Time-trends in the utilization of decentralized mental health services in Norway - A natural experiment: The VELO-project

Article (PDF Available)inInternational Journal of Mental Health Systems 4(1) · March 2010with22 Reads
DOI: 10.1186/1752-4458-4-5 · Source: DOAJ
Abstract
Background There are few reports on the effects of extensive decentralization of mental health services. We investigated the total patterns of utilization in a local-bed model and a central-bed model. Methods In a time-trend case-registry design, 7635 single treatment episodes, from the specialist and municipality services in 2003-2006, were linked to 2975 individual patients over all administrative levels. Patterns of utilization were analyzed by univariate comparisons and multivariate regressions. Results Total treated prevalence was consistently higher for the central-bed system. Outpatient utilization increased markedly, in the central-bed system. Utilization of psychiatric beds decreased, only in the central-bed system. Utilization of highly supported municipality units increased in both systems. Total utilization of all types of services, showed an additive pattern in the local-bed system and a substitutional pattern in the central-bed system. Only severe diagnoses predicted inpatient admission in the central-bed system, whereas also anxiety-disorders and outpatient consultations predicted inpatient admission in the local-bed system. Characteristics of the inpatient populations changed markedly over time, in the local-bed system. Conclusions Geographical availability is not important as a filter in patients' pathway to inpatient care, and the association between distance to hospital and utilization of psychiatric beds may be an historical artefact. Under a public health-insurance system, local psychiatric personnel as gatekeepers for inpatient care may be of greater importance than the availability of local psychiatric beds. Specialist psychiatric beds and highly supported municipality units for people with mental health problems do not work together in terms of utilization. Outpatient and day-hospital services may be filters in the pathway to inpatient care, however this depends on the structure of the whole service-system. Local integration of psychiatric services may bring about additive, rather than substitutional patterns of total utilization. A large proportion of decentralized psychiatric beds may hinder the development of various local psychiatric services, with negative consequences for overall treated prevalence.

Figures

Myklebust et al. International Journal of Mental Health Systems 2010, 4:5
http://www.ijmhs.com/content/4/1/5
Open Access
RESEARCH
BioMed Central
© 2010 Myklebust et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Com-
mons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduc-
tion in any medium, provided the original work is properly cited.
Research
Time-trends in the utilization of decentralized
mental health services in Norway - A natural
experiment: The VELO-project
Lars H Myklebust*
1
, Knut W Sørgaard
2,3
, Svein Bjorbekkmo
2
, Martin R Eisemann
3
and Reidun Olstad
1,4
Abstract
Background: There are few reports on the effects of extensive decentralization of mental health services. We
investigated the total patterns of utilization in a local-bed model and a central-bed model.
Methods: In a time-trend case-registry design, 7635 single treatment episodes, from the specialist and municipality
services in 2003-2006, were linked to 2975 individual patients over all administrative levels. Patterns of utilization were
analyzed by univariate comparisons and multivariate regressions.
Results: Total treated prevalence was consistently higher for the central-bed system. Outpatient utilization increased
markedly, in the central-bed system. Utilization of psychiatric beds decreased, only in the central-bed system.
Utilization of highly supported municipality units increased in both systems. Total utilization of all types of services,
showed an additive pattern in the local-bed system and a substitutional pattern in the central-bed system. Only severe
diagnoses predicted inpatient admission in the central-bed system, whereas also anxiety-disorders and outpatient
consultations predicted inpatient admission in the local-bed system. Characteristics of the inpatient populations
changed markedly over time, in the local-bed system.
Conclusions: Geographical availability is not important as a filter in patients' pathway to inpatient care, and the
association between distance to hospital and utilization of psychiatric beds may be an historical artefact. Under a
public health-insurance system, local psychiatric personnel as gatekeepers for inpatient care may be of greater
importance than the availability of local psychiatric beds. Specialist psychiatric beds and highly supported municipality
units for people with mental health problems do not work together in terms of utilization. Outpatient and day-hospital
services may be filters in the pathway to inpatient care, however this depends on the structure of the whole service-
system. Local integration of psychiatric services may bring about additive, rather than substitutional patterns of total
utilization. A large proportion of decentralized psychiatric beds may hinder the development of various local
psychiatric services, with negative consequences for overall treated prevalence.
Introduction
The deinstitutionalization of western mental health care
since the 1950's, has been associated with an increased
quality of life and satisfaction with new community ser-
vices for many patients [1]. It can also be attributed to
increased mortality and "transhospitalisations" of former
inpatients to marginally staffed services, jail and home-
lessness [2-4]. Contemporary research, therefore, advo-
cates a balanced approach that includes both mental
hospitals and outpatient local community services [5].
The decentralized Norwegian mental health services
represent an alternative to the central mental hospitals,
were inpatient care are mainly to take place locally, in a
system of small units at community mental health centres
[6,7]. Although this type of organization may be advanta-
geous, it has largely been left scientifically unnoticed. In
an extensive literature search, only one relevant article
was found [8]. Further, none were found on the dynamics
between local psychiatric beds and other local mental
health services. The decentralized Norwegian Mental
* Correspondence: lars.henrik.myklebust@unn.no
1
Psychiatric Research Centre of Northern Norway, University Hospital of North
Norway, Tromsø, Norway
Full list of author information is available at the end of the article
Myklebust et al. International Journal of Mental Health Systems 2010, 4:5
http://www.ijmhs.com/content/4/1/5
Page 2 of 7
Health Services, therefore, represent a relevant scientific
focus.
Norwegian Mental Health Services
After a slow and uncoordinated initial phase of deinstitu-
tionalisation [9], a white paper on a national public sys-
tem of mental health services was presented to The
Norwegian Parliament [10]. The system were divided in
three administrative levels:
• The 1
st
level of Municipality services, staffed by
semi-specialized personnel and GPs.
• The 2
nd
level of local Community Mental Health
Centres (CMHC).
• The 3
rd
level of Central Mental Hospitals (CMH).
Both the 2
nd
and 3
rd
level-services are staffed with
psychiatrists, clinical psychologists and specialist
nurses.
The CMHC are the core component of the system, with
duties to provide and coordinate services in geographi-
cally defined sectors [7]. Local adaptation has led to
diversity regarding their structure and size, keeping,
within the frames of national policies of health insurance,
case registration, and clinical standards [11].
The VELO-project
In 2005, the neighbouring CMHCs of Vesterålen and of
Lofoten drew scientific attention because of noticeable
organizational dissimilarities, whilst at the same time a
strong resemblance in the catchment areas' characteris-
tics and needs [12]. This opportunity to study mental
health services, in a close to natural experiment, has been
well described in a previous publication [13].
In short, the main differences between the two CMHCs
are in their number of outpatient services and local psy-
chiatric beds. The CMHC of Vesterålen has 1 outpatient
clinic (which incorporates both -consultations and vari-
ous forms of day-care treatments), and 20 psychiatric
beds at 3 fully staffed units. The CMHC of Lofoten has 2
outpatient clinics (1 until 2005) and 2 fully staffed day-
hospitals (1 until 2005), and a maximum of 6 beds avail-
able at the local general hospital. Consequently, the staff
sizes of the centres are respectively 77 versus 29 man-
labour-years. Both may also refer patients to the Nord-
land County's only Central Mental Hospital (CMH)
located in the distant town of Bodø.
In the following, the service-system of Vesterålen will
be termed as a "local-bed model", in contrast to a "cen-
tral-bed model" of Lofoten. Their outline is illustrated in
Figure 1.
Theoretical perspectives and hypotheses
Several theoretical perspectives may be relevant in the
study of mental health services.
The influential Stage model on patients' pathways to
care, is built upon 5 sequential stages, with 4 correspond-
ing filters of selective permeability [14]. Of particular
interest for our study is the 4
th
-filter between specialist
outpatient services and hospitals, which regulates the
flow of patients in and out of psychiatric beds. One deter-
minant of its permeability, is the availability of services,
were the inverse relationship of geographical distance
and bed-utilization is known as " Jarvis' Law" [15,16].
Another is the work of outpatient services, although the
literature is inconclusive as to whether these reduce or
increase bed utilization [17,18].
The recent Hydraulic model [19] postulates that the
"pressure of morbidity" will tend to equalize throughout a
system of services and that the behaviours of one compo-
nent will affect (all) the others. There are, to our knowl-
edge, no studies that have explicitly tested for the
implications of this model. With regard to our study, the
model may broaden the dynamic scope into also consider
the highly supported municipality units' relation with
specialist psychiatric services.
A somewhat different line of reasoning stems from the
phenomenon of "supply-induced demand", that suggests
a positive association between provision and utilization
of hospital beds [20,21]. This phenomenon is also sug-
gested to be relevant for mental health inpatient care [22],
but how it is affected by activities of complementary ser-
vices has not been studied.
Figure 1 Outline of the total mental health services in the two
sectors of Vesterålen and Lofoten, County of Nordland, Norway.

ͳ
ʹͲ͵
Ǧ
͹͹ȋǦ

Ȍ
ͳ ȋʹȌ
ͳȋʹȌǦ
͸

ǦǡǡǤǡȋȌ








Myklebust et al. International Journal of Mental Health Systems 2010, 4:5
http://www.ijmhs.com/content/4/1/5
Page 3 of 7
Aims
The aim of this study was to investigate the utilization of
different types of mental health services, and the associa-
tion between them. We formulated the following scien-
tific questions, and deduced specific hypotheses from the
relevant theories:
1. Is utilization of psychiatric beds affected by geo-
graphic distance to services?
a. Hypothesis: Utilization of psychiatric beds will
be continually higher in the local-bed system than
in the central-bed system.
2. Is utilization of psychiatric beds affected by the
development of highly supported municipality units?
a. Hypothesis: Increased activities of highly sup-
ported municipality units will be associated with a
decrease in utilization of psychiatric beds in both
systems.
3. Is utilization of psychiatric beds affected by
changes in the activities of specialist outpatient- and
day-hospital services?
a. Hypothesis: An increase in utilization of outpa-
tient services and day-hospitals will be associated
with a decrease in utilization of psychiatric beds
in both systems.
b. Hypothesis: A change in utilization of outpa-
tient services will be associated with altered indi-
cations for inpatient admission in both systems.
c. Hypothesis: Total utilization of services will be
higher for patients in a local-bed system than a
central-bed system.
Methods
The study had a time-trend case-registry design, with an
observational period from 2003-2006. Approvals were
given from The Regional Ethics Committee for Research
and the Norwegian Data Protection Agency. The Direc-
torate for Health and Social Affairs released all health
professionals from their confidentiality obligations.
The psychiatric services
A standard sample of socio-demographic variables, clini-
cal information, type and amount of treatment, is rou-
tinely recorded for all patients referred to the Norwegian
specialist psychiatric services. These registries were
linked across service-levels by use of the patients' Social
Id-number. This allowed for aggregation of 7635 single
treatment episodes for 2979 individual patients. All epi-
sodes were calculated into total treatments for each year
of the observation period. Missing data was collected
from clinical journals.
The sample consisted of patients between the ages of 18
and 65, with a population size of N = 18 265 for the local-
bed system and N = 12 733 for the central-bed system.
The net change of population size over the observational
period was respectively -146 and 0 persons. Patients who
belonged to municipalities outside the catchment areas
(N = 22) or moved between the two areas (N = 8) were
excluded.
The diagnoses were grouped into 4 broad categories
according to the ICD-10 manual, except for 9 patients
with serious affective disorders who were classified into
the category of psychosis, due to relevant features of their
disorder. In order to obtain an adequate size of categories
for analysis, we collapsed less-frequent diagnoses into
"others". We retained the category of "psychiatric obser-
vation" for patients who had been discharged without a
diagnostic conclusion.
Municipality services
Data on residents in highly supported municipality units
was obtained by questionnaires to the local administra-
tive authorities. The total number of all 27 residents were
subsequently linked to the psychiatric case-registries by
their social Id-number.
Analysis
The dependent variable for analysis was utilization of ser-
vices. This was defined by the number of patients treated
and sum of treatments, for all modalities: inpatient days
& nights, outpatient consultations, days in day-hospitals
(i.e. various day-care treatments in the local-bed system),
and days & nights in highly supported municipality units.
In addition, we used the concept of "bed-equivalents" to
convey inpatient utilization into terms of single beds. 1
bed-equivalent equals 365 inpatient days & nights, con-
tinuous availability assumed.
Statistics
The two systems were compared by rates per 1000 inhab-
itants, for all parameters. Differences were tested by stan-
dard procedure for confidence intervals of proportions.
No adjustment in population-structures of gender and
age was made, due to small and non-significant differ-
ences between the two catchment areas [23]. We also per-
formed a univariate analysis of the two systems inpatient-
populations at the start (2003) and end (2006) of the
observational period. The relation between utilization of
psychiatric beds and other types of services was analysed
by a series of multivariate logistic regressions, performed
for each service-system separately, at the beginning and
end of the observational period (2003 and 2006). The
dependent variable was whether the patient was hospital-
ized or not. The covariates were gender, age, diagnosis
and utilizations of outpatient consultations or day-hospi-
tal. Outpatient consultations and days in day-hospital
were log-transformed due to skewed distribution. All
covariates were entered simultaneously.
The calculation and statistical analyses was carried out
by Microsoft Excel 12.1.3 and SPSS 16.1 software.
Myklebust et al. International Journal of Mental Health Systems 2010, 4:5
http://www.ijmhs.com/content/4/1/5
Page 4 of 7
Results
Differences in patterns of utilization between the two sys-
tems are reported in tables, while changes within the sys-
tems are reported in text only (p-levels).
Total prevalence of patients treated
The total treated prevalence of patients was consistently
higher in the central-bed system, except during 2005. The
difference between the two systems increased from 13.8%
in 2003 to 17.2% in 2006. Within each system, the net
increase was of 14% (p < .000) for the central-bed system,
and 10% (p < .05) for the local-bed system. See upper part
of Additional file 1.
Psychiatric beds
For the number of psychiatric inpatients per 1000 inhab-
itants, there was no difference between the two systems
in the observational period, except for a 32% higher rate
in the central-bed system in 2004. There was no net
change within either of the systems. For bed-equivalents,
the difference between the two systems increased from
none in 2003, to a 15.5% lower rate for the central-bed
system in 2006. Within the systems, the net reduction
was of 10.1% (p < .000) for the central-bed system, while
no change for the local-bed system (see Additional file 1).
Highly supported municipality units
For utilization of beds in highly supported municipality
units, the local-bed system consistently outnumbered the
central-bed system, which did not have such until 2006.
Within the systems, the net increase in bed-equivalents
was about 40% for the local-bed system, from 0.71 per
1000 in 2003, to 1.04 in 2006 (p < .000). For the central-
bed system, the net increase was from 0 per 1000 in 2003,
to 0.31 per 1000 in 2006 (see Additional file 1).
Total inpatient utilization
When psychiatric beds and beds in highly supported
municipality units were combined, the local-bed system
consistently utilized inpatient care at nearly twice the rate
of the central-bed system. For the rate of patients, there
was no difference over the observational period, except
for a 28% higher rate for the central-bed system in 2004.
Within the systems there was no significant net change
over the observational period. For bed-equivalents, there
was a 45% higher rate for the local-bed system than the
central-bed system both in 2003 and 2006. Within the
systems, there was a net increase of 24% for the local-bed
system, 17% for the central-bed system (p < .000). See
lower part of Additional file 1.
Outpatient consultations
For the ratio of outpatients compared to the total sample
of patients, the difference between the two systems
changed from an 8% higher rate for the local-bed system
in 2003, to none in 2006 Within the systems, there was a
net increase of 5% for the central-bed system (p < .000),
while there was no change for the local-bed system. For
outpatients per 1000 inhabitants, the difference between
the two systems increased from none in 2003, to a 15%
higher rate for the central-bed system in 2006. Within the
systems, there was a net increase of 19% (p < .000) for the
central-bed system, while no change for the local-bed
system. For the number of consultations per 1000 inhab-
itants, the difference between the systems was a 28.3%
higher rate for the central-bed system in 2003, decreasing
to a 22.9% higher rate in 2006. Within the systems, there
was a net increase of 5.3% for the central-bed system, and
of 10% for the local-bed system (p < .000). See upper part
of Additional file 2.
Day-hospital
For the ratio of day-hospital patients compared to the
total sample of patients, there was no difference between
the two systems during the observational period. For day-
hospital patients per 1000 inhabitants, the difference
between the two systems increased over the observa-
tional period from none in 2003, to a 90% higher rate for
the central-bed system in 2006. Within the systems, there
was a net increase of 380% (p < .000) for the central-bed
system, while there was no change for the local-bed sys-
tem. For number of day-hospital days per 1000 inhabit-
ants, the difference between the two systems was reduced
over the observational period from 95.0% higher rate for
the central-bed system in 2003, to 70.0% in 2006. Within
the systems, there was a net increase of 87.5% (p < .000)
for the local-bed system and 21.0% (p < .000) for the cen-
tral-bed system (see Additional file 2).
Characteristics of inpatient populations
The distribution of age, gender and diagnosis of inpa-
tients at the start (2003) and end (2006) of the observa-
tional period are presented in Additional file 3.
For inpatients with psychosis or affective disorders,
there were no differences between the two systems.
For inpatients with substance abuse, the central-bed
system consistently admitted more than the local-bed
system, although the local-bed system more than doubled
these admissions (p < .000) over the observational period.
There was no change within the central-bed system.
For inpatients with anxiety-disorders, there was no dif-
ference between the systems at the beginning of the
observational period, however the local bed systems ulti-
mately admitted more than the central-bed system.
Within the systems, the central-bed system reduced these
admissions with more than 50% (p < .000), while there
was no change for the local-bed system.
For gender, the central bed system initially admitted
more male inpatients than the local-bed system, however
Myklebust et al. International Journal of Mental Health Systems 2010, 4:5
http://www.ijmhs.com/content/4/1/5
Page 5 of 7
there was no difference at the end of the observational
period. Initially there was no difference between the sys-
tems used for female inpatients, however the local-bed
system admitted more females than the central-bed sys-
tem at the end of the period. Within the systems there
were no significant changes for either gender.
Predictors of inpatient admission
In order to explore if admission patterns changed during
the observational period, logistic regressions were per-
formed, using inpatient admission (yes/no) as the depen-
dent variable. The first (2003) and last (2006) year of the
observational period was used for the analyses. The series
of logistic regressions on the predictors for psychiatric
inpatient admissions were all statistically significant and
distinguished between patients being hospitalised or not.
See Additional file 4 for the local-bed system, and Addi-
tional file 5 for the central-bed system.
Local-bed system
For 2003, the model of the local-bed system was highly
significant at a χ
2
109.641, df = 9 (p < .000), and explained
between15.0% (Cox and Snell R Square) and 23.5%
(Nagelkerke R Square) of the variance. For 2006, the cor-
responding values were χ
2
132.528, df = 9 (p < .000),
16.4% (Cox and Snell R Square) and 25.4% (Nagelkerke R
Square). The predictors for inpatient treatment were (in
2003): psychosis, outpatient consultations, age, and psy-
chiatric examination, and (in 2006): psychosis, substance
abuse, outpatient consultations, and psychiatric examina-
tion.
Central-bed system
For 2003, the model for the central-bed system was highly
significant at a χ
2
58.412, df = 9 (p < .000), and explained
between 10.4% and 16.0% of the variance. For 2006, the
corresponding values were χ
2
116.031, df = 9 (p < .000),
17.4% - 29.2%. The predictors for inpatient treatment
were (in 2003): psychosis, substance abuse, and affective
disorders, and (in 2006): psychosis, substance abuse, and
affective disorders, with the addition of outpatient con-
sultations as a negative
predictor.
In sum, the characteristics of inpatient population and
predictors for inpatient admissions suggest that indica-
tions for utilization of psychiatric beds expanded mark-
edly over time in the local bed model only, and ultimately
diverged between the two systems.
Discussion
Several findings emerged from our comparative study on
time-trends of utilization in a "central-bed model" versus
a "local-bed model" of services. Treated prevalence was
consistently higher for the central-bed system of services.
Utilization of highly supported municipality units
increased in both systems. Outpatient utilization
increased markedly in the central-bed system, while there
was a modest increase in the local-bed system. Utilization
of psychiatric beds decreased only in the central-bed sys-
tem. The total pattern of utilization was additive in the
local-bed system, in contrast to a more substitutional
association in the central-bed system. Only severe diag-
noses predicted inpatient admission in the central-bed
system, whereas also anxiety-disorders and outpatient
consultations predicted inpatient admission in the local-
bed system. Characteristics of the inpatient populations
changed markedly over time, only in the local-bed sys-
tem.
Geographic availability and utilization of psychiatric beds
The substantial changes in utilization of psychiatric beds
in the central-bed system over the relative short 4-year
observational period, suggest that geographical availabil-
ity is not important as a filter in patients' pathway to inpa-
tient care. It is in accordance with preliminary findings
from Italy [24]. It may be that the longstanding phenome-
non of "Jarvis' law" are rooted in data that stem from the
beginning of the 1900's [25,26] which can probably not be
generalized to contemporary societies.
One interpretation may be that economical factors
such as modern health insurance-systems are more cru-
cial than geographical distance for utilization of inpatient
services. This further leads to a greater relative impor-
tance of local psychiatric personnel as gatekeepers for
inpatient care, rather than the presence of local beds.
Under a public health-insurance system, local psychia-
trists or GPs will certify hospitalization regardless of the
individual patients' financial or behavioural abilities for
transport.
Utilization of psychiatric beds may therefore rather be
considered as affected by other factors that change more
rapidly over time than geographic availability, for
instance the workings of other services.
The association between psychiatric beds and beds in
highly supported municipality units
The decreasing utilization of psychiatric beds in the cen-
tral-bed system may be associated with an increase of
highly supported municipality units. However, in the
local-bed system, utilization of both these types of inpa-
tient services increased over the observational period.
The most plausible interpretation is therefore that utili-
zations of these two types of services do not affect each
other, despite their similar features of 24 hours of care.
They may serve different needs, or their use is uncoordi-
nated because they belong to different administrative lev-
els. A hydraulic analogy is consequently of limited
relevance for this relation, which may be better character-
ized by the phenomenon of supply-induced demand. This
Myklebust et al. International Journal of Mental Health Systems 2010, 4:5
http://www.ijmhs.com/content/4/1/5
Page 6 of 7
in accordance with the trend of "reinstitutionalization"
evident in recent studies, that have combined utilization
of several types residential care [27-29]. Utilization of
psychiatric beds may therefore be affected by the activi-
ties of other psychiatric services, i.e. outpatient and/or
day-hospital treatments.
The association between psychiatric beds and outpatient
services
For the central-bed system, the increase in utilization of
outpatient services is contingent with a decrease in utili-
zation of psychiatric beds. For the local-bed system there
is instead a positive association in these two types of ser-
vices. Structural aspects of the whole system of services,
rather than only the needs of the patients, may therefore
be relevant to explain the diverging results on the rele-
vance of outpatient services as a filter on the pathway to
inpatient care [17,22,30,31]. One hypothesis may be that
if both in- and outpatient services are integrated in terms
of local management or clinical relations, patterns of total
utilization will be additive rather than substitutional. This
is further supported by our results on the differences in
predictors for inpatient admission, were an association
between outpatient and inpatient treatment is only found
in the local-bed system. Also, the tendency to admit
patients with anxiety-disorders and those under psychiat-
ric examination are found only in the local-bed system.
Further, a decentralization of psychiatric beds may be
so resource-demanding that they hinder the development
of other local services. Small local-bed units with 24 -
hour staffing need a relatively high number of personnel
organized in shifts. These persons are linked to a geo-
graphical site, with specialized buildings and facilities for
psychiatric patients. This may decrease the flexibility of
the health-care system, and reduce the possibility to con-
vert resources into different types of services. Our results
on higher total treated prevalence for the central-bed sys-
tem compared to the local-bed system support this line of
reasoning. The set of services in the central-bed system
may be more balanced and relevant to the total needs of
the local population, compared to the local-bed systems'.
Consequently, heavy emphasis on decentralized psychiat-
ric beds may interfere with the national objectives of
increased availability through provision of various local
services.
Strengths and limitations
A major advantage of the study was the opportunity to
explore patterns of health service utilization in a close to
natural experiment. Both the geographical and demo-
graphic similarities of the catchment areas, the absence of
private service providers, and the overarching national
clinical standards and public health insurances help to
rule out confounding variables in interpretation of the
results.
A related advantage was the permission to use the 11-
digit Social Id-number for all patients. This allowed for
analysis on the level of individual patients throughout the
total service-system and is to our knowledge the first
time in Norwegian health-service research.
The results emerge in the cultural context of the Nor-
wegian society, which may limit generalizations to other
countries. These limitations are on the other hand inher-
ent in all health-service research and advice us to inter-
pret all internationally published results cautiously. Our
results may be of special relevance to other modern soci-
eties with similar structures of settlements and geogra-
phy.
Conclusions
Geographical availability is not important as a filter in
patients' pathway to inpatient care, and the association
between distance to hospital and utilization of psychiatric
beds may be an historical artefact. Under a public health-
insurance system, local psychiatric personnel as gate-
keepers for inpatient care may be of greater importance
than the availability of local psychiatric beds. Specialist
psychiatric beds and highly supported municipality units
for people with mental health problems do not work
together in terms of utilization. Outpatient and day-hos-
pital services may be filters in the pathway to inpatient
care, however this depends on the structure of the whole
service-system. Local integration of psychiatric services
may bring about additive, rather than substitutional pat-
terns of total utilization. A large proportion of decentral-
ized psychiatric beds may hinder the development of
various local psychiatric services, with negative conse-
quences for overall treated prevalence.
Additional material
Competing interests
The authors declare that they have no competing interests.
Additional file 1 Utilization of residential treatment/care: psychiatric
beds and highly supported living units in the period of 2003 - 2006,
comparing a psychiatric local-bed system to a central-bed system. *P
< .001 and **P < .05 when systems are compared.
Additional file 2 Total treated prevalence, utilization of outpatient
and day-hospital services over the period of 2003 - 2006, a local-bed
system versus a central-bed system. *P < .001 and **P < .05 when mod-
els are compared.
Additional file 3 Inpatient-population characteristics in a local-bed
system versus a central-bed system. The years of 2003 and 2006.
Additional file 4 Predictors of inpatient treatment (no/yes) in a local-
bed system. The years of 2003 and 2006. Logistic regression model.
Additional file 5 Predictors of inpatient treatment (no/yes) in a cen-
tral-bed system. The years of 2003 and 2006. Logistic regression model.
Myklebust et al. International Journal of Mental Health Systems 2010, 4:5
http://www.ijmhs.com/content/4/1/5
Page 7 of 7
Authors' contributions
All authors participated in the design of the study. LHM wrote the manuscript.
LHM and RO performed the statistical analysis. All authors read and approved
the final manuscript.
Acknowledgements
The study was initiated and funded by the Northern Norway Regional Health
Authorities.
Author Details
1
Psychiatric Research Centre of Northern Norway, University Hospital of North
Norway, Tromsø, Norway,
2
The Nordland Hospital Trust, Bodø, Norway,
3
The
University of Tromsø, Institute of Psychology, Faculty of Health Sciences,
Tromsø, Norway and
4
The University of Tromsø, Institute of Clinical Medicine,
Faculty of Health Sciences, Tromsø, Norway
References
1. Thornicroft G, Bebbington P: Deinstitutionalisation: From hospital
closure to service development. The British Journal of Psychiatry 1989,
155:739-754.
2. Lamb HR, Bachrach LL: Some perspectives on deinstitutionalization.
Psychiatric Services 2001, 52:1039-1045.
3. Hansen FH, Haga D, Vandvik E, Roland P, Bryne H, Aarseth HP:
Samhandling og desentralisering. Forslag til overordnet strategi for de
regionale helseforetakene. Bodø: Helse Nord RHF, Helse Sør RHF, Helse
Øst RHF, Helse Vest RHF, Helse Midt-Norge RHF; 2004.
4. Hansen V, Jacobsen BK, Arnesen E: Cause-specific mortality in
psychiatric patients after deinstitutionalisation. Br J Psychiatry 2001,
179:438-443.
5. Thornicroft G, Tansella M: Components of a modern mental health
service: a pragmatic balance of community and hospital care:
Overview of systematic evidence. British Journal of Psychiatry 2004,
185:283-290.
6. The Directorate for Health and Social Affairs: National Strategic Plan on
Mental Health. Volume 1. Oslo: The Directorate for Health and Social
Affairs; 1999.
7. The Directorate for Health and Social Affairs: Community Mental Health
Centres - facing the primary care and backed by specialized hospital
services. Oslo: The Directorate for Health and Social Affairs; 2006.
8. Gebhardt RP, Schmidt-Michel PO: [An acute psychiatric ward moves into
the community. An empirical test of the satellite model]. Nervenarzt
2002, 73:1088-1093.
9. Lavik NJ: Trends in the organization of mental health services in
Norway. International Journal of Mental Health 1987, 16:164-169.
10. Ministry of Health and Care Services: Report to the Norwegian
Parliament nr.25: Openness & entirety. About psychiatric illness and
the services. Ministry of Health and Care Services. Oslo; 1997.
11. Ruud T, Lidal E, Røme K, Sitter M, Hagen H: Distriktspsykiatriske sentre i
Norge 2003. Oslo: SINTEF; 2004.
12. Nymann A: Selection and use of need-indexes for mental health
services in the catchment areas of two Community Mental Health
Centres in the County of Nordland, Norway. In Matergrade Thesis The
University of Tromsø, Faculty of Social Science, Institute of Psychology;
2006.
13. Myklebust LH, Sorgaard K, Bjorbekkmo S, Nymann A, Molvik S, Olstad R:
Bed utilization in two differently organized community mental health
services in Northern Norway: the VELO-project. Soc Psychiatry Psychiatr
Epidemiol 2008.
14. Goldberg D, Huxley P: Mental illness in the community The University of
Tromsø, Faculty of Social Science, Institute of Psychology; 1980.
15. Smith CA, Wright D, Day S: Distancing the mad: Jarvis's law and the
spatial distribution of admissions to the Hamilton Lunatic Asylum in
Canada, 1876-1902. Social science & Medicine 2007, 64:2362-2377.
16. Sohler KB, Thompson JD: Jarvis' Law and the Planning of Mental Health
Services. Public Health Reports 1970, 35:.
17. Burgess P, Bindman J, Leese M, Henderson C, Szmukler G: Do community
treatment orders for mental illness reduce readmission to hospital? An
epidemiological study. Social Psychiatry & Psychiatric Epidemiology 2006,
41:574-579.
18. Burns T, Catty J, Dash M, Roberts C, Lockwood A, Marshall M: Use of
intensive case management to reduce time in hospital in people with
severe mental illness: systematic review and meta-regression. British
Medical Journal 2007, 335(7615):336.
19. Thornicroft G, Tansella M: The Mental Health Matrix Cambridge, United
Kingdom; 1999.
20. Roemer MI: Bed supply and hospital utilization: a natural experiment.
Hospitals 1961, 35:36-42.
21. Ginsburg PB, Koretz DM: Bed availability and hospital utilization:
estimates of the "Roemer effect". Health Care Financ Rev 1983, 5:87-92.
22. Hansson L, Sandlund M, Gostas G, Kastrup M, Muus S, Nieminen P,
Zandren T, Oiesvold T, Saarento O: The Nordic comparative study on
sectorized psychiatry - Utilization of psychiatric hospital care related to
amount and allocation of resources to psychiatric services. Social
Psychiatry and Psychiatric Epidemiology 1996, 31:327-335.
23. Myklebust LH, Sorgaard K, Bjorbekkmo S, Nymann A, Molvik S, Olstad R:
Bed utilization in two differently organized community mental health
services in Northern Norway: the VELO-project. Soc Psychiatry Psychiatr
Epidemiol 2009, 44:550-557.
24. Tansella M: Integrating Evaluative Research and Community-based
Mental Health Care in Verona. The Role of the Case Register. In The
Quality Register Conference 2008 Tromsø, Norway; 2008.
25. Jarvis E: The influence of distance from and proximity to an insane
hospital, on its use by any people. Boston Medical Surgery Journal
1850:209-222.
26. Smith CA, Wright D, Day S: Distancing the mad: Jarvis's law and the
spatial distribution of admissions to the Hamilton Lunatic Asylum in
Canada, 1876-1902. Social science & Medicine 2007, 64:2362-2377.
27. Turner T, Priebe S: Forget community care--reinstitutionalisation is here.
British Journal of Psychiatry 2002, 181:253.
28. Priebe S, Badesconyi A, Fioritti A, Hansson L, Kilian R, Torres-Gonzales F,
Turner T, Wiersma D: Reinstitutionalisation in mental health care:
comparison of data on service provision from six European countries.
British Medical Journal 2005, 330:123-126.
29. Priebe S, Turner T: Reinstitutionalisation in mental health care. British
Medical Journal 2003, 326:175-176.
30. Mattioni T, Di Lallo D, Roberti R, Miceli M, Stefani M, Maci C, Perucci CA:
Determinants of psychiatric inpatient admission to general hospital
psychiatric wards: an epidemiological study in a region of central Italy.
Soc Psychiatry Psychiatr Epidemiol 1999, 34:425-431.
31. Tyrer P, Jayakody M: A randomised control study of close monitoring of
vulnerable psychiatric patients. The Lancet 1995, 345:756-759.
doi: 10.1186/1752-4458-4-5
Cite this article as: Myklebust et al., Time-trends in the utilization of decen-
tralized mental health services in Norway - A natural experiment: The VELO-
project International Journal of Mental Health Systems 2010, 4:5
Received: 12 June 2009 Accepted: 31 March 2010
Published: 31 March 2010
This article is available from: http://www.ijmhs.com/content/4/1/5© 2010 Myklebust et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.International Journal of Mental Health Systems 2010, 4:5
    • "As the deinstitutionalization process unfolded , however, policy planners and healthcare providers in North America and Western Europe began to realize the unanticipated consequences of this revolution in the mental health field. The deinstitutionalization movement is deemed successful if one focuses only on the benchmarks found in administrative datasets, typically used for reimbursement purposes, or in census of mental health facilities: closure of hospitals and asylums; cuts in the number of beds; decrease in rates of inpatient admission, bed rotation factor, average length of stay, and number of residents71727374. Yet, many countries continue to rely on mental hospitals as the main hubs of mental health care. "
    [Show abstract] [Hide abstract] ABSTRACT: Background Policies generate accountability in that they offer a standard against which government performance can be assessed. A central question of this study is whether ideological imprint left by policy is realized in the time following its adoption. National mental health policy expressly promotes the notion of deinstitutionalization, which mandates that individuals be cared for in the community rather than in institutional environments. Methods We investigate whether mental health policy adoption induced a transformation in the structure of mental health systems, namely psychiatric beds, using panel data on 193 countries between 2001 and 2011. Results Our striking regression results demonstrate that late-adopters of mental health policy are more likely to reduce psychiatric beds in mental hospitals and other biomedical settings than innovators, whereas they are less likely than non-adopters to reduce psychiatric beds in general hospitals. Conclusions It can be inferred late adopters are motivated to implement deinstitutionalization for technical efficiency rather than legitimacy reasons.
    Full-text · Article · Nov 2014