European Psychiatry 30 (2015) Contents lists available at

736–742 ScienceDirect

of studies support modern assertive health service models.

ARTICLEINF
imited for parts of continental Europe, as well as for the
O e outcome parameter. Method: We conducted a
C lled trial including adult patients with a schizophreniform disorder
nts on the Global Assessment of Functioning Scale (GAF).
d controls (TAU, n = 142) were assessed every six-month within
Article history: Received 25
ces in rural areas. Mental and functional state were rated using
January 2015 Received in
g Scale (BPRS) and the GAF. Functional limitations and
revised form 12 April 2015
e were patient-rated using the WHO-Disability Assessment
Accepted 16 April 2015
I) and the Medication Adherence Report Scale (MARS). We
Available online 21 May 2015
d models. Results: The GAF and BPRS of both groups improved
ase in the intervention group was significantly higher. In contrast,
Keywords: Assertive
WHODAS–II and MARS – neither showed a stable temporal
outreach Integrated care
nce between groups. Conclusion: Our findings only partly support
Complex interventions
ention, because of conflicting results between clinician- and
Schizophrenia Intensive
ly, the benefits of AO need to be further evaluated.
case management
European Psychiatry

jo u rn al h om epag e: ​h ttp ://ww w.eu ro p s y- jo ur n

al.co m

Original article Clinical and functional outcome of assertive outreach for patients with schizophrenic disorder:
Results of a quasi-experimental controlled trial

D. Ka ̈stner a​,​b​,​*​, D. Bu ̈ chtemann a​, I. Warnke a​​ ,​c​, J. Radisch a​​ ,​c​, J. Baumgardt a​​ ,​e​, S. Giersberg a​​ ,
K. Kopke a​​ , J. Moock a​​ , W. Kawohl a​​ ,​d​, W. Ro ̈ssler a​,​d​,​f

a ​Competence Tandem Integrated Care, Innovation Incubator, Leuphana University Lu ̈neburg, Lu ̈neburg, Germany b ​Department of Psychosomatic Medicine and
Psychotherapy, University Medical Centre Hamburg-Eppendorf and Scho ̈n Clinic Hamburg Eilbek,, Hamburg, Germany c ​Medical Psychology Unit, Hannover
​
Medical School, Hannover, Germany ​d Department for Psychiatry, Psychotherapy and Psychosomatics, Psychiatric Hospital, University of Zurich, Zurich, Switzerland ​e

Department of Social Work, Faculty of Business and Social Sciences, University of Applied Sciences, Hamburg, Germany f​ ​Institute of Psychiatry, Laboratory of Neuroscience

(LIM 27), University of Sao Paulo, Sao Paulo, Brazil

© ​2015 Elsevier Masson SAS. All rights

reserved.

//dx.doi.org/10.1016/j.eurpsy.2015.04.003
1. Introduction 4-9338/​© ​2015 Elsevier Masson SAS. All
ts reserved.
Schizophrenia is among the leading disorders causing reatment ​[28]​. These relapses are, in turn, suspected to cause
disability ​[39]​. It is associated with relapse ​[33] ​and long-lasting llness progression and further psychosocial and biological harm
negative effects on various life domains, such as impairments in 16]​. Only 14–20% of the patients diagnosed with schizophrenia
psycho- logical, social and occupational functioning ​[45]​, and ecover completely ​[33]​.
somatic health ​[35]​, as well as with increased mortality compared Against this background, it is evident that the
to the general population ​[33]​. Relapses and rehospitalisation are equirements for an effective care exceed the mere treatment of
often related to poor adherence, which is estimated to occur in the he psychiatric symptoms. Instead, it is recommended that patients
majority of patients with schizophrenia within the course of their with severe mental illnesses like schizophrenia are treated within
modern assertive community-based care systems. Depending on
he concrete composition, these interventions are labelled
* Corresponding author. University Medical Center
Assertive Community Treatment, Intensive Case Management,
Hamburg-Eppendorf, Depart- ment of Psychosomatic Medicine
ntegrated Care, or Assertive Outreach (AO) ​[10,14]​. Key
and Psychotherapy, Martinistr. 52, 20246 Ham- burg, Germany.
components of all these interventions are flexible, team-based,
Tel.: +49 40 7410 5505 1.
assertive care delivery, small case-loads, regular home visits,
E-mail address: ​d.kaestner@uke.de ​(D. crisis services, case management, psycho-education, and
Ka ̈stner). esponsibility for health
 Downloaded for Anonymous User (n/a) at University of Nahdlatul Ulama Surabaya from
ClinicalKey.com by Elsevier on October 04, 2019. ​For personal use only. No other uses without permission. Copyright ©2019. Elsevier
Inc. All rights reserved.
D. Ka ̈stner et al. / European Psychiatry 30 (2015) 736–742 737
and social needs ​[49]​. Assertive community-based models were
fragmented. Health services, social care, and rehabilitation are developed in the 1970s in the US and subsequently
implemented
financed through different funds. Outpatient medical care for and evaluated in the US and the UK. Recently,
nation-wide
severely mentally ill patients is mainly constituted of visits to implementation processes have begun in The
Netherlands and in
office-based psychiatrists, who work as independent entrepre- Denmark ​[34,48]​, whereas for other parts of
continental Europe
neurs. Team-based approaches within the community scarcely only a few projects have been studied and reported on
exist, and the cooperation of the office-based psychiatrists with ​[7,18,22]​. The latter holds true for Germany, where
even though
other professions, as well as across sectors, is individually and assertive community-based models are guideline
recommended,
regionally diverse. Possibilities of a prompt integration of care reality is dominated by a fragmented health care
system
additional resources to prevent or treat crises outside the hospital consisting of independently working psychiatrists
within the
are limited ​[8]​. In principle, case management functions could be outpatient care ​[13]​. To our knowledge, findings
from only three
performed by different professionals (e.g., social workers, outpa- German pilot projects for patients with
schizophrenia in specific
tient clinics, or ambulatory nursing services), as it is legally urban catchment areas have been published
[24,25,40,41,44]​.
established in Germany. Yet, these services are seldom standar- However, the need for modern assertive
community-based
dised and only a minority of severely mentally ill patients receives models is supported by the majority of
randomized-controlled
case management in Germany ​[13]​. trials and naturalistic studies with respect to objective (hospitali- zation, days
hospitalized, independent living, contact to the health
2.2. Study design and population care system) and subjective outcome parameters ​[14,42]​. Among the subjective
parameters, functioning and psychopathology are
Detailed descriptions of intervention and TAU, design, and among the most frequently studied. Regarding
functioning,
instruments were published in the study protocol by Bramesfeld reviews and recent single studies favour assertive
community-
et al. ​[8]​. We conducted a quasi-experimental controlled trial, based models ​[6,14,18,25,42,48]​. Despite the
comparably large
where due to practical circumstances the implementation of the evidence base, the clinical meaning of the
significant, yet relatively
AO intervention backdated the beginning of the evaluation by a small changes is unclear ​[14]​. Current evidence on
psychopathol-
few months. The recruitment took place in the practices of ogy is also mainly in support of assertive models
[3,6,15,18,25,42]​,
participating psychiatrists from May 2011 to June 2012. Due to the yet conflicting results exist ​[1,48,46]​. Finally,
the last outcome
earlier start of the real-life AO-implementation, a time difference parameter relevant in the present study,
pharmacological adher-
of a few months between AO inclusion and study inclusion might ence, was found to be a clinically important
aspect, yet it is
have occurred with patients who were recruited in the very first typically neglected in studies ​[14]​. The few existing
investigations
months (1–2 months of recruitment: n = 27 AO patients, 1– show mixed results ​[4]​, some favouring assertive
community-
3 months of recruitment: n = 50 AO patients). The allocation to based models ​[25,42]​, whereas others do not
[18,31]​. Conflicting
intervention vs. control was determined by the health insurance results were often explained by differences within
design, setting,
affiliation of the patient. This quasi-experimental design was a intervention, and control condition; namely treatment
as usual
pragmatic decision due to the fact that only two large health within a specific region ​[9,23]​.
insurance companies held the integrated care contract required to Giventhelimited
availabledata,particularlyonpharmacological
receive AO. These two companies together insure about one third adherence, as well as the regionally lacking
evidence for rural parts
of the population of Lower Saxony. Insurants from various other of Germany, the present quasi-experimental trial
aims to evaluate
health insurances formed the control group. We cannot completely the effect of an already implemented AO model
on the mentioned
rule out disparities between the insured populations, but a free outcome parameters using patient- and
clinician-ratings.
choice of the insurance exists for more than 15 years and differences are likely to be decreasing. We included
patients 2. Materials and methods
who were 18 years or older with a schizophreniform disorder (ICD 10 F2) and a maximum score of 60 on the Global
Assessment of 2.1. Intervention and treatment as usual (TAU)
Functioning Scale (GAF). Prior to the inclusion into the study, all participants gave written informed consent.
Information on the The investigated AO model is an approach to improve the
flow of participants is shown in ​Fig. 1​. We could recruit 176 patients outpatient care delivery in predominantly rural
areas (Lower
receiving AO and 142 patients receiving TAU in 17 participating Saxony) by installing ambulatory psychiatric
nursing services and
study practices. Three assessments over the course of 1 year took standardising the collaboration between
office-based psychiatrists
place: baseline (t0), 6-month follow-up (t1) and 12-month follow- and ambulatory psychiatric nurses. It is a real-life
implemented AO
up (t2). The assessments consisted of various instruments intervention, which is adapted to the existing structures of
the
completed by the treating psychiatrist and the patient, as well German health care system in rural areas. Within the
context of an
as a structured interview conducted by the practice assistant with integrated care contract need-oriented case
management, 24-h
the patient. Results regarding other relevant outcome parameters, crisis service, home treatment and
psycho-education is offered by
such as in-patient days or cost-effectiveness will be reported psychiatricnurses, whomeet regularly with the
treatingpsychiatrist
elsewhere. (i.e.,theteamleaderwithintreatmentconferences).Atthebeginning
The trial was approved by the local ethics board and is of the intervention, a treatment plan is developed for each
patient.
registered as an International Standard Randomised Controlled This plan is subsequently adapted as discussed
within the treatment
Trial (ISRCTN34900108). conferences. Another part of the intervention is a treatment guideline, which is binding to
the service providers upon signing
2.3. Clinician-rated instruments the integrated care contract. This guideline is used to assure fidelity. A management
association (IVPNetworks GmbH), which contracted
The Brief Psychiatric Rating Scale (BPRS) ​[36] ​is a widely used the statutory health insurance companies and the
service providers,
instrument to measure current psychopathology (e.g., depression,
assistedwiththeimplementation.Theassociationalsotook carethat
anxiety, hallucinations, and suicidality), especially in patients with service providers perform according to the
treatment guideline.
schizophrenia. It consists of 18 items and was completed by the The standard mental health care delivery in
Germany
treating psychiatrist following an appointment with the patient. (treatment as usual – TAU) is of high quality, but the
system is
Thereby, the presence and severity of different symptoms is judged
Downloaded for Anonymous User (n/a) at University of Nahdlatul Ulama Surabaya from ClinicalKey.com by
Elsevier on October 04, 2019. ​For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
D. 738
Ka ̈stner et al. / European Psychiatry 30 (2015) 736–742 Assessed for eligibility (n = 45 4) ​Excluded :
Not meeting inclusion criteria (n = 18 )
Missing information: inclusion criteria (n=1)
Missing information: intervention vs. control (n=3)
Nonresponder (e.g., time constraints, concerns about confidentiality, ph ysician change) (n = 11 4: n=15
AO; n= 96 TAU,
Allocated ​(n = 318) n=3
​ missing information AO vs. TAU) Intervention (n = 176) Control (n = 14 2)
Lost to follo w up
Lost to follo w up
(n=43)
(n=37)
Discontinued interventio n (n = 5)
Fig. 1. Flow of participants through each stage of the trial.
on a 7-point-scale ranging from 0–6 (not present–extremely
and reliable measure in studies and clinical practice, even if self- severe). The psychometric properties and
sensitivity to change of
report may be influenced by social desirability. The psychometric the instrument have been proven ​[32]​. We used
the total score (i.e.,
properties of the German version were evaluated in a sample with mean score) and the score of the schizophrenia
subscale as a
chronically ill patients and found to be satisfactory. Here, we measure of global psychiatric symptom severity and of
severity of
calculated the sum score (range: 5–25). The data showed a schizophrenic symptoms, respectively. This subscale
consists of
typically skewed distribution. Consequently, we dichotomized the 10 items and has a satisfactory reliability among
experienced
data into adherers and non-adherers with a cut-off score of 25. This raters, as well as proven validity ​[2]​.
Additionally, we calculated
procedure is common, although the cut-off definitions vary. We scores for the positive and negative syndromes and
general
chose the sample median ​[26]​. symptoms ​[32]​. These scales originate from Kane et al. ​[21] ​and it was found that
they have a strong correlation with the PANSS
2.5. Statistical analysis syndrome scales ​[5]​.
The Global Assessment of Functioning (GAF) is a scale to rate the
We conducted a non-responder analysis comparing sex, age, overall level of psychiatric symptoms and functioning,
ranging
baseline GAF score, and receipt of AO of study participants and from 0 -100. It is a quick measure extensively used
in clinical
patients fulfilling inclusion criteria, yet rejecting a study partici- practice and research. It is reliable and valid within
samples of
pation (i.e., non-responders). Differences between non-responders patients with severe mental illnesses ​[20,43]​.
and responders were analysed using Chi​2 ​and Mann–Whitney U- tests. Furthermore, a dropout and a missing value
analyses were 2.4. Patient-rated instruments
carried out in order to report the rate of patients who were lost to follow-up, and to ensure that data are missing at
random. Here, a We applied the patient-administered 12-item short version of
logistic regression with the dependent variable complete case (yes/ the World Health Organization–Disability
Assessment Schedule II
no) and the already mentioned predictor variables (sex, age, (WHODAS–II) ​[50]​. The instrument assesses health and
disability
baseline GAF, receipt of AO) was conducted. An analogous in the general population and clinical samples.
WHODAS–II covers
procedure was applied when analysing missing values within the six ICF domains (cognition, mobility, self-care,
getting along,
the outcome parameters, whereas the predictor variables did not life activities, and participation), which are to be
evaluated on a 5-
contain missing values. t-tests and Chi​2 ​tests were used to compare point scale ranging from ‘‘none’’ to ‘‘extreme’’
(0–4). The short
baseline characteristics between intervention and control group. version allows calculation of a sum score (range:
0–48), which
To evaluate the effect of the treatment, we calculated three- describes the extent of functional limitations. The
instrument has
level mixed models for the outcomes BPRS (psychopathology: total been employed in numerous studies and is
capable of measuring
score, subscale schizophrenia, positive and negative syndrome, change and the clinical effectiveness of
interventions. The
and general symptoms), GAF (functioning), WHODAS–II (func- psychometric properties of the instrument
evaluated in a
tional limitations), and MARS (medication adherence). Thereby, we comparable sample of patients with
schizophrenia were found
accounted for missing values as well as for the non-independence to be good ​[29]​.
of the data due to repeated measurement and clustering of patients The Medication Adherence Report Scale (MARS)
is a self-report
within practices (level 3: 17 practices, level 2: 318 patients). of adherence to the prescribed pharmacological
treatment. The
Regarding missing values, it was found that mixed modelling non-adherent behaviour is described by five items
with a scale
without any ad hoc imputation yielded the most powerful tests ranging from ‘‘always’’ to ‘‘never’’ (1–5). It is
regarded as a practical
within various different scenarios ​[12]​. We included main effects
Downloaded for Anonymous User (n/a) at University of Nahdlatul Ulama Surabaya from ClinicalKey.com by
Elsevier on October 04, 2019. ​For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
D. Ka ̈stner et al. / European Psychiatry 30 (2015) 736–742 739
for time (time effect t1 and t2 to compare baseline with the
of the control group (time effect t1 and t2). The group effects respective follow-up) and receipt of AO, as well as
interaction
signify baseline differences between the control and intervention terms between time and AO. Additionally we used
sex, age (with
group. Finally, the interaction terms between time and receipt of the reference of middle-aged male patients), and
baseline GAF
AO (time ​Â ​group) represent the effects of the AO intervention. score as covariates. The baseline GAF was included
because
Regarding the statistical models, we can conclude that GAF and all missing values in all outcome parameters, but
the GAF itself were
BPRS scores improved significantly, also for patients receiving TAU. associated with this variable. In addition to
these fixed factors,
These improvements were evident after 6 months and remained at time was incorporated as a random factor at level
2. To visualize
the 1-year follow-up time-point. Improvements in GAF and all effects, we used marginal plots, which display the
predicted means
BPRS scores after 1 year (t2) of patients receiving AO (intervention of the outcome variables when holding all other
variables at the
group – IG) were significantly larger compared to patients treated reference point (i.e., middle-aged male patients
with average GAF
in TAU. For example, the reference group of middle-aged male score at baseline). Linear models were calculated for
continuous
patients gained an average of 6.54 GAF points within AO- outcome parameters, a logistic model for the
dichotomized MARS
treatment, compared to 2.29 points in TAU over 1 year. The ​[37,38] ​(xtmixed and xtmelogit). All analyses were
carried out
marginal plots illustrate that both groups had nearly parallel using Stata 12 MP.
increases within the first half year. Afterwards, only the interven- tion group continued to improve. None of the
group effects, except
3. Results
the effect for the GAF, were significant, indicating a lower baseline functioning of the intervention group when
controlling for age and
3.1. Non-responder, dropout, and missing value analysis
sex.​In contrast to the mentioned effects for the clinician-rated
Non-responder analysis revealed that study participants (n = 318) and non-responders (n = 114) did not differ
significantly in any variable, except receipt of AO (n = 96 TAU, n = 15 AO; n = 3 missing information; ​x2​ ​= 58.71;
P ​< ​.001). Thus, patients receiving TAU were more likely to opt against study participation. With respect to the
total sample, 74.8% of the participants completed all assessments (complete cases, ​Fig. 1​). Patients with missing
assessments and complete cases did not differ significant-
Downloaded for Anonymous User (n/a) at University of Nahdlatul Ulama Surabaya from ClinicalKey.com by
Elsevier on October 04, 2019. ​For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
parameters, both patient-rated variables [i.e., functional limita- tions (WHODAS–II) and pharmacological adherence
(MARS)] showed neither a stable temporal improvement nor a difference between groups. There was a significant
reduction in functional limitations (WHODAS–II) after 6 months for the control group, but the effect did not
reappear after 1 year. No time, group, or interaction effect reached statistical significance for the MARS ratings.
ly with respect to age, sex, or receipt of AO. However, there was a statistically non-significant tendency for dropouts
to show a lower
4. Discussion baseline GAF (z = –1.95; p = .051). Missing values within WHO- DAS–II, MARS, and BPRS scores
were also associated with lower
Previous evidence has shown that assertive community-based functioning at baseline (GAF). Thus, data were
missing at random
models might be effective in fulfilling the complex needs of and we included baseline GAF as a covariate in
subsequent
patients with schizophrenia. However, at least in parts of analyses of the respective outcome parameters.
continental Europe and particularly in rural areas, their imple- mentation and evaluation is limited. Moreover,
pharmacological 3.2. Baseline characteristics
adherence is an important but rarely investigated outcome parameter. The present prospective study demonstrated
positive Baseline characteristics of the total sample, as well as for each
effects of an AO model in predominantly rural parts of Germany on group, are presented in ​Table 1​. About half of
the patients were
the clinician-rated parameters mental and functional state. female and the mean age of patients was ​$ ​45 years.
Furthermore,
However, from the patients’ perspective, neither the functional the intervention and control group were comparable
with respect
state nor the pharmacological adherence had a significantly more to these socio-demographic variables, as well as
diagnosis and
favourable course in the AO compared to the control group. family situation.
The results concerning the clinician-rated improvements in mental and functional state are in line with most of the
previous 3.3. Multilevel mixed models
findings on these outcome parameters ​[25,42]​. From the perspec- tive of the clinicians, patients in the AO group
have clearly Results of the mixed effects models are shown in ​Table 2​. ​Fig. 2
benefited from the integration of services and the support through shows the corresponding marginal plots for
middle-aged male
regular home visits, crisis services, and psycho-education. patients and the outcome parameters: BPRS total score,
GAF,
Interestingly, the AO model showed an effect not only on WHODAS–II, and MARS. The time effects indicate the
difference
 functioning and general and negative symptomatology, but also between baseline assessment and respective
follow-up for patients
on positive symptoms, usually indicative of the acute phase of
Table 1 Baseline characteristics of the total sample and intervention (AO) and control group (TAU).
Characteristic Total sample AO TAU Test statistics p value
Male (%) 169 (53.1) 94 (53.4) 75 (52.8) ​x2​​ = 0.0 0.916 Mean age (SD) 45.4 (12.6) 44.5 (12.6) 46.4 (12.5) t = 1.3
0.180 Children (%)​a ​117 (38.1) 66 (38.8) 51 (37.2) ​x2​​ = 0.1 0.775 Partner (%)​a ​109 (35.2) 60 (34.9) 49 (35.5) ​x2​​ = 0.0
0.909 Diagnosis​a ​F20.0 (%) 267 (85.8) 145 (84.8) 122 (87.1) ​x2​​ = 0.4 0.809 F20.1–F20.9 (%) 41 (13.2) 24 (14.0) 17
(12.2) F22–F25 (%) 3 (1.0) 2 (1.2) 1 (0.7)
a ​Variable contains missing values (n ​61⁄4 ​318).
D. 740 Ka ̈stner et al. / European Psychiatry 30 (2015) 736–742 Table 2 Results of the
multilevel mixed effects models.

Coefficient (standard
error)​a
Group effect
Outcome parameter​b ​Constant Time effect
AO
t1
Group effect
Time ​Â g​ roup
AO
t2-AO
Time ​Â g​ roup
Time effect
t1-AO
t2
Time ​Â g​ roup
Time effect
t1-AO
t2
Time ​Â g​ roup
Time effect
t1-AO
t2
Time ​Â g​ roup
Group effect
t1-AO
AO
Time ​Â g​ roup
Group effect
t1-AO
AO
Downloaded for Anonymous User (n/a) at University of Nahdlatul Ulama Surabaya from
 ClinicalKey.com by Elsevier on October 04, 2019. ​For personal use only. No other uses without permission. Copyright ©2019. Elsevier
Inc. All rights reserved.
BPRS total score 1.70 –.29 (.05)​*** ​–.23 (.07)​*** ​.01 (.06) –.06 (.07) –.30 (.09)​** ​BPRS schizophrenia 1.59 –.28 (.06)​*** ​–.21 (.07)​**
.00 (.07) –.04 (.08) –.26 (.10)​** ​BPRS positive 1.27 –.21 (.06)​*** ​–.22 (.07)​** ​.04 (.08) –.11 (.08) –.20 (.10)​* ​BPRS negative 1.93
–.36 (.06)​*** ​–.22 (.07)​** ​–.07 (.08) .08 (.08) –.27 (.10)​** ​BPRS general 2.09 –.33 (.08)​*** ​–.26 (.09)​** ​.06 (.10) –.14 (.10) –.48
(.12)​*** ​GAF 44.84 1.92 (.85)​* ​2.29 (.94)​* ​–2.04 (1.02)​* ​.47 (1.14) 4.23 (1.25)​** ​WHODAS–II 13.44 –1.49 (.59)​* ​–.80 (.69) .29 (.92)
1.44 (.79) .29 (.93) MARS -.18 .32 (.34) .30 (.35) –.02 (40) .06 (.45) .63 (.49)

a - Reference is t0 and the control group (TAU). All models control for sex and age (with the reference of middle-aged male patients)
and baseline GAF. b - The range of the scores is as follows: BPRS 0–6; GAF 0–100; and WHODAS–II 0–48. Because of the
logarithmic function the MARS coefficients are not as interpretable as the others. Lower values indicate an improvement for BPRS
and WHODAS–II scores, whereas GAF and MARS scores have to be interpreted the other way around; t1 = 6 month follow-up; t2 =
1 year follow-up.
* p​ ​< ​.05. ​** p​ ​<
.01. ***
​ ​p ​< ​.001.
have remained undetected. Jerrell and Ridgely ​[19] ​reported some
psychosis. However, in light of the divergence between the differences between self- and observer-rated results with respect to
clinician-rated and patient-rated results derived from instruments functioning subscale scores and a greater consistency of
aiming to assess comparable clinical parameters (WHODAS–II, observer-ratings, yet they did not discuss these differences further.
GAF – functional limitations and global functioning), the above- Presently, the sensitivity to change of the different instruments
mentioned findings must be interpreted with caution. Such a GAF and WHODAS–II) might have also played a role. Although
divergence has rarely been reported within the existing literature he instrument developers ​[47] ​stated that the sensitivity to change
evaluating AO models. This might be due to the fact that only a of the WHODAS–II is at least comparable to other functioning
very few studies have thus far incorporated both self- and cales, another study ​[29] ​also found differing results
observer- ratings on functioning ​[17,19]​, so that divergent
perceptions could
Fig. 2. Marginal plots for BPRS (total score), GAF, WHODAS–II and MARS. **​ ​ Significant interaction terms: courses with a
statistically significant effect of the intervention at the marked assessment. Please note that the graphs do not cover the full range of
the scales in order to enable visual inspection. For information on the range of the scales and the direction of interpretation see notes
below ​Table 2​.
D. Ka ̈stner et al. / European Psychiatry 30 (2015) 736–742 741
between WHODAS–II and GAF. Similarly to the present findings,
Acknowledgement only GAF scores showed significant improvements over time, both in observer-ratings. Besides a
‘‘true’’ divergence between percep-
The study is part of a project from the Innovation-Incubator at tions and the mentioned measurement issues, a bias
towards
the Leuphana University Lu ̈ neburg, which is funded by European expectancy of the treating psychiatrists
(Pygmalion effect) is an
Funds for Regional Development and the Federal State of Lower option in explaining these results. Complex
interventions under
Saxony. We thank the management association IVPNetworks naturalistic conditions usually make concealment
impossible ​[11]​,
GmbH for their support in the recruitment of study practices. and at the beginning of new interventions, service
providers might carry an enthusiasm of the early phase ​[23]​. This could have contributed to an overestimation of the
improvements of the AO
References group by the psychiatrists.
Regarding self-rated pharmacological adherence, the lack of
[1] ​Aagaard J, Mu ̈ ller-Nielsen K. Clinical outcome of assertive community treat- significant findings might be
due to a high level of adherent behaviour at baseline, as indicated by the skewness of the variable.
ment (ACT) in a rural area in Denmark: a case–control study with a 2-year follow-up. Nordic J Psychiatry
2011;65(5):299–305. ​[2] ​Andersen J, et al. The Brief Psychiatric Rating Scale. Dimension of ​Previous studies
demonstrating positive effects of assertive models
schizophrenia – reliability and construct validity. Psychopathology ​on medication adherence included patients with
first-episode schizophrenia, patients with a history of medication non-adher-
1989;22(2–3):168–76. ​[3] ​Bak M, et al. An observational, ‘real life’’ trial of the introduction of assertive community
treatment in a geographically defined area using clinical rather ​ence, or patients with dual diagnosis ​[25,27]​. In
contrast, studies
than service use outcome criteria. Soc Psychiatry Psychiatr Epidemiol ​such as ours that do not showing such effects
[18,31]​, typically use less stringent inclusion criteria regarding the risk of non-
2007;42(2):125–30. ​[4] ​Barkhof E, et al. Interventions to improve adherence to antipsychotic medica- tion in
patients with schizophrenia – areview of the past decade. Eur Psychia- ​adherence. Therefore, it might be beneficial
to define patients at
try 2012;27(1):9–18. ​risk for non-adherent behaviour and target interventions more precisely.
There are several strengths and limitations of the present study.
[5] ​Bell M, et al. The Positive and Negative Syndrome Scale and the Brief Psychi- atric Rating Scale. Reliability,
comparability, and predictive validity. J Nervous Mental Dis 1992;180(11):723–8. ​[6] ​Bjorkman T, Hansson L. Case
management for individuals with a severe mental ​Strengths lie within the comparably large sample size and the
numerous participating study practices. Furthermore, the natural- istic approach considers existing treatment
structures and allows a
illness: a 6-year follow-up study. Int J Soc Psychiatry 2007;53(1):12–22. ​[7] ​Bjo ̈rkman T, Hansson L, Sandlund
M. Outcome of case management based on the strengths model compared to standard care. A randomised controlled
trial. Soc Psychiatry Psychiatr Epidemiol 2002;37(4):147–52. ​sustainable implementation in rural areas after further
evalua- tions. This is often not the case with trials on AO models, which investigate pilot projects and include
patients within in-patient
[8] ​Bramesfeld A, et al. Effectiveness and efficiency of assertive outreach for Schizophrenia in Germany: study
protocol on a pragmatic quasi-experimental controlled trial. BMC Psychiatry 2013;13:56. ​[9] ​Burns T. Case
management and assertive community treatment in Europe. ​settings located at a university clinic in an urban area or
within a
Psychiatr Serv 2001;52(5):631–6. ​limited number of newly established specific teams ​[23]​. In contrast, the regional
expansion of the integrated care contract
[10] ​Burns T, Firn M. Assertive outreach in mental health: a manual for practi- tioners. Oxford, New York: Oxford
medical publications, Oxford University Press; 2002 [345 p.]. ​corresponding to the presently investigated
intervention suggest
[11] ​Campbell M, et al. Framework for design and evaluation of complex inter- ​that it has a good acceptability and
feasibility, as well as a potential for an easy transfer into the everyday practice, especially in rural
ventions to improve health. BMJ (Clinical research ed) 2000;321(7262):694–6. ​[12] ​Chakraborty H, Gu AH. Mixed
model approach for intent-to-treat analysis in longitudinal clinical trials with missing values. RTI Press publication;
2009 ​areas ​[30]​. Our study also has some limitations. The patient- and
[No. MR-0009-0903]. ​clinician-ratings were not assessed by the same instrument. For instance, it would have been
interesting to administer the WHODAS in the patient- and the clinician-rated form to get more
[13] ​DGPPN. S3-Leitlinie Psychosoziale Therapien bei schweren psychischen Erk-
rankungen. Berlin, Heidelberg: Springer Berlin Heidelberg; 2013. ​[14] ​Dieterich M, et al. Intensive case
management for severe mental illness.
Cochrane Database Syst Rev 2010;6:10. ​information on the stability of the clinician rating. Furthermore,
independent raters for observer-ratings (i.e., study nurses or physicians) are favourable, but could not be realised in
our study
[15] ​Drukker M, et al. A real-life observational study of the effectiveness of FACT in
a Dutch mental health region. BMC Psychiatry 2008;8:p93. ​[16] ​Emsley R, Chiliza B, Asmal L. The evidence for
illness progression after relapse
in schizophrenia. Schizophr Res 2013;148(1–3):117–21. ​due to practical reasons (long distances between the
numerous study practices). Finally, the observation period of one year might have been too short to also detect
changes in the patient-ratings.
[17] ​Harrison-Read P, et al. Heavy users of acute psychiatric beds: randomized controlled trial of enhanced
community management in an outer London borough. Psychol Med 2002;32(3):403–16. ​[18] ​Huguelet P, et al.
Effects of an assertive community program in patients with ​In sum, our study gives insight into an AO model in a
rural
severe mental disorders and impact on their families. Psychiatry Clin Neurosci ​German area with positive findings
on clinician-rated functional and mental state, suggesting that further evaluation is warranted.
2012;66(4):328–36. ​[19] ​Jerrell JM, Ridgely MS. Comparative effectiveness of three approaches to serving people
with severe mental illness and substance abuse disorders. J Nerv Ment Dis 1995;183(9):566–76. ​[20] ​Jones SH, et
al. A brief mental health outcome scale-reliability and validity of ​5. Conclusion
the Global Assessment of Functioning (GAF). Br J Psychiatry 1995;166(5): 654–9.
We evaluated an AO intervention in predominantly rural
[21] ​Kane JM, et al. Clozapine in treatment-resistant schizophrenics. Psychophar-
macol Bull 1988;24(1):62–7. ​regions in Germany and found positive effects of the investigated
[22] ​Kilian R. Evaluation von Programmen zur integrierten psychiatrischen Ver- ​AO model on mental and
functional state rated by the treating psychiatrist. Because of a divergence between these results and the self-ratings
of the patients, the present findings only partly support
sorgung. Zum Problem der Gewinnung von Evidenz fu ̈ r die Wirksamkeit und Effizienz komplexer Interventionen
in der Gesundheitsversorgung Nerven- heilkunde 2013;32(5):287–92. ​[23] ​King R. Intensive case management: a
critical re-appraisal of the scientific ​the AO model. Against this background of ambiguous results and considering
the regional importance of this AO model, we strongly recommend further research concerning its effectiveness.
evidence for effectiveness. Adm Policy Ment Health 2006;33(5):529–35. ​[24] K ​ issling W. Das ‘‘Mu ̈ nchner
Modell’’–Krankenkassen finanzieren Compliance–
Programme. Psychoneuro 2008;34(9):410–5. ​[25] ​Lambert M, et al. Assertive community treatment as part of
integrated care
versus standard care. J Clin Psychiatry 2010;71(10):1313–23. ​[26] ​Mahler C, et al. Assessing reported adherence to
pharmacological treatment ​Disclosure of interest
recommendations. Translation and evaluation of the Medication Adherence Report Scale (MARS) in Germany. J
Eval Clin Pract 2010;574–9.
The authors declare that they have no conflicts of interest
[27] ​Manuel JI, et al. Does assertive community treatment increase medication adherence for people with
co-occurring psychotic and substance use disor- ​concerning this article.
ders? J Am Psychiatr Nurses Assoc 2011;17(1):51–6.
Downloaded for Anonymous User (n/a) at University of Nahdlatul Ulama Surabaya from ClinicalKey.com by
Elsevier on October 04, 2019. ​For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
D. 742 Ka ̈stner et al. / European Psychiatry 30 (2015) 736–742 [28] Marder SR. Monitoring treatment and
managing adherence in schizophrenia. J
[39] R​ o ̈ssler W, et al. Size of burden of schizophrenia and psychotic disorders. Eur Clin Psychiatry
2013;74(10):pe21.
Neuropsychopharmacol 2005;15(4):399–409. ​[29] ​Mas-Expo ́ sito L, et al. The World Health Organization Short
Disability Assess-
[40] ​Schmid P, Steinert T, Borbe R. [Implementing models of cross-sectoral ment Schedule: a validation study in
patients with schizophrenia. Compre-
mental health care (integrated health care, regional psychiatry budget) hensive Psychiatry 2012;53(2):208–16.
in Germany: systematic literature review]. Psychiatr Prax 2013;40(8): ​[30] ​Mohamed S. Adaptation of intensive
mental health intensive case manage-
414–24. ment to rural communities in the Veterans Health Administration. Psychiatr Q
[41] ​Schmidt-Kraepelin C, Janssen B, Gaebel W. Prevention of rehospitalization in 2013;84(1):103–14.
schizophrenia: results of an integrated care project in Germany. Eur Arch ​[31] ​Morken G, Grawe RW, Widen JH.
Effects of integrated treatment on antipsy-
Psychiatry Clin Neurosci 2009;259(S2):205–12. chotic medication adherence in a randomized trial in recent-onset
schizo-
[42] S​ cho ̈ ttle D, et al. Integrated care in patients with schizophrenia. Curr Opin phrenia. J Clin Psychiatry
2007;68(4):566–71.
Psychiatry 2013;26(4):384–408. ​[32] ​Mortimer AM. Symptom rating scales and outcome in schizophrenia. Br J
[43] ​Schwartz RC. Concurrent validity of the Global Assessment of Functioning Psychiatry 2007;191(50):ps7–14.
Scale for clients with schizophrenia. Psychol Rep 2007;100(2):571–4. ​[33] ​NICE. Psychosis and schizophrenia in
adults: the NICE guideline on treatment
[44] ​Spill B, et al. Patientenzufriedenheit mit einem Programm der Integrierten and management.. London: NICE;
2014.
Versorgung fu ̈ r die Indikationen Schizophrenie und Depression. Psychiatr Prax [34] ​Nordentoft M, et al. From
research to practice: how OPUS treatment was
2013;40(3):142–5. accepted and implemented throughout Denmark. Early Interv Psychiatry
[45] ​Switaj P, et al. Disability and schizophrenia: a systematic review of experi- 2013.
enced psychosocial difficulties. BMC Psychiatry 2012;12:193. ​[35] ​Oud MJ, Meyboom-de Jong B. Somatic diseases
in patients with schizophrenia
[46] ​Sytema S, et al. Assertive community treatment in the Netherlands: a ran- in general practice: their prevalence
and health care. BMC Fam Pract
domized controlled trial. Acta Psychiatr Scand 2007;116(2):105–12. 2009;10:32.
[47] U ​ stu ̈ n TB, et al. Developing the World Health Organization Disability Assess- [36] ​Overall JE, Gorham DR.
The Brief Psychiatric Rating Scale. Psychol Rep
ment Schedule 2.0. Bull World Health Organ 2010;88(11):815–23. 1962;10:799–812.
[48] ​Van Duin D, et al. Bridging the science-to-service gap in schizophrenia care in ​[37] ​Rabe-Hesketh S, Skrondal
A. Multilevel and longitudinal modeling using stata
the Netherlands: the Schizophrenia Quality Improvement Collaborative. Int J 3 ed Volume 1: Continuous
Responses, Vol. 1. Tex: College Station, Stata Press;
Qual Health Care 2013;25(6):626–32. 2012.
[49] ​van Os J. Schizophrenia treatment: content versus delivery. Acta Psychiatr ​[38] ​Rabe-Hesketh S, Skrondal A.
Multilevel and longitudinal modeling using stata
Scandin 2009;119:29–32. 3 ed Volume 2: Categorical Responses Counts and Survival, Vol. 2. Tex: College
[50] World Health O., Measuring Health and Disability: Manual for WHO Disability ​Station, Stata Press; 2012.
Assessment Schedule WHODAS 2.0.
Downloaded for Anonymous User (n/a) at University of Nahdlatul Ulama Surabaya from ClinicalKey.com by
Elsevier on October 04, 2019. ​For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.