebler and Pietzcker (1987b) in that
social factors explained 32% o f symptomatic outcome. Prudo and Blum advise
such findings are also of clinical importance as they suggest the necessity o f
intervention in areas of social functioning which has been demonstrated by others
to be positively impacted by psychosocial treatment.
Verghese, Rajkumar, Sethi, and Trivedi (1989) also followed a cohort
included in the IPSS study in India for 2 years. These authors found that
medication compliance, length o f illness, change in level of religious activities,
28
dangerous behavior, agitation, and socio-economic difficulties were the factors
most predictive of outcome (of particular interest, the finding concerning
medication compliance has also been corroborated by the McCreadie et al. (1989)
who found that only 32% of clients with poor outcomes in a 2 year study o f first
episode schizophrenics received medication continuously over the trial period).
On the other hand, it should be noted these variables accounted for about 22% of
the explained variance in outcome. Additionally, these authors reported that
participants coming from a rural background had a better prognosis than those
from urban settings. Verghese et al. (1989) suggest this finding may be an
indicator that persons from rural settings have better social support and are better
tolerated.
In a retrospective study, Dion and Dellario (1988) classified 312 persons
diagnosed with schizophrenia into positive, negative, and mixed subtypes as
described by Andreasen (cited in Dion and Dellario). Results indicated there were
no significant differences across demographic and outcome variables such as
vocational status and number of hospitalizations between the groups. On the other
hand, there were significant differences between the groups in skills of daily living.
Concomitantly, the data also pointed to a favorable course and more positive
outcome and for persons with schizophrenia classified with positive symptoms.
These findings are also o f particular significance as it may lead to an enhanced
ability to target clients for specific interventions with greater efficiency.
In another long-term outcome study, Jonsson and Nyman (1991) predicted
14-17 year outcome in 110 consecutive first admission schizophrenic clients in a
Swedish sample. Data were gathered from National Health Insurance Records,
and interviews and questionnaires of clients and relatives. A discriminant analysis
categorizing clients into good and poor outcomes was calculated using 107
dichotomous predictor items that had been gathered at the time o f the index
admission. Nine variables were associated with good outcomes, while fourteen
were associated with poor outcomes. The prognostic variables that demonstrated
the highest predictive power were inclusive of premorbid functioning in social and
occupational domains, and symptomatology. Additionally, predictions of social
competence and total outcome were possible, while symptomatology was slightly
significant and economic self-sufficiency was not predictable.
Moller, Schmid-Bode, and von Zerssen (1986) have also investigated the
predictive validity of four prognostic scales in two long-term studies of
schizophrenic clients. These authors concluded that the Strauss-Carpenter
Prognostic Scale (SCPS) was the most efficacious in prediction o f long-term
outcome. Concomitantly, they suggest that a multi-dimensional approach such as
that utilized by the SCPS produces a better prognostic value than scales that
employ less comprehensive concepts, such as premorbid adjustment. Similarly,
Gaebel and Pietzcker (1987) found the SCPS to be superior in outcome prediction
in comparison to three other prognostic scales (Vaillant, Stephens, and Phillips,
respectively) when outcome was measured by the Strauss and Carpenter Outcome
30
Scale (SCOS) and the Clinical Global Impressions scale in a German sample.
Their findings supported the earlier findings of Strauss and Carpenter (1977, 1974,
1972) that were suggestive o f "open linked systems" of functioning and that
clinical outcome in a given area of functioning is related to previous levels of
functioning in that domain; these findings have also been corroborated more
recently by Carpenter and Strauss (1991). Gaebel and Pietzcker found that (a)
some individual items contributed to higher explained variance in the total 21 -item
score, and (b) the prognostic power of the SCPS was strongest in the area o f
social functioning. In particular, their data showed that quality o f social contacts
in the year before index had the most prominent significance for most of the
outcome criteria.
The relationship between psychopathological, psychobiological, and clinical
psychological variables and outcome has also been examined. Straube, Wagner,
Foerster, and Heimann (1988) found that psychobiological and clinical
psychological variables had the highest predictive power with respect to short-term
and medium-term outcome. Their data revealed that psychobiological variables
could explain approximately 70% of the variance in respect to days spent in the
hospital and relapse 2 years after index admission, while clinical psychological
variables explained 61% of the variance in 2 year relapse. The psychopathological
variables were found to be non-significant. Unfortunately, this study did not
examine the influence of the predictors in a stepwise regression analysis to
determine the unique contribution of each domain in respect to outcomes.
Similarly, Brekke, Raine, Ansel, Lenz, and Bird (in press) have investigated the
relationships between psychobiological, clinical psychological, and psychosocial
variables. Their data indicated significant correlations across these different
groups o f variables, as well as, a significant amount of variance explained in
several psychosocial functioning outcome indicators. Although these studies were
discussed here it should be noted that the variables used extend beyond the
parameters of the proposed investigation. Their utility is encouraged for future
investigations concerning prognostic factors and outcome.
From the review above it may be concluded that there is ample evidence
concerning the relationship of client characteristics and outcome. However, a
caveat is that few o f the studies identified examined these relationships when
treatment occurred in a community support program or in a case management
setting. Further research is needed to explore how client characteristics relate to
service characteristics. Additionally, while the proposed study will not investigate
psychobiological and clinical psychological variables, continued research using
these client specific characteristics is warranted given the evidence presented by
Straube et al. (1988) and Brekke et al. (in press). It is quite possible that these
indicators may offer the best method of predicting client outcomes across multiple
domains in future research.
Conclusions
The findings reviewed above offer support for CSP's in that they offer
support for systems o f care and the various modalities that constitute a CSP.
32
However, some of the evidence presented tends to lend equivocal support for
either CSP's or case management. It should be recognized that much o f the
equivocal nature can be attributed to the "lens" (e.g., outcome definitions) through
which one views these findings. For instance, if one is interested in providing care
in the least restrictive setting and quality o f life, it would appear there is more than
ample support for such programs. On the other hand, if one is concerned primarily
with cost effectiveness the findings are somewhat cloudy. Furthermore, much of
the cost effectiveness data is related to use of hospitalization as an outcome
measure, in spite of the fact that it is fairly well acknowledged that such a measure
is inefficient.
Another issue lies in the conceptualization o f service variables. While there
is agreement that these variables be measured there has not been consistent
agreement among authors on operational definitions. Concomitantly, while there
has been a call in the literature over the past decade to include these variables in
programmatic research, few studies have actually done so. Therefore, the need for
more programmatic implementation data continues and consensus agreement must
be reached on conceptual and operational definitions.
There are essentially seven broad and vastly important conclusions that can
be reached about CSP's. First, persons with severe and persistent mental illnesses
can survive in the community when supportive community intervention is available.
Concomitantly, community interventions should be individualized to meet a clients
specific needs rather than just "plugging" them into an available slot in a system of
care. Second, if this population is to adequately survive in the community there
must be an assurance o f continuity of care. The significance of this concept is that
(a) there must not be gaps in treatment, and (b) interventions must be continual
over a few years at the very least, not simply a few months. These conclusions are
based on: (a) the fact that clients experience poorer adjustment when there are
gaps immediately after hospital discharge (Tessler, 1987); and (b) the
preponderance o f evidence that indicates when interventions are stopped, gains
attributed to treatment and the ability to function in the community erode. Third,
although there is building evidence concerning the importance of service intensity,
the review above indicated we know very little about what levels of service
intensity are required to ensure positive client outcomes. Further research
concerning this concept is clearly warranted as such evidence has direct
implications for program funding. Similarly fourth, we know little about which
clients benefit from the different aspects of CSP intervention. Additional data is
necessary to answer this question. Fifth, it must be recognized that at this point in
time temporary hospitalization is an aspect of a continuum of care for this
population. Hence, short-term periods o f hospitalization should not be looked at
as treatment failure. This also points to the importance of using other more
appropriate indicators of outcome, such as psychosocial functioning,
symptomatology, and quality o f life in CSP evaluation research. Sixth, attempts
must be made to identify the specific effects of the various treatment domains in
relation to client outcomes. And lastly, as noted above agreement must be reached
on operational definitions of continuity of care (both cross-sectional and
longitudinal), comprehensiveness, and service intensity.
STUDY HYPOTHESES
Although this study was intended to be primarily exploratory in nature,
there are four highly relevant explanatory hypotheses that may be tested
concerning participant and service characteristics that were derived from the above
literature. The principal study hypotheses are as follows:
H I: Participant characteristics (e.g., gender, ethnicity, age at study entry,
prognostic indicators) will be predictive of various participant outcomes.
H2: Service characteristics (continuity of care, service comprehensiveness, and
service intensity) will positively impact various participant outcomes.
H3: Groups o f participant and service characteristics will be predictive of
various areas of participant functioning.
H4: Specificity o f treatment as measured by the intensity o f services in specific
service domains will be related to improvement in associated areas of
participant functioning.
Analyses
As noted above there has been a substantial call in the literature for future
research to combine both process and outcome variables to determined which
service elements are critical to which participant outcomes over time (Solomon,
1992; Holloway, McClean, and Robertson 1991; Olfson, 1990; Taube, Morlock,
Bums, and Santos, 1990; Test, 1984). Additionally, it appears important to look
at participant characteristics to determine their impact upon various outcomes in
community treatment. The model guided method of program implementation
suggested by Brekke (1987) and Brekke and Test (1992) was used to glean
knowledge from the data collected. For ease of review I have repeated the study
36
questions, listed the proposed statistics that were implemented in the analyses, and
listed the variables o f interest (see Figure 1 for definitions) that were utilized in the
analyses. The following analyses were conducted:
Explanatory Questions
1. Given participant characteristics, which participant variables have the most
significant impact upon various participant outcomes in community
treatment at 6 and 12 months?
Experimental Hypothesis: Participant characteristics will be predictive of
various participant outcomes.
Statistical Procedure: Multiple Regression.
Dependent (Outcome! Variables: Participant change scores in
symptomatology, psychosocial functioning, and subjective participant
experience.
Independent (Participant) Variables: Gender, ethnicity, age at study entry,
Strauss and Carpenter prognostic scale.
2. Given service characteristics, which service variables have the most
significant impact upon various participant outcomes in community
treatment at 6 and 12 months?
Experimental Hypothesis: Continuity o f care (longitudinal continuity),
comprehensiveness o f services (cross-sectional continuity), and service
intensity will positively impact various participant outcomes.
Statistical Procedure: Multiple Regression.
Dependent (Outcome! Variables: Participant change scores in
symptomatology, psychosocial functioning, and subjective participant
experience.
Independent (Service! Variables: Longitudinal continuity (number o f 30
day periods without staff-participant contact); cross-sectional continuity
(number o f different types of contact that account for at least 3% o f
contact time); service intensity (total frequency and duration of all
contacts).
37
3. Do groups of participant and service variables exist that predict various
participant outcomes in community treatment at 6 and 12 months?
Experimental Hypothesis: Groups o f participant and service characteristics
will be predictive o f various areas of participant functioning.
Statistical Procedure: Multiple Regression.
Dependent (Outcome) Variables: Participant change scores in
symptomatology, psychosocial functioning, and subjective participant
experience.
Independent (predictor) Variables related to Services: Longitudinal
continuity (number of 30 day periods without stafF-participant contact);
cross-sectional continuity (number o f different types o f contact that
account for at least 3% of contact time); service intensity (total frequency
and duration o f all contacts).
Independent (predictor) Variables related to Participants: Gender,
ethnicity, age at study entry, Strauss and Carpenter prognostic scale.
4. Given the various domains (i.e., living skills, vocational rehabilitation)
provided in this model CSP, which service specificity variables are
associated with corresponding participant outcomes in community
treatment at 6 and 12 months?
Experimental Hypothesis: Specificity of treatment as measured by the
intensity of services in specific service domains will be related to
improvement in associated areas of participant functioning.
Statistical Procedure: Pearson Correlation.
Outcome Variables: Participant change scores in treatment specific areas
o f functioning (i.e., work functioning, independent living, symptomatology,
psychosocial functioning, and subjective participant experience).
Specificity of Service Variables: Type of intervention measured by total
duration of all contacts.
Exploratory Questions
1. What is the nature o f the relationship between participant and service
variables in community treatment at 6 and 12 months?
Statistical Procedure: Pearson Correlation.
Participant Variables: Gender, ethnicity, age at study entry, Strauss and
Carpenter Prognostic Scale, baseline symptomatology measures.
Service Variables: Longitudinal continuity of care, total frequency and
duration o f all contacts, cross-sectional continuity.
METHOD
This study examined data concerning schizophrenic participants who were
admitted to a "model" CSP. As noted above, the data utilized for this dissertation
came from a model CSP that is part of a 3 year multi-site investigation where data
collection was concluded in September 1994. All participants in the study
reported here completed a minimum o f 1 year o f treatment at the Portals program.
The study focused on service and participant characteristics that are hypothesized
to impact treatment outcomes over the first year of the study. Additionally,
relationships between participant characteristics and service characteristics were
explored during the same period.
Participants
The participants in this study were a sub-sample of participants at a model
CSP from the larger multi-site study who had completed 1 year of treatment.
Participants included were from consecutive admissions to the CSP over an 8
month period, and all participants in the study were in the publicly-funded Los
Angeles County mental health system.
Since all participants in this study were in the L.A. County mental health
system, according to county policy they all had a County case manager assigned to
them. However, once a participant was admitted to Portals, the County case
manager turned all responsibility over to the agency, and their only contact from
that period was to ensure the participant was still in the Portals program. This
eliminated multiple treatment interference.
40
The participant selection criteria was as follows:
Diagnosis. Participants selected for the study met criteria for schizophrenia
or schizoaffective disorder according to Research Diagnostic Criteria (RDC;
Spitzer, Endicott, and Robbins, 1978). RDC criteria were selected because they
represent clear and explicit criteria of demonstrated reliability (Spitzer, Endicott,
and Robbins, 1978) and were a major influence in the development of DSM-III.
Residence. Participants must have been residents of Los Angeles for at
least three months prior to study admission.
Age. Participants must fall within the range of 18-60 years o f age.
Exclusions. Participants with mental retardation, organic brain syndrome,
or a primary diagnosis of substance abuse were be excluded.
Design
A quasi-experimental follow-along design was utilized in the multi-site
study from which data for the current analyses were derived. There was no
attempt to interfere with the normal admission/selection procedures to the
treatment programs. Repeated measures o f participant functioning were taken at
index admission, and every 6 months over a 3 years. A 15 year history was also
obtained from existing data. The current study used the first 12 months o f data on
subjects who were admitted to the Portals site. Treatment implementation data
were collected on a daily basis for all participants while they were in the Portals
program.
41
Entry into the Main Study
Entry into the study proceeded as follows: Within 2 days of admission into
Portals, all participants who met the initial selection criteria (diagnosis of
schizophrenia, schizoaffective disorder, or schizotypal personality disorder in their
charts) were informed by staff members about the study, and asked if study
personnel may talk with them. If the participant agreed to talk with study
personnel, the nature of the study was described to them, and they were asked if
they would agree to participate on a voluntary basis. Upon agreement to
participate and signing a consent form, the first interview was scheduled to be
within 48 hours of that time.
Diagnostic Interview
Within two weeks o f entry into the study each participant had a diagnostic
interview. Participants were diagnosed in this interview according to RDC criteria
using the Schedule for Affective Disorders and Schizophrenia (SADS; Endicott
and Spitzer, 1978) by a trained Ph.D. level clinician.
Description o f Treatment Program
The CSP selected from the multi-site investigation for this project was
Portals, which is a complex of old residences located in urban Los Angeles. At
Portals clients are admitted continuously to one or more of the program's four
separate rehabilitation components: the transitional living program; the apartments
program, the vocational training program , or the socialization program. All of the
components but the apartments program are located within one complex of
42
buildings. The apartments program is comprised of several satellite apartments
throughout the area surrounding the Portals complex. A description o f each
component is as follows:
Transitional Living Program (TLPT TLP is housed in one main building.
There are 14 beds available and the maximum stay is 4 months. There are six staff
who provide 24 hour coverage. The clients cook, clean, budget monies, and do all
the chores associated with independent living. The goal is to develop independent
living skills. Staff provide training, milieu structure, and case management
services.
Apartments Program (API. The AP is made up of several separate
apartments located in the area surrounding Portals. There are no on-site staff per
se. However, two Master's level clinicians housed in Portals provide on-site
services as needed, in addition to case management services. There are 34
apartment beds available, with clients living independently in pairs. The clients all
know one another and meet weekly for group sessions in one of the apartments.
The maximum duration of stay is 2 years. The staff monitor living conditions, as
well as client progress, and provide individualized services as necessary to maintain
client apartment living.
Vocational Training Program (VTPT This aspect of the Portals program
consists of several various components geared toward improving clients levels of
functioning. There are clerical, maintenance, and food preparation units located
on-site. These units comprise the first level of VTP intervention, and clients
43
assigned to them maintain Portals itself. Staff provide on-site training and
supervision. The second level is the volunteers program. Clients are assisted and
maintained in volunteer work in the public or private sector in the community. The
Corporate Cookie is the third level, which is a cookie store that is staffed by
Portals clients, and owned and operated by the program. Lastly, there is
competitive employment in the community.
Social Program fSPV The SP consists o f a social activity center and a
social club. Twenty to thirty clients are in the activity center at any given time.
The center provides structured daily activities for clients who are generally among
the lowest functioning at Portals. Two staff provide services at the center. On the
other hand, the social club is an evening and weekend service that houses dances,
games and generally provides a place for clients to socialize. There may be
upwards of 200 clients participating in the social club, and they do not need to be
formally admitted to Portals to do so. In fact, many come from board and care
facilities in the Los Angeles area.
Additional Characteristics. There are approximately 75 clients in the
various levels o f the program at any one time. There are 25 staff to provide
administrative services. Staff education ranges from Baccalaureate to Ph.D.
Services are provided on-site or in the community environment according to client
needs. Services are offered on an indefinite basis.
Sources of D ata
Interviews
Interviews were a primary source of data. All participants were
interviewed face-to-face by research staff at index admission (time of study entry),
and subsequently at 6 and 12 months. Instruments that were administered are
discussed below. An additional interview was conducted on a one-shot basis 1-2
weeks after the baseline interview to administer the Demographic Interview Form
(DIF; Test, et al., 1991; Stein and Test, 1980).
Training of Interviewer
Since the majority of outcome data came from face-to-face interviews, the
issue o f interviewer training was critical. The interviewer was trained to deliver
the Community Adjustment Form verbally, and to make clinical judgements for the
Brief Psychiatric Rating Scale. While the Index o f Self Esteem and Satisfaction
with Life Scale are all self-report measures, the interviewer w as also ready to assist
patients who had difficulty with them.
The majority o f the instruments used in the principal investigation have
been used at the Program for Assertive Community Treatment (PACT) research
site in Madison, Wisconsin for the past 15 years. The senior research associate
and interviewer from PACT trained the principal investigator and interviewer of
the multi-site project in the use o f these instruments.
To ensure that reliability was maintained during the conduct of the multi
site investigation, reliability checks were conducted every 6 months via simulated
45
interviews. In addition, the principal investigator also periodically re-interviewed
participants to check the reliability of the use of the instruments in the field. If
problems arose, they were discussed, and the interviewer was retrained were
necessary.
Blindness o f the Interviewer. It is essential that every effort be made to
minimize the possibility of interviewer or rater bias in the collection o f data. While
it was not possible to keep interviewers/coders blind to the treatment the
participants were receiving or to the programs they were involved in, the following
steps were taken to minimize bias. First, the interviewer hired was kept blind to
the real purposes of the multi-site study. The interviewer was told that the
purpose o f the study was to delineate patient problems that led them to contact an
agency, and to look over time at the participants new and recurrent problems, and
what services they received for them. Additionally, the interviewer was informed
that the programs involved were agencies that agreed to the conduct of the multi
site investigation, interview participants, and represent index admissions from
which to begin participant tracking. Second, contact between interviewer and
treatment program staff was kept to a minimum, although some contact was
necessary for coordinating participant interviews. Lastly, as was discussed earlier,
the instruments used and the training of the interviewer were designed to minimize
the need o f subjective interviewer ratings, therefore minimizing bias.
46
Measurement Instruments
The measurement instruments used in this study are presented below with
reference to their reliability and validity when it was available in published
literature. Issues related to interviewer reliability in the use o f interview-based
instruments have been discussed above and will not be reiterated here. The
specific instruments and the times when they were administered or used are
outlined in Figure 1. All instruments are described below.
Participant Characteristics and Outcome
The instruments described below were those given through research staff
interviews. The Demographic Interview Form was administered in a one-shot
interview 1-2 weeks after admission. All of the remaining instruments in this
category were administered at index and every 6 months throughout the study
period.
Demographic Interview Form (DIF't
This instrument has also been used in previous studies, and measures socio
demographic characteristics of participants and their psychiatric history, as well as
their work and social history of functioning prior to study entry. These data are
gathered using a semi-structured interview format.
Community Adjustment Form (CAF1
This is a major assessment instrument that has been developed, refined, and
tested psychometrically over approximately the past 20 years by other
investigators. Earlier versions were used in previous studies and were found to
47
discriminate between treatment and comparison groups (Test and Stein, 1980). A
preliminary reliability study (Johnson, cited in Brekke, 1988) has revealed high
inter-rater agreement. Brekke et al. (1993) reported the interobserver reliability
(kappa or intraclass correlation coefficient) ranged from 0.96-1.0 for the principal
sample. Procedures for reliability training and assessment have been discussed
above.
Information for the CAF is obtained in the course o f a semi-structured
interview with the participant. The measurement focus of the CAF is the
frequency with which certain objective behaviors or events occur over a specified
time period. The three primary areas drawn upon are participant functioning,
treatment received, and selected social/environmental factors.
Participant functioning variables assessed include time in institutions,
residential setting, employment, social relations, activities o f daily living, drug and
alcohol use and arrests. Treatment information assessed includes medication
usage, contacts with mental health providers, and participation in social and
vocational rehabilitation programs and activities. Social/ environmental factors are
inclusive o f contact with family and friends, social support, and income and
economic support. Data collected in some o f these areas represent a summation of
events over a 6 month period. On other data, the information is recorded
sequentially across the 6 month time period making these data continuous across
time, and allowing for the sequence in which events occurred to be maintained.
48
Quality o f Life Scale (OLSI
The intrapsychic deficit measure used in this study is the intrapsychic
foundations subscale o f the QLS (Heinreichs, et al 1984). This subscale was
selected because these intrapsychic deficits are seen as a core aspect of the
disorder o f schizophrenia, and the other functioning deficits are believed to emerge
from them (Heinreichs, et al. 1984). The interobserver reliability (ICC) on the
items ranged from 0.85-0.97 (mean=0.91); the alpha was 0.83.
Brief Psychiatric Ratine Scale (BPRSI
This is a widely used 22-item rating scale o f psychiatric symptomatology
developed by Overall and Gorham (1962) with established reliability and validity.
Interviewer training procedures were discussed earlier. The interrater reliability
(ICC) on the 22 items ranged from 0.74 to 1.0, with on average o f 0.92.
Cronbachs alpha for the present study was 0.80.
Satisfaction with Life Scale (SLSI
The SLS is an 18-item self-report scale that measures client satisfaction in
domains such as work, living situation, and number of friends. It has been used
previously by Stein and Test (1980), and was adapted from work done by
Fairweather et al. (1979). The instrument has also demonstrated utility in
quantitatively measuring subjective experience in community treatment of
schizophrenia (Brekke, et al., 1993). An earlier validity study done in Wisconsin
using TCL data (cited in Brekke, et al., 1989) revealed that it successfully
discriminated between client groups and "normals". Brekke (1988) indicated that
49
data from a pilot study revealed the scale discriminated successfully between
recent CSP admissions and first year graduate students. An alpha o f 0.89 for the
principal sample has previously been reported (Brekke, et al., 1993).
Index of Self-Esteem (ISE)
The ISE (Hudson, 1982) is a 25-item self-report measure that has been
designed for use in clinical evaluations. Brekke et al. (1993) also found the
instrument useful in measuring participant subjective experience. The ISE has a
clinical cut-off score, and its reliability and validity are discussed in Hudson
(1982). Brekke et al. (1993) reported an alpha of 0.87 for the principal sample in
using this instrument.
Measures Derived From Interview Schedules
The Global Assessment Scale, Role Functioning Scale, and Strauss and
Carpenter Outcome Scale ratings were derived from the CAF and BPRS data
using protocols developed and tested by the principal investigator and two
graduate assistants. Once the protocols were specified and tested, one additional
rater was trained and interrater reliability checks (using two or three raters) were
conducted. The reliability (intraclass correlations) ranged from .54 to 1.0, with a
mean of .91; 9 o f 10 possible correlations were over .9 (Brekke, 1992).
Respective scale scores were derived at index and every 6 months throughout the
study period.
50
Global Assessment Scale (GAS)
The GAS is a single rating scale for evaluating the overall functioning of a
participant on a continuum from psychiatric sickness to health (Endicott, Spitzer,
Fleiss, and Hertz, 1979). It specifies a time period for evaluation, which results in
the assignment of a single value from 1 to 100 on an anchored scale in increments
of 10 by the evaluator that indicates the participants respective level of functioning.
The GAS was selected because o f its high reliability and validity (Endicott, et al.,
1979).
Strauss and Carpenter Prognostic Scale (SCPS1
The SCPS is a 14-item scale that measures discrete dimensions such as
work history, social class, family history of mental illness, personal relationships,
symptom onset and use of institutions, precipitating events, and symptomatology
(Strauss and Carpenter, 1972). The instrument has been widely used as a
prognostic measure for over 20 years and has been found to be superior in
prognostication in comparison to other prognostic scales (Gaebel and Pietzcker,
(1987). For the purposes of this investigation, only the summed score will be
utilized rather than its discrete dimensions in consideration o f sample size and
parsimony.
Strauss and Carpenter Outcome Scale (SCOSI
The SCOS (Strauss and Carpenter, 1974) has four discrete dimensions,
each o f which is rated from 0-4 based on stated criteria for each. The four
dimensions are duration of non-hospitalization, social contacts, useful employment,
51
and absence of symptoms. A composite score may also be derived by obtaining a
sum for all dimensions. The variables of interest on the SCOS will be non
hospitalization and the composite score (also see below).
Role Functioning Scale (RFS1
The RFS is an instrument that was developed to complement the GAS
(Goodwin, Sewell, Cooley, and Leavitt, 1993). Green and Gracely (1987) have
suggested that the RFS is more relevant to a chronically mentally ill population.
The purpose of the RFS is to obtain ratings of client level of functioning in the
domains o f independent living/self care, work functioning, family relations, and
social functioning. The RFS rates each dimension from 1-7 based on an anchoring
description for each o f the seven levels. A composite score of these domains is
also derived by summation. Because of the breadth of the rating scale on the RFS
compared to SCOS, the various discrete domains o f the RFS will also be examined
as separate outcome variables.
Measuring Program Implementation
Daily Contact Log (DCL)
This is a form completed daily by each program staff member that has
client contact, or contact with other agency staff on behalf of the client. It
provides a record of staff contacts with clients during each day. Aspects of
treatment recorded are whether or not a contact occurred, the place o f contact,
who was involved, whether it was a group or individual contact, duration in
minutes, and content o f the contact (i.e., psychotherapy, case management,
52
medication, social-recreational, training in activities of daily living, employment
training, etc.) The results of the use o f this from are reported in Brekke (1987),
and Brekke and Test (1992, 1987). The DCL has been used with high inter
observer reliability and validity (Brekke, 1987).
As shown in Figure 1, the following variables are derived from the DCL
and will be used as primary service variables:
Continuity o f Care. The number o f gaps in service will represent the
concept o f longitudinal continuity of care (Brekke and Test, 1992), Gaps in
service will be measured by counting the number of 30 day periods in which there
was not any contact between program participants and treatment staff.
Treatment Intensity. Brekke and Test (1992) have suggested that
treatment intensity may be measured in two distinct ways. First, in this study
intensity will be represented by measuring in actual minutes the duration of each
different type o f service contact that occurs between treatment staff and program
participants. The duration o f all contacts will be summed to obtain the total
duration o f treatment. Second, intensity will also be measured by counting the
actual number o f service contacts that occur between program participants and
treatment staff. The total frequency will then be obtained by summing the
occurrence o f all contacts over time.
Treatment Comprehensiveness. Treatment comprehensiveness (cross-
sectional continuity of care) will be defined as the number o f different types of
services that program participants received (Brekke and Test, 1992). Specifically,
53
comprehensiveness will be measured by counting the number of different types of
services that account for at least 3% of total contact time.
54
RESULTS
Sample Characteristics
Sample Size
Results from a study that was conducted concurrently with this dissertation
indicated that study attrition for the multisite investigation was 12% after 1 year
(Brekke, Long, and Nesbitt, under review). Treatment attrition, however, was
somewhat higher than anticipated. The data indicated that treatment attrition at
Portals was 19% after 6 months and 45% after 12 months.
To compensate for treatment attrition, rules were developed in
collaboration with the principal investigator for consistency with future studies
utilizing these data. For sample inclusion in the present study participants must
have had contacts through at least 75% of the total contact time from index (i.e.,
136 days in treatment for the 6 month and 274 days in treatment for the 1 year
samples). Similarly, some participants did not have complete interviews or were
missing an interview entirely. There was also one participant that did not have
any recorded contacts during the first 181 days of treatment, but was consistently
seen during the second half of the year and was included in the 1 year analyses.
Lastly, a few participants that were included in the 6 month sample dropped out of
the 1 year sample. After accounting for missing interviews the final samples sizes
were 41 participants for the 6 month sample, and 30 participants for the 1 year
sample.
Demographics
As can be seen in Table 1, the demographic characteristics o f participants
varied minimally from the 6 month sample to the 1 year sample. Study participants
were predominantly young males, who scored in the low-middle end of the
spectrum on the SCPS at index. Table 2 offers descriptive data regarding baseline
levels of functioning. High scores on the BPRS represent higher symptoms, while
higher scores all other measures represent higher levels of functioning. These data
show that participants were moderately symptomatic, and had relatively low levels
o f psychosocial functioning based on the range of possible scores on those
instruments. In term s of self-esteem and satisfaction with life, participants scores
were in the mid-range on those measures. As in the case o f demographic
characteristics, these data indicate there was virtually no variation between the 6
and 12 month samples in terms o f particpant levels o f functioning. Therefore, the
analyses at 6 and 12 month outcome were conducted using samples with nearly
identical characteristics.
Characteristics of Service Delivery
As shown in Table 3, participants received an average o f 93.2 contacts
(SD=74.4) with a mean duration of 31 minutes (SD=9.4) during the first 6 months
o f treatment. The mean frequency of contacts for the entire 12 month study
period was 125.4 (SD=100.5), while the mean duration of all contacts was 30.0
minutes (SD=8.6). These data indicate that participants received nearly 75% of
services during the first 6 months of treatment. This amounts to nearly three times
56
as much service in terms of both total frequency and duration over the first 6
months o f the study, as compared with the latter 6 month period.
The intercorrelations between service process variables are offered in Table
4. These data indicate significantly high levels of association between these
variables. Essentially, participants with a high total duration of contact had both a
higher frequency of contacts and more comprehensive services, with fewer gaps in
service continuity. As shown in Table 4 these relationships intensified at 12 month
outcome. Concerns regarding the use of these variables in model building will be
addressed in a subsequent section.
Detailed data that describe duration and frequency of contacts by type are
offered in Table 5 (please note that some participants who were not in the 6 month
sample because o f missing data at 6 month outcome were included in the 1 year
sample which may account for differences observed in ADL contacts). These data
demonstrate that service delivery to study participants in this setting was most
intensive in the areas o f vocational rehabilitation, activities of daily living, and
supportive individual contact for the first 6 months of the study. The data indicate
further that frequency of service delivery remained highest in these areas during the
second 6 months of treatment, but that there was also a marked increase in total
frequency and total duration o f case management and social recreation contacts for
that period.
Service Process Characteristics and Baseline Participant Functioning
57
Correlations between baseline levels o f functioning and service process
variables at both 6 months and 12 months were conducted to determine if program
staff were targeting participants for services based on participants respective
individual levels of functioning at index. These data are presented in Table 6. At 6
months the data indicated that higher baseline ratings on the RFS independent
living scale were associated with receiving less comprehensive services. Higher
baseline levels of work functioning on the RFS were also associated with receiving
a higher frequency of treatment contacts during the first 6 months o f program
participation. The data also indicated that higher baseline RFS social functioning
ratings were associated with receiving services that were more comprehensive
during that period.
On the other hand, the 12 month data indicated that higher ratings on the
GAS at baseline were associated with receiving a lower total duration of treatment
over the course of the study. A higher self-reported ISE rating was also associated
with receiving a higher total duration o f treatment after 12 months. These findings
appear to indicate that staff were not favoring participants with high levels of
baseline functioning.
Effectiveness o f Services
In general, Portals participants have been found to demonstrate superior
outcomes in comparison to (a) an alternative CSP, and (b) board and care with
standard case management (Brekke, Long, and Nesbitt, under review). While not
a focus o f the present study, paired t-tests were used to test the efficacy of the
58
treatment model on all outcome indicators of participant functioning in the present
study. One-tailed tests o f significance were used. These results are offered in
Table 7. After 6 months o f treatment the outcome data indicated significant
improvement in both clinical and psychosocial functioning. Additionally, change
was in the desired direction in the domain of subjective experience although
findings were non-significant. The 12 month outcome data also demonstrate
significant improvement in the clinical, psychosocial, and subjective experience
domains. However, it is important to note that change in the BPRS was not in the
desired direction at both 6 month and 12 month outcomes. The change score
results on the BPRS indicated that participants experienced an increase in
symptoms over the study period.
Level of Association Between Participant Characteristics
and Change Score Outcomes
In all analyses on participant outcomes, change scores were generated from
baseline to 6 months and from baseline to 1 year. These change scores were the
outcome variables used in all subsequent analyses. Participants race
(Caucasian/non-Caucasian) and gender (male/non-male) were dummy coded.
None of the participant characteristics were found to be significantly
intercorrelated with other participant characteristics in either the 6 month or 1 year
samples. For all analyses between participant characteristics and change score
outcomes 2-tailed tests o f significance were used.
59
At 6 month outcome, the only significant correlation noted was between
age at study entry and improvement in satisfaction with life (r=.41, p=008). These
data indicate that older participants were more satisfied with their station in life
after 6 months of treatment. Correlations between participant characteristics and 1
year change scores are offered in Table 8. The associations noted at 6 months
between age at study entry and improvement in satisfaction with life was not
maintained over the course of the 1 year study period. However, these data
indicate that being male and younger at study entry were associated with
improvement in independent living as measured by the RFS. Being male was also
associated with improvement in RFS global functioning.
Levels o f Association Between Service Process Variables
and Change Score Outcomes
Table 9 presents the correlations between service process variables and
participant change score outcomes at 6 and 12 months. One-tailed tests of
significance were used for these analyses as it was expected that service delivery
would positively affect outcomes, while gaps in service were expected to have a
negative impact on outcome. The 6 month findings indicated that fewer gaps in
services were associated with more improvement in SCOS use of institutions,
GAS, and the RFS global functioning and independent living scales. Higher total
duration and frequency of services were associated with greater gains at 6 month
outcome on those measures. Higher comprehensiveness o f service was also found
to be significantly related to improvement in RFS independent living. No
60
significant relationships were noted between service process variables and
improvement in subjective experience.
Levels o f association between service process variables and participant
change score outcomes at 12 months are also found in Table 9. These data offer
additional evidence of the association between service process variables and
participant change score outcomes in both clinical and psychosocial domains.
After 1 year o f treatment fewer gaps in service were associated with less o f an
increase in symptoms on the BPRS, and more improvement on the QLS Deficit
Syndrome Subscale, SCOS use o f institutions and total scores, GAS ratings, and
RFS total and work functioning scales. Conversely, a higher total duration and
frequency of services were associated with less of an increase in BPRS
symptomatology, and more improvement in terms of SCOS total and use of
institutions scores, GAS ratings, and RFS total, work, and independent living
scales. As in the 6 month analyses, there were no significant relationships between
service process characteristics and subjective experience variables.
In summary, these data indicate that service process variables are clearly
associated with participant improvement at both 6 month and 12 month outcome.
The general pattern in the data indicates that participants who experienced higher
gaps in services had less improvement on clinical and psychosocial measures, while
participants that were the recipients of a higher total duration and frequency of
treatment experienced greater gains in clinical and psychosocial functioning.
Exploratory Analyses
Participant Characteristics and Service Process Variables
Pearson Product Moment Correlations were also used to explore levels of
association between participant characteristics and service process variables at
both 6 month and 12 month outcomes. As shown in Table 10, the 6 month data
show a significant negative association between the SCPS and gaps in service after
6 months o f service delivery. These data suggest that participants who had a
better prognosis experienced fewer gaps in service.
In addition, a significant negative association was noted between race and
gaps in service. This finding suggests that Caucasian participants had fewer gaps
in service than non-Caucasians over the first 6 months o f treatment. On the other
hand, none of the associations noted between participant characteristics and
service process variables were noted in the 12 month data. Therefore, these
particular findings should be approached with caution until such time that these
analyses are replicated with a larger sample where statistical manipulation of
ethnicity is possible.
Hypothesized Analyses
Participant Characteristics and Change Score Outcomes
Because these analyses were theory driven, the analytic strategy
implemented utilized a two-stage multiple regression procedure to determine the
relative importance o f hypothesized participant characteristics. In the first stage all
hypothesized participant characteristic variables were entered into the equations
62
using forced entry, while the second stage utilized only those variables that were
noted with t-score probability values less than or equal to . 10 on the first stage. If
only one or no participant characteristic variables were noted with t-score
probability values less than or equal to . 10 on the first stage, multivariate analyses
for that outcome indicator were terminated. This process was repeated on all
outcome indicators for both the 6 month and 1 year samples. For parsimony, only
regression models that approached or were statistically significant will be reported.
R e s u l t s
At 6 month outcome no multivariate models were found to support the
experimental hypothesis in either the clinical, psychosocial, or subjective
experience domains. However, partial support for Hypothesis 1 was generated at
12-month outcome in terms o f improvement in the psychosocial domain. As noted
in Table 11, being younger at study entry and being male were significant (p=.02)
in explaining 26% of the variance in improvement in independent living as
measured by the RFS. In this explanatory model gender demonstrated slightly
more relative importance than age. This was the only statistically significant model
that was found when participant characteristics were used to explain change score
outcomes.
Service Process Variables and Change Score Outcomes
The two-stage analytic strategy discussed above was discarded to
determine the impact of putative service process characteristics on participant
outcomes because of the high intercorrelations between service process variables
63
that were noted in Table 5. These high correlations are an indication o f
multicollinearity. Because o f the near perfect correlations observed between total
duration and frequency at both 6 month and 12 month outcomes, and because both
variables measure the same concept (intensity) these service process variables are
essentially interchangeable. Therefore, total duration w as selected fo r inclusion in
all ensuing service process analyses as it is (a) a finer measure of the concept of
treatment intensity and (b) to decrease the effects of multicolinearity . An
additional reason to select one measure o f intensity was related to the relatively
small sample size (N=30) at 12 month outcomes. From a conservative
psychometric standpoint sample size would preclude entry of more than 3-
variables into a regression equation. Concomitantly, this decision would also
ensure testing stability of the conceptual model across the 6 month and 12 month
samples.
Additional difficulty was also encountered with high multicolinearity
between the remaining service process variables. Significant negative correlations
were also noted between duration and gaps in service at both 6 month and 12
month outcomes. Darlington (1990) and Lewis-Beck (1980) have noted that in
such cases the predictors used in multiple regression analyses may fail to obtain
statistical significance, although the predictors may actually be associated with the
dependent variable in the tested sample. Essentially, the intercorrelated predictor
variables cancel the effects o f each other. Unfortunately, there are no clear-cut
solutions to this problem. Darlington and Lewis-Beck indicate that while it makes
empirical sense to increase sample size to thwart the effects of multicolinearity, this
does not make practical sense in applied social science research when samples are
fixed. On the one hand, Lewis-Beck and Darlington both suggest a possible
solution to multi-colinearity is to eliminate an offending variable and to conduct
the analyses with the remaining variables. On the other hand, Lewis-Beck suggests
a more conservative strategy and proposes the preferred method is to substitute
the remaining offending variable for the first and to re-estimate the equation.
These results are then compared with the hypothesized equations, allowing for
better assessment of the relative importance of the independent variables.
Therefore, in an effort toward conservatism and to glean the most
knowledge from these data the following analytic strategy was developed and
utilized on all change score outcome indicators at both 6 months and 12 months.
First, the multiple regression equations were estimated using forced block entry
and the three originally hypothesized service process variables (total duration, gaps
in service, and comprehensiveness) were entered into the equation as independent
variables. Second, the equations were estimated again using only total duration
and comprehensiveness as independent variables. Finally, the equations were re-
estimated a third time using only gaps in service and comprehensiveness as
independent variables. It was only in this cautious manner that the saliency of
explanatory variables could best be estimated.
Because these are the first outcome analyses that used the concepts of
continuity o f care, treatment intensity, and service comprehensiveness as
65
explanatory variables the Type I error rate on the F-test of significance was relaxed
to . 10 to enhance knowledge building. In models where the three-variable
equation was significant, that model is presented along with significant two-
variable estimates so that the explanatory power o f independent variables may be
more clearly identified. Additionally, due to the exploratory nature of these
analyses, significant two variable estimates that were observed in attempting to
limit the effects of multicolinearity are offered. However, it should be noted that
these two variable solutions may be overspecified. Therefore, these findings
should be approached with extreme caution.
The significant models that explain change score outcomes are presented
within their conceptual domain (e.g. clinical, psychosocial, or subjective
experience). Within each respective domain the 6 month findings are offered, and
are then followed by the 12 month findings.
Clinical Change Score Outcomes
6 Month Outcomes. Partial support was found for the experimental
hypothesis at 6 month outcome. As shown in Table 12 these results indicate that
the number o f gaps in service and comprehensiveness approached significance
(p=.09) in explaining 12% o f the variance in use o f institutions as measured by the
SCOS. This finding suggests participants with fewer gaps in treatment and less
comprehensive services experienced more improvement in use of institutions at 6
month outcome. In this model gaps in service was the only significant variable in
terms of explanatory power.
12 Month Outcomes. Further support for the experimental hypothesis was
generated at 12 month outcome. As shown in Table 13, more gaps in service,
lower total duration, and more comprehensiveness were significant (F= 3.12;
p=.04) in explaining 27% of the change score variation in the BPRS. The two
variable estimates show that gaps in service and comprehensiveness were
significant (F=4.7; p=.02) in explaining this variation, with only gaps in service
contributing significant explanatory power. The two variable estimate that
included total duration and comprehensiveness was also significant (F=3.0; p=.07)
in explaining BPRS outcome variation, with total duration retaining the most
relevance in that 2-variable model. Therefore, it appears that gaps in service was
the most relevant variable in terms o f explaining outcome variation on the BPRS at
12 month outcome, while total duration offered limited explanatory power, and
comprehensiveness did not offer a statistically significant contribution.
Concomitantly, as the data on program effectiveness indicated, participants
showed an increase in symptoms at the end of the 12 month study period when
measured by the BPRS. Because of the positive direction of the slope o f gaps in
service and the negative direction of the slope of total duration, a reasonable
inference from these data is that increased levels of symptomatology may be
attributed primarily to experiencing more gaps in service, although a low total
duration of treatment may be o f secondary importance.
Regarding use of institutions, a significant two variable equation was also
noted concerning improvement in SCOS use of institutions at 12 month outcome.
67
As shown in Table 14, the data revealed that total duration of treatment and
comprehensiveness were significant (F=2.6, 09) in explaining 16% o f the
change score improvement on the SCOS use o f institutions scale. These data
suggest that participants that received a higher total duration of treatment and
more comprehensive service experienced more improvement. In this model total
duration o f service was the most salient in terms of explanatory power.
Psychosocial Change Score Outcomes
6 Month Outcomes. Partial support for the experimental hypothesis was
also generated in the psychosocial domain at 6 month outcomes. As shown in
Table 15, a higher total duration and less comprehensiveness approached
significance (F=2.7; p= 08) in explaining 13% o f the change score variation in
improvement on the GAS. In this model total duration showed the most relative
importance.
Table 16 presents a significant three-variable model (F=2.2, p=.10) that
explains 15% of the variation in improvement in the RFS global functioning scale
at 6 month outcome. On the one hand, none o f the variables in the three-variable
model initially appear to have significant t-scores. On the other hand, this model is
an example o f the effects of multicolinearity discussed above. Further examination
o f the two-variable models indicated that more gaps in service and better
comprehensiveness were significant (F=2.8, p=.08) in explaining 13% o f outcome
variation, with gaps in service retaining the explanatory power. The model that
included total duration and comprehensiveness generated a non-significant model.
68
Therefore, these data suggest that participants that experienced fewer gaps in
services demonstrated more improvement in RFS global functioning.
The data also produced a significant 3-variable explanatory model
regarding independent living as measured by the RFS at 6 month outcome (Table
17). This solution was significant (F=10.0, p=.0001) in accounting for 45% o f the
variation the RFS independent living change score, with total duration showing the
most relative importance in that model. Additionally, this finding is supported by
the two variable equations which showed that total duration and
comprehensiveness alone were significant (F=13.6, p<0001), with only total
duration contributing to the explanatory power in that model. The model that
combined gaps in service and comprehensiveness was statistically significant
(F=7.8, p=.001), with gaps in service retaining the explanatory power. Therefore,
although somewhat nebulous, it seems that receiving a higher total duration o f
treatment is the most important variable, followed by experiencing fewer gaps in
service in explaining improvement on the RFS independent living scale at 6 month
outcome.
12 Month Outcomes. At 12 month outcome three 3-variable equations
were found to support the experimental hypothesis in terms of SCOS total score,
GAS, and RFS total score outcomes. As indicated in Table 18, receiving a higher
total duration, with fewer gaps in service, and less comprehensiveness were
significant (F=2.6; p=.07) in explaining 23% of the variation in improvement in the
total SCOS score, with comprehensiveness appearing to be most relevant in this
equation. Again, further examination o f the 2-variable equations suggest this
finding may be attributed to the effects of multicolinearity. Also shown in Table
18 are the 2-variable equations. These data also clearly demonstrate the cancelling
effect o f multicolinearity on contributing explanatory variables. From these
additional data it can be seen that a higher total duration and less
comprehensiveness were significant (F=4.0; p=.03) in explaining 23% of the
improvement in the SCOS total score, with total duration showing the most
relative importance in the model. Additionally, gaps in service and
comprehensiveness were significant (F=2.5, p=. 10) in explaining 16% of the
variation in SCOS total score improvement, with gaps demonstrating more relative
importance than comprehensiveness in that model. Therefore, in consideration of
the findings of the two-variable models, it appears that receiving a higher total
duration o f service is the most relevant variable, while experiencing fewer gaps in
service demonstrated secondary relevance in explaining improvement in the SCOS
after 1 year of treatment.
An explanatory model for improvement in the GAS was also generated by
total duration, gaps in service, and comprehensiveness. These findings are offered
in Table 19. This model was significant (F=3.4, p=.03) in explaining 29% o f the
variation in improvement on the GAS. However, as in the case of the 6 month
RFS global functioning findings above,'it was not possible to clearly detect the
relative importance o f predictors due to non-significant t-values for each process
variable. On the other hand, further examination of the two variable models
70
indicated that experiencing a higher total duration and more comprehensiveness
were significant (F=5.1, p=.01) in explaining 27% of the variation in GAS
improvement, with only duration demonstrating relative significance in that model.
Additionally, experiencing fewer gaps in service and more comprehensiveness
were also significant (F=3.9, p=. 03) in accounting for 23% o f the variation in GAS
improvement, with only gaps in service retaining significant explanatory pow er in
that model. Therefore, these data appear to indicate that receiving a higher total
duration of treatment with fewer gaps in service, respectively, carry relative
importance in explanatory power in GAS improvement scores.
Table 20 presents a significant (F=2.6, p=.07) 3-variable model that
appears to explain 23% of the improvement in global functioning on the RFS at 12
month outcome. In that model total duration appears to carry most of the
explanatory power. This finding was also supported by the 2-variable equations,
which indicated that a higher total duration and more service comprehensiveness
were significant (F=4.0, p=.03) in explaining 23% of the variation, with only a
higher total duration contributing significant explanatory power in terms o f RFS
global functioning. The model that included gaps in service and
comprehensiveness produced non-significant findings. Therefore, it appears that
receiving a higher total duration o f treatment is the principal explanatory
component in explaining improvement variation in RFS global functioning.
As shown in Table 21, a two variable model inclusive o f gaps in service and
comprehensiveness was significant (F=3.2, p=.06) in explaining 19% of the
variation in work improvement on the RFS. From these data it appears that having
fewer gaps in service was the most relevant factor in regard to explaining
improvement in work functioning after 12 months of service delivery.
Subjective Experience Change Score Outcomes
While service process variables generated numerous multivariate
explanatory models in the clinical and psychosocial domains, there were no models
o f significance that emerged when improvement in participant subjective
experience was the outcome variable.
Service Process Variable and Outcome Summary
In summary, the data support provide partial support for the experimental
hypothesis. This support is found in both the clinical and psychosocial domains. It
also appears that for the most part the findings are more robust at 12 months,
rather than at 6 month outcomes. In terms of symptomatology, it appears that
gaps in service contributes the most power in explaining change score variation on
the BPRS. Regarding use o f institutions, gaps in service also explains the variation
in improvement at 6 months, while a higher total duration contributes the most
explanatory power at 12 month outcome. Additionally, it appears that total
duration of services may be the most efficient variable in explaining improvement
in the domain o f psychosocial functioning. The only exception to this was a two
variable model that suggested that gaps in service explained significant variation in
improvement in RFS work functioning at 12 month outcome. Lastly, as in the
case of participant characteristics no multivariate models emerged at either 6 or 12
72
month outcomes that lend support for the experimental hypothesis in terms of
improvement in subjective experience.
Participants Characteristics. Services Process Variables.
and Change Score Outcomes
In an effort to build upon the previous analyses and to develop a more
comprehensive predictive model, further analyses were conducted using both
participant and service characteristics. For those instances in which a participant
characteristic(s) and service characteristic(s) both had a significant t-score on a
given outcome indicator in any of the previous analyses, those combinations of
variables were entered in a single multiple regression equation. Because of the
issue o f multicolinearity noted above, in circumstances where both total duration
and gaps in service were significant in 2-variable models, the variable with the
most significant t-score was selected for use in this phase of model building.
Forced block entry was also utilized for these analyses. Significant findings were
observed in both the clinical and psychosocial outcome domains using combined
models. In comparison to the models offered above, the combined models offer an
improvement in explanatory power in most cases. These results are offered below.
Clinical Change Score Outcomes
Although there was no support in the 6 month data for a more
comprehensive model in the clinical domain, the 1 year findings offered partial
support for the experimental hypothesis that participant and services characteristics
are predictive of clinical outcome in the community treatment o f schizophrenia.
Improvement in the SCOS use of institutions was the only clinical outcome
indicator to offer support. As noted in Table 22 being younger at study entry and
a higher total duration o f contact was significant (F=5.6, p=.009) in explaining
29% of the variation in improvement use of institutions at 1 year outcome. This
equation indicates that being younger at study entry and receiving a higher total
duration o f contact contribute significantly to explaining outcome variation in use
of institutions. In this model both variables contributed significant explanatory
power, although total duration showed slightly more relative importance than age
at study entry.
Psychosocial Change Score Outcomes. Partial support for the
experimental hypothesis was also noted in the data in terms o f explaining variation
in improvement in RFS independent living at 6 month outcome. These data are
offered in Table 23. It should be noted that this was the only outcome indicator
producing a significant combined model in the 6 month data in any domain. The
results indicated that race and total duration of all contacts were significant
(F=15.0, p=.0001) in explaining 44% of the variation in the change score on the
RFS independent living scale. These data suggest that being Caucasian and
receiving a higher total duration o f treatment contributed to explaining the 6 month
outcome variation in terms o f improvement in RFS independent living. In this
model only total duration offered a significant contribution in explaining variation
in the change score.
74
In terms of improvement at 1 year psychosocial outcomes, two multivariate
models were found to explain improvement in global levels of psychosocial
functioning. These data are presented in Tables 24 and 25. The first model (Table
24) pertains to improvement in the GAS, in which both a higher total duration o f
all contacts and being of younger age at study entry were significant (F=8 .2,
p=.002) in explaining 38% o f the variation in improvement at 12 month outcome.
In this model both variables contributed significant explanatory power, with total
duration explaining more variation in the model than age at study entry.
Additionally, as shown in Table 25, a higher total duration of all contacts and being
younger at study entry also produced a significant model (F=6.6, p=.005) that
accounted for 33% of the variation in improvement in the RFS total score. While
both total duration and age at study entry were significant in explaining outcome in
each o f these models, it should be noted that total duration was slightly more
significant in both cases.
Specificity of Treatment and Participant Change Score Outcomes
Pearson Product Moment Correlations were used to test study hypotheses
at both 6 months and 12 months that were related to specificity o f treatment. One
tailed tests o f significance were used as the assumption is made that services will
positively impact participant outcomes. The specific treatment contacts that were
measured included total duration of: (a) independent living contacts (represented
by both ADL and living arrangement contacts), (b) social-recreation contacts, and
(c) work/vocational contacts. Two non-specific treatment factors, individual
75
supportive contact and case management contacts, were also included in these
analyses. The outcome variables utilized were the participant change scores that
were used in the analyses above. The 6 month findings are offered in Table 26,
while the 12 month results are presented in Table 27. Because o f the exploratory
nature o f these data, the matrices also include significant correlations between
specific hypothesized areas of service delivery and various other outcomes in the
three domains measured. As these are the first analyses to explore the association
between type of service and outcome, it is particularly interesting to note the
significant levels o f association between the total duration of various services and
participant improvement in divergent areas o f functioning.
Independent Living Contact
Partial support was generated for the experimental hypothesis in terms of
total duration of independent living contact at both 6 month and 12 month
outcomes. As indicated in Table 26 the 6 month data show significant positive
associations between a higher duration of ADL contact and improvement on the
RFS independent living scale. A higher duration of ADL contact was also
significantly associated with improvement on the RFS global functioning scale,
although this may be a statistical artifact resulting from the high association
between ADL contacts and the RFS independent living scale component.
Additionally, the findings indicated that a higher duration of living arrangement
contact was significantly associated only with improvement on the RFS
independent living scale at 6 months. Therefore, these findings suggest that
76
participants with more independent living contact after 6 months of treatment
experienced more independent living, and possibly better global functioning.
As shown in Table 27, the associations noted between a higher duration of
ADL contacts and improvement on the RFS global and independent living scales
were maintained at 12 month outcome. A higher duration o f ADL contact was
also significantly related to improvement on the GAS and the RFS work
functioning scale after 1 year of service delivery. Numerous significant
associations were also noted between a higher total duration of living arrangement
contact and improvement in both clinical and psychosocial outcomes at 1 year.
After 1 year o f service delivery higher living arrangement contact was significantly
associated with participants showing improvement in the deficit syndrome, SCOS
use of institutions and global functioning, GAS, and RFS work and independent
living functioning. Consequently, these data suggest that more independent living
contact over the total study period was related to improved clinical and
psychosocial functioning.
Social Recreation Contacts
As shown in Table 26, the 6 month findings offer additional partial support
for the experimental hypothesis in terms of total duration o f social/recreational
contacts. These data indicate a significant positive association between a higher
total duration o f social/recreational contacts and improvement in RFS social
functioning. The 6 month findings indicate further that higher social recreation
contact was also significantly related to improvement in both the clinical and
77
psychosocial domains. These additional data show that more social recreation
contact was significantly associated with improvement in SCOS use o f institutions,
GAS ratings, and RFS global and independent living scales. These findings
suggest that more social recreation contact was related to less use of institutions,
better global functioning, and more independent living after 6 months o f treatment
At 12 month outcomes (Table 27) the association noted between a higher
total duration o f social/recreational contact and RFS social functioning outcomes
was not significant, although the data continue to indicate significant positive
associations with improvement in SCOS use o f institutions and the GAS.
Additionally, a higher total duration of social/recreation contact was negatively
associated with the BPRS. This particular finding suggests that more
social/recreation contact was related to less o f an increase in symptoms, better
global functioning, and less use o f institutions over the 1 year study period.
Work/Vocational Contacts
Surprisingly, work/vocational contacts were not found to be significantly
related to improvement in work functioning at either 6 month or 12 month
outcomes. Nevertheless, it is interesting to note that a higher duration o f work
contact was significantly related to improvement in both the RFS global and social
functioning scales at 6 months. These relationships became stronger and continued
at 12 month outcome, as well as, more work/vocational contact being significantly
associated with improvement on the GAS. Therefore, it seems that more
work/vocational contact was related to better global and social functioning after 1
78
year o f service delivery.
Non-Specific Factor Contacts
Total duration of individual supportive contact (1:1) and case management
contacts were selected to represent non-specific factors that impact participant
outcomes. In this program it should be noted that case management contacts were
used to monitor and discuss program progress with participants. These particular
types o f contact were selected because they represent non-specific factors.
The 6 month findings (Table 26) indicated that a higher total duration of
individual supportive contacts was significantly associated with improvement in
SCOS use of institutions and RFS independent living. As shown in Table 27, a
higher total duration of individual supportive contact after 1 year of treatment
continued to be associated with improvement in SCOS use o f institutions, as well
as, the SCOS global scale. A significant negative association was also observed
between a higher duration o f individual supportive contact and the BPRS. Hence,
it appears that more individual supportive contact over the total study period is
related to less of an increase in symptoms, more improvement in independent
living, and less use of institutions.
Regarding total duration of case management contact, the 6 month findings
indicate that more case management contact was significantly associated with
improvement in RFS independent living, SCOS use o f institutions, and the deficit
syndrome. At 1 year outcome the association between more case management
contact and improvement in SCOS use o f institutions continues, while there is no
79
association with the deficit syndrome. However, significant associations are noted
between a higher total duration of case management contact and improvement on
virtually every measure o f psychosocial functioning at 1 year outcome.
Consequently, these particular findings indicate the importance of the relationship
between non-specific treatment factors and improved participant functioning in
both the clinical and psychosocial domains.
Results Summary
These findings suggest that the services delivered were effective in
improving participants functional status. Improvement that was noted after 6
months of treatment was maintained during the subsequent 6 months with a lower
intensity of services and less continuity of care. Very importantly, numerous
associations were noted between service process characteristics and improvement
in functional status at both 6 and 12 month outcomes. The most prevalent
associations were observed between a higher total duration o f treatment and
improvement, and more gaps in services and less improvement.
In terms o f the explanatory power of service process variables, the findings
indicated that gaps in services were more salient in accounting for outcome
variation in two o f three models that explained clinical functioning, while total
duration of treatment was more relevant in models that explained variation in
improvement in psychosocial change score outcomes. The exceptions to this
conclusion were (a) total duration retaining explanatory power in 12 month SCOS
80
use o f institution change scores, and (b) gaps in service explaining the majority of
variation in both 6 month and 12 month RFS work functioning change scores.
Only a few associations were observed between participant characteristics
and change score outcomes. The significant associations were between being male
and/or being younger at study entry and change score outcomes. Similarly, only
one multivariate model emerged when participant characteristics were used
exclusively to explain change score outcomes.
On the other hand, models that combined both service process variables
and participant characteristics explained more outcome variation than models that
included only service process variables or participant characteristics variables.
These combined multivariate models suggested that the total duration of treatment
was more salient in explaining change score outcome variation than participant
characteristics.
Support was also generated for specificity o f treatment. These findings
suggested that social/recreation and independent living contacts were associated
with improvement in associated areas of functioning, as well as, on several other
non-specific measures of clinical and psychosocial functioning. Similarly,
individual supportive contacts and case management contacts were also associated
with improvement in both clinical and psychosocial functioning.
DISCUSSION
Perhaps the most important aspect of these findings pertain to the concepts
of continuity of care, intensity of services, and treatment comprehensiveness in
accomplishing improvement in participant functional outcomes. These service
process characteristics have been hypothesized for nearly a quarter century to be
important in effecting participant outcomes, yet this is the first study to test the
efficacy o f these concepts on one sample. These findings stand apart from
previous work in community treatment as they are the first to look at (a)
improvement in participants functional status, (b) the levels o f association between
service process variables and improvement in outcomes, (c) the impact o f these
variables in terms of explaining functional status outcomes across three distinct
domains. Similarly, and not necessarily of secondary importance, were the findings
regarding specificity of treatment which indicated both specific and non-specific
effects upon participant functional status outcomes.
The findings indicated that the treatment program was effective in
improving participants functional status in the clinical, psychosocial, and subjective
experience domains. Participants received a relatively high frequency (nearly 4
times weekly), of various types o f contact that were of short duration (30
minutes), with few gaps in continuity over the first 6 months o f treatment. During
the ensuing 6 months service contacts continued to be of short duration, although
the frequency was slightly more than once per week and the number of gaps
trebled. Despite these changes in frequency o f contact and continuity, the initial
82
significant gains observed after 6 months of treatment were maintained at
significant levels during the latter half o f the 1 year study period. This finding
extends the previous literature on treatment intensity and, in effect, allows us to
combine two earlier findings into a new plausible hypothesis. On one hand,
Dietzen and Bond (1993) have suggested programs that maintain a minimum o f
service intensity may reduce hospital use among participants who were frequently
hospitalized. On the other hand, Snowden and Clancy (1990) concluded that there
may be diminishing returns with continued high levels o f intensity over extended
periods o f time. What the present findings suggest is that an initial high level of
treatment intensity combined with good continuity o f care decreased use of
institutional care while also improving participant functional status. It was also
observed that these initial improvements were sustained when less intensive service
delivery was maintained over time.
O f particular importance were the numerous significant associations
observed between service process variables and participant improvement in the
clinical and psychosocial domains at both 6 month and 12 month change score
outcomes. The most prevalent associations were noted between more intensity of
services and better outcomes, and higher levels of continuity o f care and better
outcomes. Service comprehensiveness was only associated with improvement in
independent living and only at 6 month outcome. Amazingly, continuity of care
and intensity were associated with nearly every clinical and psychosocial outcome
indicator after 1 year of treatment. The number and magnitude o f these findings
83
(13 of 44 possible correlations at 6 month outcome, and 18 of 44 possible
correlations at 12 month outcome) far exceeds what would be expected by chance.
The general pattern of the findings indicated that a higher intensity of treatment
with better continuity of care were the service process characteristics most clearly
related to participants showing improvement in clinical and psychosocial
functioning at both 6 month and 12 month outcomes.
In regard to service process characteristics and explanatory models,
multicollinearity was encountered. This corroborates the finding of Dietzen and
Bond (1993) who also noted high intercorrelations between measures o f intensity
with a much larger sample. Consequently, use of a larger sample in future
investigations may not rectify that issue. Nevertheless, the problem with
multicollinearity was solved in the current investigation by (a) eliminating
frequency of treatment and retaining total duration as the measure of intensity used
in model building, and (b) looking at multiple combinations of potential
explanatory variables in the regression models to determine the relative importance
o f each variable entered.
In this conservative and cautious manner substantial support was generated
for the experimental hypothesis that service process variables were able to explain
significant variation in participant improvement. Several multivariate explanatory
models were found that generated support in both the clinical and psychosocial
functioning domains. The results indicated that better continuity of care seemed to
have the most relative importance in explaining variation in clinical improvement,
while more treatment intensity appeared to be the most salient in explaining the
variation in psychosocial improvement. The lone exception to these findings
pertained to the utility of intensity being able to explain a significant amount of the
variation in use o f institutions at 1 year outcome, and continuity being able to
explain significant variation in improvement in work functioning after 1 year of
treatment. Overall, the explanatory models that emerged were able to explain
between 12-27% o f the improvement in clinical outcomes, and between 13-45% of
the improvement in psychosocial outcomes.
In terms o f participant characteristics and change score outcomes, only a
few relationships emerged in the data between participant variables and change
scores. While there were no significant associations in the findings between
participant characteristics and change score outcomes at 6 months, significant
relationships were noted between gender and improvement in RFS global and
independent living, and age at study entry and improvement in RFS independent
living in the 12 month sample. These relationships indicated that being male and
being younger were significantly related to improvement in independent living.
Being male was also the only participant characteristics associated with
improvement in RFS global functioning after 1 year of treatment. The implications
of these finding regarding age would suggest that it is important to intervene as
early as possible in clients course of illness. The findings concerning gender should
be approached cautiously as the samples were skewed in that there were
significantly more males in the study than females. Nonetheless, being male was
85
associated with more improvement than being female in terms of RFS total
functioning and RFS independent living.
Partial support was also generated for the experimental hypothesis in terms
of participant characteristics explaining change score improvement. The
participant specific characteristics of being male and younger at study entry
explained a significant amount of the variation in improvement in independent
living after 1 year of treatment. However, this was the only multivariate
explanatory model o f significance that emerged when only participant
characteristics were used in regression models.
On one hand, the findings concerning participant characteristics and change
score outcomes were somewhat surprising given the number of previous studies
that found participant related variables were associated with participant outcomes
over time (Verghese, et al., 1989; Prudo and Blum, 1987; Gaebel and Pietzcker,
1987; Maj, Starace, Kemali, 1987; Moller, Schmid-Bode, vonZerssen, 1986;
Strauss and Carpenter, 1972). O f particular interest, was the fact that the SCPS
was not associated with participant improvement in either the 6 month or 12
months samples, although the SCPS had been found to be superior as a prognostic
tool (Gaebel and Pietzcker, 1987; Moller, Schmid-Bode, von Zerssen, 1986), and
includes virtually every prognostic variable that had previously been found to be
related to longitudinal outcomes. On the other hand, the previous investigations
that utilized participant characteristics and prognostic factors were concerned with
absolute levels o f functioning at outcome, not with participant change scores or
86
the ability of participant characteristics to explain variation in improvement in
participant functional status.
In terms o f building a more comprehensive model that included both
participant and service process characteristics, four significant models were
observed. Three of these models explained more variation in outcome than any of
the models that solely used either service process variables or participant
characteristics. In all of the combined models that emerged when both participant
and service process characteristic variables were included, treatment intensity was
found to be the most salient variable in explaining outcome variation. These
findings indicted that participants who received a higher intensity o f service
delivery and/or were younger were associated with more improvement in use of
institutions and two measures o f global functioning at 12 month outcome.
Additionally, male participants that experienced higher service intensity after 6
months o f treatment experienced more improvement in independent living.
Based on these findings it appears that of all o f the participant
characteristic and service process variables that were tested, treatment intensity
was the single most important factor in explaining improvement in outcomes in
these samples. Nevertheless, it must be recognized that this finding does not
discount the importance of continuity of care. As the correlations between service
process variables indicated, treatment intensity and continuity o f care were closely
related variables in this particular CSP. The degree to which these variables are
generally related needs to be assessed in the context o f other treatment programs.
87
All in all, the findings concerning service process characteristics have
profound implications for future program funding, as well as, the managed care
environment. Quite simply, participants in this sample realized significant
improvement in functional status after 6 months o f treatment when there was a
high intensity of services with few breaks in continuity o f care. These gains were
maintained during the latter half o f the study period with a lower intensity of
treatment and somewhat less overall continuity o f care. More treatment intensity
was most closely associated with better functional status outcomes and explained
more improvement, while less continuity of care was associated with poorer
functional status outcomes and explained less improvement. Consequently, in
consideration of the trend toward capitated funding it does not make fiscal sense to
discontinue care when termination o f services will result in a more than likely
increase in use of institutions, and a concomitant decrease in functional status that
will require increases in other types of service provision.
While it was discussed above that participants showed marked
improvement in functional status in the clinical, psychosocial, and subjective
experience domains, there were increases in symptomatology noted on the BPRS
at both baseline to 6 month and baseline to 12 month outcomes. However, a
possible alternative explanation of this particular finding is that the demands of
participating in treatment may have led to an increase in participant perceived
stress, which in turn led to increases in symptomatology. This finding also lends
additional support to findings which suggested that increases in functional status
88
co-occur with increases in symptoms (Brekke, et al., in press; Tarrier and Turpin,
1992; Test, 1984). The increases in symptom levels were also significantly
associated with participants experiencing more breaks in continuity of care, as well
as, receiving services that were less intensive. Additionally, and of primary
importance, a multivariate model emerged that demonstrated that more breaks in
continuity o f care had the most relevance in explaining variation in BPRS change
scores over the 1 year study period.
In continued contrast to the strong findings in the clinical and psychosocial
domains, limited improvement was observed in participant subjective experience
outcomes. In this matter, the results indicated significant participant improvement
was observed only in self-esteem, and only at 12 month outcome. This finding
lends support to the suggestion of Rosenfeld and Neese-Todd (1993) that it may
take longer than 1 year to observe gains in subjective experience outcomes. It was
also noted that participant characteristics and service process characteristics were
neither associated with, nor held any explanatory power in terms of participant
subjective experience outcomes. This suggests that these characteristics may also
be distinct and not affected by functional status (Brekke, et al., 1993).
A concomitant finding that pertains to the issue o f program efficacy
concerns the limited associations that were observed between participant
characteristics and service process variables. Only a few associations were
observed between these distinct groups of variables at 6 month outcomes. No
significant associations were noted between these groups of variables in the 12
89
month sample. Additionally, there were no clear associations between baseline
functioning and service process characteristics. Therefore, it appears that it was
unlikely that program staff were selecting participants that had the most promise of
showing improvement and targeting them for treatment.
Also of particular importance were the results pertaining to specificity of
treatment within the treatment program. In this matter partial support was found
for the experimental hypothesis. The findings indicated that specific treatment
efforts in the domains o f independent living and social/recreational activities were
significantly related to improvement in associated areas o f participant functioning.
The data indicated further that these specific forms of service contact w ere also
associated with improvement on clinical and other psychosocial measures. The
findings regarding social recreational treatment substantiate the findings o f Prudo
and Blum (1987) who advised that intervening in social domains may have clinical
relevance.
On one hand, the findings did not indicate there was an association
between work/vocational treatment and improvement in w ork functioning. On the
other hand, this finding concerning work/vocational contact specificity is not
particularly surprising as Brekke (1992) has speculated that a more complex
behavioral repertoire may be required to show improvement in work functioning
than in social domains. Consequently, it may necessitate multiple forms o f
treatments for participants to develop such a repertoire. This suggestion is also
supported by the findings concerning independent living contacts and improvement
90
in work outcomes in the present investigation. Additionally, work/vocational
contacts were related to participants demonstrating improvement in both social
and global functioning. Therefore, these contacts appear to have value in assisting
participants to improve in other areas of functioning.
Concerning non-specific treatment factors, the findings suggested the utility
o f both individual supportive contact and case management contacts in this
treatment program. To avoid confusion with other operational definitions o f case
management, it is important to reiterate that in this program case management
contacts were those staff/participant encounters that (a) monitored progress and/or
motivated participants for continued participation, and (b) resolved conflicts that
participants encountered in the program. The results indicated that supportive
individual treatment and case management were both independently significant in
relation to participants demonstrating improvement in both the clinical and
psychosocial domains at 6 and 12 month outcomes. Of particular relevance, both
o f these types o f treatment were associated with participants demonstrating a
decrease in use o f institutions. Additionally, case management contacts were
significant at both 6 and 12 month outcomes in terms of effecting improvement in
participants level o f independent functioning.
These findings are highly relevant for program planners who are called
upon to develop programs that decrease utilization of institutionalized care.
Participants that had higher levels o f case management contact also demonstrated
significant improvement on every psychosocial measure except social functioning
91
after 1 year of service delivery. However, it is extremely important to reiterate
that case management was only one aspect of treatment in this program, and it
should not be seen as a substitution for comprehensive services. Additionally,
there were a range o f services available to participants within the treatment
program. What this finding does suggest is the importance of closely monitoring
participant progress in the case management process, as well as, having various
types o f services readily available so that treatment efforts to improve participant
levels o f functioning may be realized.
The findings concerning specificity o f treatment also provide additional
support for the concept of "open-linked" systems of outcome that were proposed
by Strauss and Carpenter (1991, 1977, 1974). Specifically, distinct types of
treatment were found to impact other areas o f functioning. These findings indicate
that different functional outcome domains are closely related. Moreover, these
results offer substantial support in regard to the necessity to measure participant
outcomes across several domains. Only by measuring outcomes across multiple
domains will beneficial, although unexpected, treatment effects be discovered. In
this manner the utility o f specific treatments may be better judged, and the
potential o f "throwing out the baby with the bath water" will be decreased. The
implications of such findings are of exceptional importance as service providers
encounter managed care environments that may seek to eliminate seemingly
inconsequential modes o f treatment.
92
Limitations o f the Study
There are several limitations to this study that merit discussion. First, this
study used small samples. Consequently, there may have been additional
significant findings that did not emerge because of the small sample sizes. On the
other hand, when viewed in this context the bi-variate findings are seemingly more
robust than the multi-variate findings. Therefore, the bi-variate findings add
significant information to current knowledge. Second, there were significantly
more males than females in each o f the samples. As a result the findings
concerning gender should be approached cautiously. Third, because of the
relatively small sample sizes ethnicity was dummy coded as Caucasian/Non-
Caucasian. It should be noted that while this does not reflect the true impact of
ethnicity, it does reflect the presence of minority status. Therefore, the findings
concerning ethnicity should be approached and generalized cautiously. Fourth,
this study used currently accepted definitions o f service process variables that have
not been previously tested in relation to participant outcomes. Therefore, caution
must be taken in comparing these findings to previous conceptualizations o f these
variables. Finally, we do not know the degree to which these findings are
generalizable to other community support programs.
Future Research
Regarding future research, it is essential for continued investigations that
evaluate both participant and service process characteristics using agreed upon
operational definitions. These investigations should measure participant
93
improvement and associated outcomes across multiple domains of functioning
using standardized outcome measures so that (a) findings may be compared across
studies, and (b) non-specific effects o f treatment may continue to be discovered.
Additionally, by measuring service process characteristics the replication o f
successful programs may be enhanced.
Future investigations should continue to test the efficacy of participant
characteristics and prognostic variables in their ability to explain longitudinal
improvement. By conducting similar research on a larger sample future
investigations could test the explanatory and predictive efficacy of the components
of the SCPS in relation to change score outcomes. Additionally, given the recent
findings o f Brekke et al. (in press) and Straube et al. (1988) such investigations
could include neuropsychological and psychobiological measures. In this manner
new combinations o f variables or groups of novel variables could be identified that
may better predict participants respective ability to improve in functional status
domains. It is possible that such endeavors may reveal new prognostic techniques
that are more efficient than current measures.
Continued investigations are also needed in the area o f service process
variables. The present findings should be replicated, and future investigations
should consider looking at participant outcomes for periods longer than 12
months. It is likely that subjective experience effects may be realized in studies of
a longer duration than the present investigation. Although individualization of
treatment has been hypothesized to impact participant outcomes, this service
94
process characteristic has only recently been operationalized and explored by Arns
and Linney (1995). Additionally, controlled manipulation is warranted to
investigate the effects o f (a) varying levels of treatment intensity, and (b) varying
degrees of program comprehensiveness. More research is also needed in terms of
specificity o f treatment. While the current investigation was concerned with
improvement on associated measures o f functioning, future investigations could
also determine the explanatory value o f those variables on related outcomes.
Concomitantly, Factor Analysis or Cluster Analysis could also be employed to
determine if there are specific factors or clusters o f treatment types that effect
functional outcomes. Such investigation should also consider the impact on non
specific outcomes.
Conclusion
This has been the first outcome investigation in the community treatment of
schizophrenia to look at service process characteristics and to use the participant
as the unit o f observation to monitor improvement in functional status.
Consequently, this study has been able to (a) provide useful data for program
replication, (b) examine the relationships between participant characteristics,
service process characteristics, and clinical, psychosocial, and subjective
experience outcomes, and (c) begin to answer the question of, "What treatment
elements, affect what outcomes, in the community treatment o f schizophrenia."
The findings offer support for the explanatory value o f these divergent
groups of variables and extend previous research. Extensive support was found
for the concept o f service intensity in that more intensive services effected
improvement in participant functional status. Similarly, more continuity of care
also led to better outcomes. Support was also found for specificity o f treatment
and associated outcomes. Specific treatments were found to be significantly
associated with improvement in a wide range of participant functional outcomes.
Non-specific treatment factors were also found to be related to improvement in
various participant outcomes. As this study is the first of its type, it is essential
that it be replicated in future community treatment research efforts.
96
Figure I. Definition of Major Variables
I. Participant Variables
Sex
Race
Age at study entry
Prognostic Indicators
II. Service Variables
Intensity
a. Total Duration
b. Total Frequency
Continuity of Care
Comprehensiveness
Specificity
a. Type by Duration
b. Type by Frequency
III. Dependent Variables
Clinical Variables
a. Symptoms
b. Deficit Syndrome
c. U se o f Institutions
Psychosocial Variables
a. G lobal Functioning
b. General Role functioning
c. W ork Functioning
d. Social Functioning
e. Independent Living
Subjective Experience Variables
a. Satisfaction with Life
b. S elf Esteem
Definition
Years
Score on Strauss and Carpenter Prognostic Scale
Total duration of all service contacts
Total number of all service contacts
# of periods in which there was not staff to participant
contact for periods greater than 30 days.
# of services that accounted for more than 3% of
contact time
Total number of minutes o f service by type
Total frequency of contacts by type
Total Score on BPRS
Total score on Deficit Syndrome Scale
Score on Strauss and Carpenter Hosp. Scale
Score on GAS
1 .Total Score on Strauss and Carpenter Outcome Scale
2.Total Score on RFS
Score on RFS Work Scale
Score on RFS Social Scale
Score on RFS Independent Living Scale
Score on Satisfaction with Life Scale
Score on Index of Self Esteem
97
Table 1
Demographic Characteristics
Characteristic
6 Month
(N=41)
N %
1 Year
(N=30)
N
Gender
Race
female
male
Caucasian
Black
Latino
Asian
Native American
Other
11
30
20
10
8
2
1
26.8
73.2
48.8
24.4
19.5
4.9
2.4
8
22
16
7
4
1
1
1
Mean SD Mean
Age at Study Entry 33.02 7.7 33.2
Strauss and Carpenter 28.3 3.9 28.7
Prognostic Scale
%
26.7
73.3
53.3
23.3
13.3
3.3
3.3
3.3
SD
7.4
4.7
98
Table 2
Levels o f Participant Baseline Functioning for 6 Month and 12 Month Samnles
6 Month Sample 1 Year Sample
Mean SD Mean SD
Clinical Variables
BPRS (22 Item) 39.5
Deficit Syndrome 17.8
Strauss & Carpenter (hosp) 3.1
Psychosocial Variables
Strauss& Carpenter (total) 2.5
Global Assessment Scale 31.4
Role Functioning Scale (total) 7.6
RFS Work Functioning 1.9
RFS Social Functioning 2.8
RFS Independent Living 2.9
Subjective Experience Variables
Satisfaction with Life 42.1
Index of Self Esteem 85.0
13.4 39.0 14.6
5.7 16.9 5.4
1.2 3.1 14
2.1 2.5 2.0
9.8 32.0 11.0
2.6 7.5 2.6
1.2 1.9 1.3
1.9 2.6 1.7
1.0 2.9 1.1
15.4 40.4 12.4
13.4 87.3 11.9
Table 3
Service Process Variable Descriptive
6 Months 1 Year
(0-180 days) (0-365 days)
N=41 N=30
Variable Mean SD Mean SD
Total Duration of 3144.8 2769.8 4004.8 3580.3
All Contacts (in minutes)
Total Frequency of 93.2 74.4 125.4 100.5
All Contacts
Average Duration of 31.0 9.4 30.0 8 . 6
All Contacts (in minutes)
Continuity of Care 0.9 1.1 2.9 2.5
(# gaps in service >30days)
Comprehensiveness (3% ) 5.2 1.4 4.9 1.6
99
Table 4
Correlations Between Service Process Variables at 6 Months
1 2 3 4
1. # 3 0 Day Gaps
in Service -
2. Total Duration
o f All Contacts
. 5 9 ***
3. Total Frequency
of All Contacts
_ 5 7 *** .95*** —
4. Comprehensiveness
(>3% of Total Time) ns .36* ns -
2 -tailed significance: * p<= .05 ** p<=.01 *** p < = . 0 0 1
Correlations Between Service Process Variables at 1 Year
1 2 3 4
1. # 3 0 Day Gaps
in Service -
2. Total Duration
of All Contacts
. 7 4 ***
3. Total Frequency
o f All Contacts
_ 7 4 *** 9 5 ***
—
4. Comprehensiveness
(>3% of Total Time) ns .44** .40* -
2-tailed significance: *p<=.05 ** p<=.01 *** p<= 001
100
Table 5
Mean Contact Duration and Frequency bv Type of Service at 6 Months
Duration Frequency
(in minutes)
Type of Contact N Subjects Mean SD Mean SD
Vocational 36 25.0 7.1 31.1 28.8
Living Arrangement 30 2 1 . 8 8 . 0 4.2 2.9
Physical 1 2 20.7 18.7 1 . 8 1 . 8
Medication Management 2 1 19.7 8 . 6 5.7 7.8
Social/Recreation 35 40.1 18.4 1 1 . 6 9.2
Case Management 34 34.0 13.6 10.5 7.8
1:1 Supportive Contact 41 19.0 6.5 16.3 16.5
Activities of Daily Living 29 43.3 16.6 29.7 27.5
Crisis 2 0 25.9 23.8 2.7 1.9
Basic Living Skills 16 32.2 31.0 5.2 5.1
"Other" Contacts 5 36.5 15.4 1 . 2 0.5
Drug/Alcohol 1 2 0 . 0 0 . 0 1 . 0 0 . 0
Mean Contact Duration and Freauencv bv Tvoe of Service at 12 Months
Type of Contact
Duration
(in minutes)
N Subjects Mean SD
Frequency
Mean SD
Vocational 29 24.9 7.6 44.1 40.2
Living Arrangement 2 2 22.7 9.8 5.4 4.5
Physical 1 0 17.4 15.4 2.4 2 . 1
Medication Management 17 18.8 7.0 7.9 8 . 8
Social/Recreation 28 37.0 13.9 17.1 18.0
Case Management 26 34.0 17.1 16.2 1 2 . 8
1:1 Supportive Contact 29 19.4 9.4 20.9 19.9
Activities of Daily Living 2 2 45.4 26.6 26.6 29.7
Crisis 1 2 27.5 21.5 2.7 1.7
Basic Living Skills 1 2 33.7 21.5 5.7 5.3
"Other" Contacts 7 32.9 20.5 1.4 .5
Drug/Alcohol 2 13.1 9.7 2.5 2 . 1
101
Table 6
Correlations Between Service Process Variables and Participant Baseline Functioning at 6 Months (N =41)
# 30 Day Total Total Comprehensiveness
Gaps in Srv. Duration Frequency (>=3% Ttl Contact
Time)
Clinical Variables
BPRS (22 Item)
Deficit Syndrome
S&C Hospitalization
Psychosocial Variables
S&C Total Score
GAS
RFS Total
RFS Work
RFS Social
RFS Ind. Living
Subjective Experience
Satisfaction with Life
Index of Self Esteem
2-tailed significance: * p<=.05
Correlations Between Service Process Variables and Participant Baseline Functioning at 12 Months (N=30)
# 30 Day Total Total Comprehensiveness
Gaps in Srv. Duration Frequency (>=3% Ttl Contact Time)
Clinical Variables
BPRS (22 Item)
Deficit Syndrome
S&C Hospitalization
Psychosocial Variables
S&C Total Score
GAS -.39*
RFS Total
RFS Work
RFS Social
RFS Ind. Living
.34*
.30*
-.36*
Subjective Experience
Satisfaction with Life
Index of Self Esteem .36*
2-tailed significance: * p<=.05
Table 7
Paired T-Test's Using 6 Month Samnle On All Participant Outcome Indicators
6 Month Outcomes
Mean
Change
SD t-score p-value
( 1-tailed)
Clinical Variables
BPRS (22 Item) -1.3 10.4 -0.83 ns
Deficit Syndrome - 1 . 8 4.9 -2.35 . 0 0 1 2
Strauss & Carpenter (hosp) -0.7 1.3 -3.62 .0005
Psychosocial Variables
Strauss& Carpenter (total) -2 . 2 2 . 8 -5.12 . 0 0 0 1
Global Assessment Scale -7.8 14.4 -3.45 .0005
Role Functioning Scale (total) -2 . 6 3.8 -4.39 . 0 0 0 1
RFS Work Functioning - 1 . 2 1 . 6 -4.93 . 0 0 0 1
RFS Social Functioning -0.7 2.3 -2.03 .025
RFS Independent Living -0.7 1 . 2 -3.68 .0005
Subjective Experience Variables
Satisfaction with Life -1.5 13.2 -0.7 ns
Index of Self Esteem -2.5 13.8 - 1 . 2 ns
Paired T-Tests Using 12 Month Sample On All Participant Outcome Indicators
12 Month Outcomes
Mean
Change
SD t-score p-value
( 1 -tailed)
Clinical Variables
BPRS (22 Item) -5.9 13.8 -2.34 .0014
Deficit Syndrome -0 . 2 5.6 -0.19 ns
Strauss & Carpenter (hosp) -0.7 1.4 -2.89 .0035
Psychosocial Variables
Strauss& Carpenter (total) -1.9 2.5 -4.04 . 0 0 0 1
Global Assessment Scale -6 . 6 18.4 -1.95 .03
Role Functioning Scale (total) -3.2 3.9 -4.49 . 0 0 0 1
RFS Work Functioning -1.3 2 . 0 -3.53 .0005
RFS Social Functioning -1.4 2 . 0 -3.73 .0005
RFS Independent Living -0.5 1 . 6 -1.81 .04
Subjective Experience Variables
Satisfaction with Life -1.3 14.5 -0.50 ns
Index of Self Esteem -5.4 14.9 -1.97 .029
Table 8
Correlations Between Participant Characteristics and 1 Year Change Score Outcome V ariables
R ace Gender Age Prognosis
Clinical Variables
BPRS (22 Item)
Deficit Syndrome
S&C Hospitalization
Psvchosocial Variables
S&C Total Score
GAS
RFS Total
RFS Work
RFS Social
RFS Ind. Living
.36*
.39* -.38*
Subiective F.XDerience
Satisfaction with Life
Index of Self Esteem
2 -tailed significance: * p<=,05
104
Table 9
Correlations Between Service Process Variables and 6 Month Change Score Outcome Variables
# 3 0 Day Total Total Comprehensiveness
Gaps in Srv. Duration Frequency (>=3% Ttl Con.Time)
Clinical Variables
BPRS (22 Item)
Deficit Syndrome
S&C Hospitalization -.31* .28* .30*
Psvchosocial Variables
S&C Total Score
GAS
RFS Total
RFS Work
RFS Social
RFS Ind. Living
-.29*
-.35**
-.53***
.35**
.32*
.65***
.34*
.30*
.56*** .27*
Subjective Experience
Satisfaction with Life
Index of Self Esteem
1 -tailed significance: * p<=.05 ** p<== .0 1 *** p< = . 0 0 1
Correlations Between Service Process Variables and 12 Month Chanee Score Outcomes
# 30 Day
Gaps in Service
Total Total Comprehensiveness
Duration Frequency (>=3% Ttl Con. Time)
Clinical Variables
BPRS (22 Item) .46** -.35* -.35*
Deficit Syndrome -.33*
S&C Hospitalization -.37* .40** .36*
Psvchosocial Variables
S&C Total Score -.32* .36*
GAS -.46** .52** .44**
RFS Total -.31* .46** .33*
RFS Work -.43** .35*
RFS Social
RFS Ind. Living .37*
Subiective Experience
Satisfaction with Life
Index of Self Esteem
1 -tailed significance: * p<=.05 ** p<=.01
105
Index of Self Esteem
1 -tailed significance: * p<=.05 ** p<=.01
Table 10
Correlations Between Particinant Characteristics and 6 Month Service Process Variables
Variable
Comprehensiveness
# 30 Day Gaps Total Total
in Service Duration Frequency
S&C Prognosis -.30*
Age at Study Entry
Gender
Race -.41**
2 -tailed significance: *p<=. 05, **p<=.01
Table 11
Gender. A ee at Studv Entrv. and RFS IndeDendent I .ivine 12 Month Outcomes
Variable Beta T T-Significance
Gender .332455 2.036 .05
Age at Study Entry -.321330 -1.911 .07
R2 =.26 F=4.6 F-Significance=,02
Table 12
Gans in Service. Comnrehensiveness. and 6 Month SCOS Use of Institutions Outcomes
Variable Beta T T-Significance
Gaps in Service -.350630 -2.214 .03
Comprehensiveness -.160495 -1.013 .32
r2= 1 2 F=2.5 F-Significance=.09
106
Table 13
Total Duration. Comprehensiveness, and 12 Month BPRS Outcomes
Variable Beta T T-Significance
Total Duration -.141541 -0.529 .60
Comprehensiveness .260605 1.389 .18
Gaps in Service .431757 1.715 . 1 0
R2 =.27 F=3.1 F-Significance=.04
Gaos in Service and ComDrehensiveness Model
Variable Beta T T-Significance
Gaps in Service .526772 3.031 .005
Comprehensiveness .226434 1.303 . 2 0
R2 =.26 F=4.7 F-Significance=.02
Total Duration and Comnrehensiveness Model
Variable Beta T T-Significance
Total Duration -.469598 -2.420 . 0 2
Comprehensiveness .276499 1.425 .17
R2 =. 18 F=3.0 F-Significance=.07
Table 14
Total Duration. ComDrehensiveness. and 12 Month SCOS U se of Institution Outcc
Variable Beta T T-Significance
Total Duration .395129 2 . 0 1 2 .05
Comprehensiveness .016664 0.085 .93
R2 =. 16 F=2.6 F-Significance=,09
Table 15
Total Duration. ComDrehensiveness. and 6 Month GAS Outcomes
Variable Beta T T-Significance
Total Duration .368269 2.263 .03
Comprehensiveness -.046983 -.289 .77
R2 =. 13 F=2.7 F-Significance=.08
107
Table 16
Total Duration. ComDrehensiveness. Gans in Service, and 6 M onth RFS Total Outcomes
Variable Beta T T-Significance
Total Duration .212824
Comprehensiveness -.130958
Gaps In Service -.255169
1.094
- .805
-1.353
.28
.43
.18
R2 =. 15 F=2.2 F-Significance=. 10
Gans in Service and ComDrehensiveness Model
Variable Beta T T-Significance
Gaps in Service -.368994 -2.341 .03
Comprehensivenss -.085016 -.539 .59
R2 =.13 F=2.7 F-Significance=.08
Table 17
Total Duration. Comnrehensiveness. GaDs in Service, and 6 M onth RFS Indenendi
Variable Beta T T-Significance
Total Duration .506872 3.224 .003
Comprehensiveness .022073 .168 .87
Gaps in Service -.218434 -1.434 .16
R2 =.45 F=10.0 F-Significance=.0001
Total Duration and Comnrehensiveness Model
Variable Beta T T-Significance
Total Duration .631276 4.749 . 0 0 0 0
Comprehensiveness .036749 .276 .78
R2 =.42 F=13.6 F-Significance=.0000
GaDs in Service and Comnrehensiveness Model
Variable Beta T T-Significance
Gaps in Service -.489525 -3.450 . 0 0 1
Comprehensiveness .131490 .927 .36
R2 =.29 F=7.8 F-Significance=.001
108
Table 18
Total Duration. Comprehensiveness. Gans in Service, and 12 Month SCOS Total Outcomes
Variable Beta T T-Significance
Total Duration .438994 1.602 .12
Comprehensiveness -.347048 -1.807 .08
Gaps in Service -.096458 -.375 .71
R2 =.23 F=2.6 F-Significance=.07
Total Duration and Comprehensiveness Model
Variable Beta T T-Significance
Total Duration .512284 2.714 .01
Comprehensivenss -.350599 -1.858 .07
R2 =.23 F=4.0 F-Significance=.03
Gans in Service and Comprehensiveness Model
Variable Beta T T-Significance
Gaps in Service -.391149 -2.109 .04
Comprehensiveness -.241067 -1.300 .21
R2 =. 16 F=2.5 F-Significance=.10
Table 19
Total Duration. Comprehensiveness. Gans in Service, and 12 Month GAS Outcomes
Variable Beta T T-Significance
Total Duration
Comprehensiveness
Gaps in Service
.390499
.041420
-.152983
1.478
.224
-.616
.15
.83
.54
R2 =.29 F=3.5 F-Significance=.03
Total Duration and Comnrehensiveness Model
Variable Beta T T-Significance
Total Duration
Comprehensiveness
.506738
.035788
2.772
.196
. 0 1
.85
R2 =.27 F=5.1 F-Significaince=.01
GaDs in Service and Comnrehensiveness Model
Variable Beta T T-Significance
Gaps in Service
Comprehensiveness
-.415120
.135694
-2.337
.764
.03
.45
R2 =.23 F=3.9 F-Significance=.03
Table 20
Total Duration. Comnrehensiveness. Gaps in Service, and 12 Month RFS Total Outcomes
Variable Beta T T-Significance
Total Duration .583049 2.127 .04
Comprehensiveness -.121669 -.633 .53
Gaps in Service .087570 .340 .74
R2 =.23 F=2.6 F-Significance=.07
Total Duration and Comnrehensiveness
Variable Beta T T-Significance
Total Duration .516511 2.737 . 0 1
Comprehensiveness -.118445 -.628 .54
R2 =.23 F=4.0 F-Significance=.03
110
Table 21
Variable Beta T T-Significance
Gaps in Service -.455295 -2.511 . 0 2
Comprehensiveness -.083025 -.458 .65
R2 =. 19 F=3.2 F-Significance=.06
Table 22
Total Duration. A ee at Studv Entrv. and 12 Month SCOS Use of Institutions Outcc
Variable Beta T T-Significance
Total Duration .412922 2.550 . 0 2
Age at Study Entry -.361799 -2.235 .03
R2 =.29 F=5.6 F-Significance=,009
Table 23
Total Duration. Race, and 6 Month RFS Indenendent I.ivine Outcomes
Variable Beta T T-Significance
Total Duration .614638 4.977 . 0 0 0 0
Race (wht/non-wht) .161053 1.304 . 2 0
R2 = 44 F=15.0 F-Significance=,0000
Table 24
Total Duration. A ee at Studv Entrv. and 12 Month GAS Outcomes
Variable Beta T T-Significance
Total Duration .531876 3.501 . 0 0 2
Age at Study Entry -.323131 -2.127 .04
R2 =.38 F=8.2 F-Significance=.002
Ill
Table 25
Variable Beta T T-Significance
Total Duration .47814 3.003 .006
Age at Study Entry -.336419 -2.132 .04
R2 =.33 F=6 . 6 F-Significance=.005
Table 26
Correlations Between Specificity of Treatment and 6 Month Change Score Outcomes
Type o f Service Contact
(measured by total duration in minutes)
Work Soc/Rec ADL Living 1:1 Case
Arrg Mgt
6 Month Change Score Outcomes
Clinical Variables
BPRS (22 Item)
Deficit Syndrome .28*
.32* .31*
.28*
.61*** .33* .31* .69***
Subjective Experience Variables
Satisfaction with Life
Index of Self Esteem
Strauss & Carpenter (hosp) .50***
Psychosocial Variables
Strauss& Carpenter (total)
Global Assessment Scale .65***
Role Functioning Scale (total) .27* .48***
RFS Work Functioning
RFS Social Functioning .29* .39***
RFS Independent Living .54***
1-tailed significance: *p<=,05, **p<=.01, ***p<=.001
113
Table 27
Correlations Between Specificity o f Treatment and 12 month Chance Score Outcomes
Type of Service Contact
(measured by total duration in minutes)
Work Soc/Rec ADL
12 Month Change Score Outcomes
Living
Arrg
1:1
Clinical Variables
BPRS (22 Item) -.37* -.39*
Deficit Syndrome .31*
Strauss & Carpenter (hosp) .35* .35* .31*
Psychosocial Variables
Strauss& Carpenter (total) .35* .41**
Global Assessment Scale .46** .42** .37* .33*
Role Functioning Scale (total) .34* .40**
RFS Work Functioning .36* .30*
RFS Social Functioning .42**
RFS Independent Living .41** .41**
Subjective Experience Variables
Satisfaction with Life
Index of Self Esteem
Case
Mgt
.30*
.32*
.35*
.40**
.44**
.48**
1 -tailed significance: *p<=,05, **p<=.01, ***p<=.001
114
References
Ams, P. G., & Linney, J. A. (1995). The relationship of service
individualization to client functioning in programs for severely mentally ill persons.
Community Mental Health Journal. 31(21. 127-137.
Bachrach, L. L., (1982). Assessment o f outcomes in community support
systems: Results, problems, and limitations. Schizophrenia Bulletin. 8. 39-60.
Bachrach, L. L. (1993a). Continuity o f care and approaches to case
management for long-term mentally ill patients. Hospital and Community
Psychiatry. 44(51. 465-468.
Bachrach, L. L. (1993b). The biopsychosocial legacy of
deinstitutionalization. Hospital and Community Psychiatry. 44(61. 523-524.
Berthaume, B., Bailey, D. M., Horowitz, B., & Vukovcan-McCrory, M.
(1994). Cost and Intensity of care in case manager referrals to psychiatric day
programs. Hospital and Community Psychiatry. 45(11. 62-65.
Bond, G. R., McGrew, J. H., & Fekete, D. M. (In Press). Assertive
outreach for frequent users of psychiatric hospitals: A meta-analysis. Journal of
Mental Health Administration.
Bond, G. R., & Boyer, S. L. (1988). Rehabilitation programs and
outcome. In J.A. Ciardiello & M. D. Bell (Eds.), Vocational Rehabilitation of
Persons with Prolonged Psychiatric Disorders. Baltimore, MD: Johns Hopkins
University Press.
Bond, G. R„ Miller, L. D„ Krumweid, R. D„ & Ward, R. S. (1988).
Assertive case management in three CMHC's: A controlled study. Hospital and
Community Psychiatry. 39(41. 411-418.
Braun, P., Kochansky, G., Shapiro, R., Greenberg, S., Gudeman, J. E.,
Johnson, S., & Shore, M. F. (1981). Overview: Deinstitutionalization of
psychiatric patients, a critical review of outcome studies. American Journal of
Psychiatry. 138161. 736-749.
Brekke, J. S. (1987). The model-guided method for monitoring program
implementation. Evaluation Review. 11131. 281-299.
115
Brekke, J. S. (1988). What do we really know about community support
programs? Strategies for better monitoring. Hospital and Community Psychiatry.
39(9), 946-952.
Brekke, J. S. (1992). An examination of the relationships among three
outcome scales in schizophrenia. The Journal of Nervous and Mental Disease.
180. 162-167.
Brekke, J. S., Levin, S., Wolkon, G. H., Sobel, E., & Slade, E. (1993).
Psychosocial functioning and subjective experience in schizophrenia.
Schizophrenia Bulletin. 19(3V 599-608.
Brekke, J. S., Long, J., & Nesbitt, N. (under review). The impact of
service characteristics on functional outcomes from community support programs
for persons with schizophrenia: A growth curve analysis.
Brekke, J. S., Raine, A., Ansel, M., Lenz, T., & Bird, L. (In Press).
Psychological factors and psychosocial functioning in schizophrenia.
Schizophrenia Bulletin.
Brekke, J. S., & Test, M. A. (1987). An empirical analysis o f services
delivered in a model community support program. Psychosocial Rehabilitation
Journal. 10(4). 51-61.
Brekke, J. S., & Test, M. A. (1992). A model for measuring the
implementation of community support programs. Community Mental Health
Journal. 28(3). 227-247.
Brekke, J. S., Wolken, G. H., Sobel, E., Test, M. A., Burke, S. (1989).
NIMH Grant Proposal.
Carpenter, W. T., & Kirkpatrick, B. (1988). The heterogeneity of long
term course o f schizophrenia. Schizophrenia Bulletin. 1414Y 645-652.
Carpenter, W. T., & Strauss, J. S., (1991). The prediction o f outcome in
schizophrenia: IV. Eleven-year follow-up of the Washington IPSS cohort. The
Journal o f Nervous and Mental Disease. 179(91 517-525.
Clark, R. E., & Fox, T. S. (1993). A framework for evaluating the
economic impact of case management. Hospital and Community Psychiatry.
44(5), 469-473.
116
Darlington, R. B. (1990). Regression and linear models. New York, NY:
McGraw Hill.
Dietzen, L. L., & Bond, G. R. (1993). Relationship between case manager
contact and outcome for frequently hospitalized psychiatric clients. Hospital and
Community Psychiatry. 44t9V 839-843.
Dion, G. L., & Dellario, D. (1988). Symptom subtypes in persons
institutionalized with schizophrenia: Comparison of demographics, outcome and
functional skills. Rehabilitation Psychology. 33(21. 95-104.
Dorwart, R. A. (1988). A ten-year follow-up study of the effects of
deinstitutionalization. Hospital and Community Psychiatry, 39( 3). 287-291.
Endicott, J., & Spitzer, R. L. (1978). A diagnostic interview: The
schedule o f affective disorders and schizophrenia. Archives of General Psychiatry.
35* 837-844.
Endicott, J . , Spitzer, R. L., Fleiss, J. L., & Herz, M. I. (1979). The global
assessment scale: A procedure for measuring overall severity of psychiatric
disturbance. Archives of General Psychiatry. 137. 48-51.
Fisher, W. H., Geller, J. L., Altaffer, F., & Bennett, M. B. (1992). The
relationship between community resources and state hospital recidivism. American
Journal of Psvchiatrv. 149(31. 385-390.
Franklin, J. L., Solovitz, B., Mason, M., Clemons, J. R., & Miller, G. E.
(1987). An evaluation of case management. American Journal of Public Health.
77(6), 674-678.
Gaebel, W., & Pietzcker, A. (1987a). Prospective study of course of
illness in schizophrenia: Part II. Prediction of outcome. Schizophrenia Bulletin.
13(2), 299-306.
Gaebel, W., & Pietzcker, A. (1987b). Prospective study of course of
illness in schizophrenia: Part III. Treatment and outcome. Schizophrenia
Bulletin. 13(21. 307-316.
Goodman, S. H., Sewell, D. R., Cooley, E. L., & Leavitt, N. (1993).
Assessing levels of adaptive functioning: The role functioning scale. Community
Mental Health Journal. 29(2). 119-131.
117
Green, R. S., & Gracely, E. J. (1987). Selecting a scale for evaluating
services to the chronically mentally ill. Community Mental Health Journal. 23. 91-
102.
Harding, C. M. (1988). Course type in schizophrenia: An analysis of
European and American studies. Schizophrenia Bulletin. 14(41. 633-643
Hargreaves, W. J. & Shumway, M. (1989). Effectiveness o f mental health
services for the mentally ill. In C. A. Taub, D. Mechanic, & A. A. Hohmann
(Eds.), The future o f mental health services research. Washington, D.C.: U.S.
Government Printing Office.
Holloway, F., McLean, E. K., & Robertson, J. A. (1991). Case
management. British Journal of Psychiatry. 159. 142-148.
Homstra, R. J., Bruce-Wolfe, V., Sagaduyu, K., & Riffle, D. W. (1993).
The effect o f intensive case management on hospitalization of patients with
schizophrenia. Hospital and Community Psychiatry. 44(91. 844-847.
Hoult, J., & Reynolds, I. (1984). Schizophrenia: A comparative trial of
community oriented and hospital oriented psychiatric care. Acta Psvchiatrica
Scandinavica. 69. 359-372.
Hudson, W. (1982). The Clinical Measurement Package. Chicago, IL:
The Dorsey Press.
Jonsson, H., & Nyman, A. K. (1991). Predicting long-term outcome in
schizophrenia. Acta Psvchiatrica Scandinavica. 83. 342-346.
Kay, S. R., & Lindenmayer, J-P. (1987). Outcome predictors in acute
schizophrenia: Prospective significance of background and clinical dimensions.
The Journal of Nervous and Mental Disease. 175(31. 152-160.
Kirkpatrick, B., Buchanan, R. W., McKenney, P. D., Alphs, L. D., &
Carpenter, W. T. (1989). The schedule for the deficit syndrome: An instrument
for research in schizophrenia. Psychiatry Research. 30. 119-123.
Kluiter, H., Giel, R., Nienhuis, F. J., Ruphan, M ., & Wiersma, D.
(1992). Predicting feasibility of day treatment for unselected patients referred
for inpatient psychiatric treatment: Results of a randomized trial. American
Journal of Psychiatry. 149(91. 1199-1205.
118
Lewis-Beck, M. S. (1980). Applied repression: An introduction.
Newbury Park, CA: Sage Publications.
McCreadie, R. G., Wiles, D., Grant, S., Crockett, G. T., Mahmood, Z.,
Livingston, M. G., Watt, J. A. G., Greene, J. G., Kershaw, P. W., Todd, N. A.,
Scott, A. M., Loudon, J., Dyer, J. A. T., Philip, A. E., & Batchelor, D. (1989).
The Scottish first episode schizophrenia study. VII. Two-year follow-up. Acta
Psychiatric Scandinavica. 80. 597-602.
McGlashan, T. H. (1988). A selective review of recent North American
long-term follow-up studies of schizophrenia. Schizophrenia Bulletin. 14(41. 515-
542.
Moller, H. J., Schmid-Bode, W., & von Zerssen, D. (1986). Prediction of
long-term outcome in schizophrenia by prognostic scales. Schizophrenia Bulletin.
12(2), 225-235.
Muijen, M., Marks, M., Connolly, J., Audini, B., & McNamee, G. (1992).
The daily living programme: Preliminary comparison of community versus
hospital-based treatment for the seriously mentally ill facing emergency admission.
British Journal o f Psychiatry. 160. 379-384.
Munich, R. L., & Lang, E. (1993). The boundaries of psychiatric
rehabilitation. Hospital and Community Psychiatry. 44171. 661-665.
National Institute of Mental Health (1982). A network for caring: The
community support program o f the National Institute o f Mental Health.
Proceedings o f Four Nation Conferences 1978-779. (DHHS Publication No.
ADM 81-1063). Washington, D.C.: U.S. Government Printing Office.
Olfson, M. (1990). Assertive community treatment: An evaluation of the
experimental evidence. Hospital and Community Psychiatry. 41(61. 634-641.
Overall, J. E., & Gorham, D. R. (1962). The brief psychiatric rating scale.
Pvscholoeical Reports. 10. 799-812.
Prudo, R., & Blum, H. M. (1987). Five-year outcome and prognosis in
schizophrenia: A report from the London Field Research Centre o f the
International Pilot Study o f Schizophrenia. British Journal o f Psychiatry. 150.
345-354.
119
Rosenfield, S., & Neese-Todd, S. (1993). Elements of a psychosocial
clubhouse program associated with a satisfying quality o f life. Hospital and
Community Psychiatry. 44(11. 76-78.
Sands, R. G., & Cnaan, R. A. (in press). Two modes o f case management:
assessing their impact. Community Mental Health Journal.
Santos, A. B., Deci, P. A., Lachance, K. R., Dias, J. K., Sloop, T. B.,
Hiers, T. G., & Bevilacqua, J. J. (1993). Providing assertive community treatment
for severely mentally ill patients living in a rural area. Hospital and Community
Psychiatry. 4411V 34-39.
Snowden, L. R., & Clancy, T. (1990). Service intensity and client
improvement at a predominantly black community mental health center.
Evaluation and Program Planning. 13. 205-210.
Snowden, L. R., Storey, C., & Clancy, T. (1989). Ethnicity and
continuation in treatment at a black community mental health center. Journal of
Community Psychology. 17. 111-118.
Solomon, P. (1992). The efficacy o f case management services for
severely mentally disabled clients. Community Mental Health Journal. 2813 V 163-
180.
Stein, L. (1990). Comments by Leonard Stein. Hospital and Community
EsyshiatDL41(6), 649-651.
Stein, L. I., & Test, M. A. (1980). Alternative to mental hospital
treatment, I: Conceptual model, treatment program, and clinical evaluation.
Archives of General Psychiatry. 37. 392-397.
Straube, E. R., Wagner, W., Foerster, K., & Heimann, H. (1989). Findings
significant with respect to short- and medium-term outcome in schizophrenia-A
preliminary report. Prog. Neuro-Psvchopharmacologv and Biological Psychiatry.
13,185-197.
Strauss, J. S., & Carpenter, W. T. (1972). The prediction o f outcome in
schizophrenia: I. Characteristics of outcome. Archives of General Psychiatry. 27.
739-746.
Strauss, J. S., & Carpenter, W. T. (1974). The prediction o f outcome in
schizophrenia: II. Relationships between predictor and outcome variables.
Archives of General Psychiatry. 31. 37-42.
120
Strauss, J. S., & Carpenter, W. T. (1977). The prediction of outcome in
schizophrenia: III. Five year outcome and its predictors. Archives o f General
Psychiatry. 31. 159-63.
Talbott, J. A. (1988). The chronically mentally ill: What do we now
know, and why aren't we implementing what we know? New Directions for
Mental Health Services. 37. 43-58.
Taube, C. A., Morlock, L., Bums, B. J., & Santos, A. B. (1990). New
directions in research on assertive community treatment. Hospital and Community
Psychiatry. 41(61. 642-647.
Tessler, R. C. (1987). Continuity of care and client outcome.
Psychosocial rehabilitation Journal. 11(11. 39-53.
Tessler, R., & Mason, J. H. (1979). Continuity o f care in the delivery of
mental health services. American Journal of Psychiatry. 136(10). 1297-1301.
Tessler, R. C., & Goldman, H. H. (1982) The chronically mentally ill:
Assessing community support programs. Cambridge, MA: Ballinger Publishing
Co.
T ester, R. C., Willis, G., & Gubman, G. D. (1986). Defining and
measuring continuity o f care. Psychosocial Rehabilitation Journal. 10(1). 27-38.
Test, M. A. (1979). Continuity of care in community treatment. New
Directions for Mental Health Services. 2. 15-23.
Test, M. A. (1984). Community Support Programs. In A. S. Bellack
(Ed ), Schizophrenia: Treatment. Management. Rehabilitation, (pp. 347-373).
New York, NY: Grune and Stratton.
Test, M. A. (1992). Training in community living. In R. P. Liberman
(Ed.). Rehabilitation o f the Seriously Mentally 1 1 1 , ( p p . 153-170V New York, NY:
Plenum Press.
Test, M. A., & Stein, L. I. (1978). Community treatment o f the chronic
mental patient: research overview. Schizophrenia Bulletin. 4(3). 350-364.
Verghese, A., John, J. K., Rajkumar, S., Richard, J., Sethi, B. B., &
Trivedi, J. K. (1989). Factors associated with the course of outcome of
schizophrenia in India: Results o f a two-year multicentre follow-up study. British
Journal o f Psychiatry. 154. 499-503.
121
Wherley, M., & Bisgaard, S. (1987). Beyond model programs: Evaluation
of a countywide system o f residential treatment programs. Hospital and
Community Psychiatry. 38(8). 852-857.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A longitudinal cross-ethnic study of sociocentricity in schizophrenia

PDF

A longitudinal examination of gender and ethnicity across two models of community treatment for individuals with schizophrenia and related disorders

PDF

An investigation of factors involved in the inception of blood coagulation.

PDF

Biosocial antecedents of schizophrenia-spectrum personality disorders: A longitudinal study

PDF

A behavioral -analytic model for assessing social competence in male batterers

PDF

Childhood videotaped neuromotor and social precursors of schizophrenia: A prospective investigation

PDF

A correlational study of feminist/womanist identity development and depression in women

PDF

An analysis of postacute treatment and outcome differences between Medicare fee-for-service and managed care

PDF

Cognitive predictors of intention -behavior consistency in safer sex practices

PDF

Patterns of change in the course of schizophrenic symptoms

PDF

Adolescent stress and styles of coping: Predictors and moderators of psychological distress

PDF

AN ANALYSIS OF PLUTARCH'S QUOTATIONS FROM EURIPIDES.

PDF

Childhood precursors of adult reality distortion: A prospective study.

PDF

Adolescent predictors and adult consequences of alcohol dysphoria: A longitudinal study

PDF

Anatomic and functional specificity of relationships between explicit memory dysfunction and atrophy of medial temporal structures in Alzheimer's disease

PDF

An investigation of caregiver sense of control and coping strategies in moderating the effects of caregiving demands

PDF

An integrated theoretical framework of juvenile and gang delinquency causation: A structural equation model.

PDF

Childhood Sexual Abuse and Depressive Symptoms In A Lesbian Population: An Exploratory Study.

PDF

An investigation of the distribution and mechanisms of deposition of exogenous cholesterol in the rat

PDF

Atomism and social control: Causes and consequences of alcohol use.

Asset Metadata

Creator Ansel, Mark Gerard (author)

Core Title Client characteristics, service characteristics, and psychosocial outcomes in the community treatment of schizophrenia

Degree Doctor of Philosophy

Degree Program Social Work

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag health sciences, mental health,OAI-PMH Harvest,psychology, clinical,Social Work

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Brekke, John (committee chair), [illegible] (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-87810

Unique identifier UC11354848

Identifier 9636319.pdf (filename),usctheses-c17-87810 (legacy record id)

Legacy Identifier 9636319.pdf

Dmrecord 87810

Document Type Dissertation

Rights Ansel, Mark Gerard

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA