Available via license: CC BY 4.0
Content may be subject to copyright.
Epidemiology and Psychiatric
Sciences
cambridge.org/eps
Original Article
Cite this article: Loubière S, Loundou A,
Auquier P, Tinland A (2023). Psychiatric
advance directives facilitated by peer workers
among people with mental illness: economic
evaluation of a randomized controlled trial
(DAiP study). Epidemiology and Psychiatric
Sciences 32, e27, 1–9. https://doi.org/10.1017/
S2045796023000197
Received: 27 September 2022
Revised: 10 March 2023
Accepted: 25 March 2023
Keywords:
health economics; mental health; other
psychosocial techniques/treatments;
psychiatric services; randomized controlled
trials
Corresponding author: S. Loubière,
Email: sandrine.loubiere@univ-amu.fr
© The Author(s), 2023. Published by
Cambridge University Press. This is an Open
Access article, distributed under the terms of
the Creative Commons Attribution licence
(http://creativecommons.org/licenses/by/4.0),
which permits unrestricted re-use,
distribution and reproduction, provided the
original article is properly cited.
Psychiatric advance directives facilitated by
peer workers among people with mental
illness: economic evaluation of a randomized
controlled trial (DAiP study)
S. Loubière1,2, A. Loundou1,2, P. Auquier1,2and A. Tinland2,3
1Department of Clinical Research and Innovation, Support Unit for Clinical Research and Economic Evaluation,
Assistance Publique – Hôpitaux de Marseille, Marseille, France; 2Health Service Research and Quality of Life
Center (UR 3279), Aix-Marseille University, School of Medicine, Marseille, France and 3Department of Psychiatry,
Assistance Publique – Hôpitaux de Marseille, Marseille, France
Abstract
Aims. We aimed to assess the cost-eectiveness of psychiatric advance directives (PAD) facil-
itated by peer workers (PW-PAD) in the management of patients with mental disorders in
France.
Methods. In a prospective multicentre randomized controlled trial, we randomly assigned
adults with a Diagnostic and Statistical Manual of Mental Disorders, h edition diagnosis
of schizophrenia, bipolar I disorder or schizoaective disorders, who were compulsorily hos-
pitalized in the past 12 months, to either ll out a PAD form and meet a peer worker for
facilitation or receive usual care. We assessed dierences in societal costs in euros (€) and
quality-adjusted life-years (QALYs) over a year-long follow-up to estimate the incremental
cost-eectiveness ratio of the PW-PAD strategy. We conducted multiple sensitivity analyses
to assess the robustness of our results.
Results. Among the 394 randomized participants, 196 were assigned to the PW-PAD group
and 198 to the control group. Psychiatric inpatient costs were lower in the PW-PAD group than
the control group (relative risk, −0.22; 95% condence interval, [−0.33 to −0.11]; P<0.001),
and 1-year cumulative savings were obtained for the PW-PAD group (mean dierence, −€4,286
[−4,711 to −4,020]). Twelve months aer PW-PAD implementation, we observed improved
health utilities (dierence, 0.040 [0.003–0.077]; P=0.032). ree deaths occurred. QALYs
were higher in the PW-PAD group (dierence, 0.045 [0.040–0.046]). In all sensitivity analyses,
taking into account sampling uncertainty and unit variable variation, PW-PAD was likely to
remain a cost-eective use of resources.
Conclusion. PW-PAD was strictly dominant, that is, less expensive and more eective com-
pared with usual care for people living with mental illness.
Introduction
Involuntary treatment and care is common and increasing in high-income countries, with vari-
ation by country (Wasserman et al., 2020). Compulsory hospital admissions, whether or not
associated with other coercive measures, are important causes of trauma and negative treat-
ment outcomes among people with mental disorders. Several studies show that these episodes
of deprivation of liberty constitute a very negative experience (Nyttingnes et al., 2016; Sibitz
et al., 2011), aecting quality of life (Swanson et al., 2003), with little evidence of eectiveness
in terms of health status, social functioning and use of services (Hofmann et al., 2022). e
overall cost of compulsory hospitalizations has scarcely been studied (Venturini et al., 2017).
Dierent models of interventions to reduce compulsory admissions have been developed. Of
these, psychiatric advance directives (PADs) or joint crisis plans (JCPs) are written documents
that allow adults with mental illness and with temporary decision-making incapacity to state
their will and preferences in advance, to be applied if further mental health crises impair their
decision-making capacity (Henderson et al., 2008).
Behind this common objective, PADs and JCPs dier in several ways, including their legal
force or with whom they are fullled (Atkinson et al., 2003). On the one hand, the speci-
city of PADs is that they emphasize that they are legal documents, on the other hand, JCPs
rather emphasize that they are on an agreement signed by the person, the healthcare pro-
fessionals and possibly the relatives. Authors have highlighted profound dierences in the
way people’s autonomy is represented in JCPs and PADs: more absolute for PADs advo-
cates and more assisted for JCPs advocates (Ambrosini and Crocker, 2010). Despite these
dierences, PADs and JCPs have so much in common that they are routinely categorized as a
https://doi.org/10.1017/S2045796023000197 Published online by Cambridge University Press
2 S. Loubière et al.
similar intervention and analysed together in systematic reviews,
which all rank them among the most eective interventions for
reducing compulsory admission (Barbui et al., 2021; Bone et al.,
2019; DeJong et al., 2016). A meta-analysis evaluated that PADs
and JCPs reduced compulsory admission by 25% (Molyneaux et al.,
2019).
Inherently, PADs promote user’s involvement and dialogue
(Murray and Wortzel, 2019). e benets in terms of autonomy
of PADs were at rst only theoretical but have gradually gained
ground as studies have shown that PADs improve user involve-
ment, empowerment and recovery; the therapeutic alliance and
integration of care, but these positive results are still low in evi-
dence (Nicaise et al., 2013). Recently, a randomized controlled trial
that we conducted showed that PADs facilitated by peer workers
were associated with fewer symptoms (eect size [95% condence
interval, CI]: −0.20 [−0.40 to 0.00]), higher empowerment (0.30
[0.10 to 0.50]) and higher recovery (0.44 [0.24 to 0.65]) compared
to the control group (Tinland et al., 2022).
e model of PADs that we have experimented in this trial had
the originality of being facilitated by peer workers, that is, peo-
ple with personal experiences of mental distress who are trained
to assist users in lling their PAD statement and sharing it with
relatives and psychiatrists. We observed that this model of PADs
was associated with a signicant reduction of over 32% in the pro-
portion of compulsory hospitalized people (main criterion of our
trial) (Tinland et al., 2022). As in the Molyneaux’ meta-analysis,
we found a less clear eect on voluntary hospitalizations and on
the total number of hospital admissions. e latter result raises
the question of the cost-eectiveness of PADs facilitated by peer
workers.
To our knowledge, and despite its societal importance, PAD
interventions have received little attention in terms of costs, and
only two of the randomized trials on the topic have explored
this parameter, exclusively in the UK (Barrett et al., 2013; Flood
et al., 2006), and the dierence in costs was not signicant. Despite
counting its eectiveness among people with mental illness, it is
unclear whether reducing compulsory admissions results in cost
savings or a shi in care to non-compulsory admissions amount-
ing to the same length of stay. e main objective of this study
was to conduct an economic evaluation of psychiatric advance
directives facilitated by peer workers (PW-PAD) as part of a lon-
gitudinal randomized clinical trial for people with severe mental
illness.
Methods
Trial design
e ‘Peer-Worker-Facilitated Psychiatric Advance Directive’ study
(DAiP) was a multicentre randomized controlled trial conducted
in seven mental health facilities of three cities in France (Lyon,
Paris and Marseille) between January 2019 and June 2021. Forty
treating psychiatrists checked the eligibility criteria and referred
eligible participants to research assistants, who validated inclu-
sion criteria, obtained written consent and conducted interviews at
inclusion, 6 and 12 months. According to the principle of ‘sector’,
which has organized most public psychiatric care in France since
the 1960s, the participating psychiatrists were both inpatient and
outpatient (the same team ensures continuity). Only six of them
were strictly outpatient. All psychiatrists at each participating men-
tal health facility were fully informed about the study at the time
of its implementation. No special training or incentives were given
to participating psychiatrists. e follow-up period was 12 months
aer an 18-month recruitment.
e study was registered on Clinicaltrials (NCT03630822).
e study was conducted in compliance with the Declaration of
Helsinki, sixth revision; Good Clinical Practice guidelines and
local regulatory requirements. e trial was approved by the local
ethics committee (2018-A00146-49).
Population
e inclusion criteria were being over 18 years of age; with
experience of involuntary admission to hospital within the past
12 months; with a diagnosis of schizophrenia, bipolar I dis-
order or schizoaective disorders according to Diagnostic and
Statistical Manual of Mental Disorders, h edition criteria; with
decision-making capacity assessed by a psychiatrist according to
the MacArthur Competence AssessmentTool for Clinical Research
(Appelbaum and Grisso, 1995) and with French government health
insurance. e exclusion criteria included being considered unable
to provide informed consent and being under the highest level of
guardianship. At the time of inclusion, most participants were dis-
charged from the hospital; nevertheless, a few participants were
included as inpatients.
Randomization
Participants were randomly assigned using a web-based system at
a 1:1 ratio. Treating clinicians, participants and research assistants
were aware of the assigned randomization group.
Intervention group: PW-PAD
All PW-PAD participants received the PW-PAD document from
research assistants, consisting of future treatment and support
preference options, description of early signs of relapse and cop-
ing strategies (English translation in the Supplementary material).
Depending on their wishes, the PW-PAD participants could (i)
meet a peer worker in a place of their choice; (ii) be supported
by this peer worker in draing the PAD document with as many
meetings as necessary and (iii) be supported by the peer worker
during the sharing of PAD with the healthcare agent and the psy-
chiatrist. Peer support workers were recruited specically for this
study and trained at a Recovery college (Centre de Formation
au Rétablissement – CoFoR). Regular exchanges were organized
between them (intervisions) and with the whole research team,
both remote and face-to-face.
Hard PADs were stored by the health worker or psychiatrist
depending on the participant’s choice and in electronic format if
available and requested. In case of crisis, reporting of the exis-
tence of PADs was done by patients, their entourage or informed
caregivers.
Control group
People assigned to the control group were treated as usual.
Depending on the person’s needs, the usual treatment includes
psychological, pharmaceutical and social support. In France, espe-
cially at the time of this study, JCPs were only used locally by a
few pioneering teams, and there was very little chance that peo-
ple in the control group would access to crisis plans. Nonetheless,
they received comprehensive information about the PAD con-
cept during the inclusion step and were free to ll out a PAD,
https://doi.org/10.1017/S2045796023000197 Published online by Cambridge University Press
Epidemiology and Psychiatric Sciences 3
but with no connection to the study’s peer worker. Figure S1 (see
Supplementary material) reports the number of PAD (PW-PAD or
other) completion and PAD use in each group.
Cost-eectiveness analysis
We performed a cost-utility analysis based on the societal per-
spective, including hospital, outpatient and community care, and
productivity losses due to illness. Incremental cost-eectiveness
ratios (ICERs) were expressed in terms of costs per quality-
adjusted life-year (QALY) gained, in accordance with Consolidated
Health Economic Evaluation Reporting Standards (CHEERS) and
the French National Authority for Health (HAS) guidelines for
economic evaluation (HAS, 2020; Husereau et al., 2022).
Eectiveness measure
Utilities for health states were assessed using the EuroQol-
5 Dimensions, three-level version (EQ-5D-3L) (Brooks, 1996;
Chevalier and de Pouvourville, 2013). is questionnaire is a
validated questionnaire that assesses a participant’s health status
through ve dimensions: mobility, personal care, routine occu-
pations, pain and discomfort and anxiety and depression. Each
dimension has three levels: no problems, some problems and severe
problems. e index score ranges from 0 (worst utility) to 1 (best
utility). Quality-of-life measures at baseline, 6 and 12 months
were summed as QALYs using an area under the curve approach
(Hunter et al., 2015; Husereau et al., 2022) and compared between
the PW-PAD and control groups.
Costs measure
We considered all direct and indirect healthcare costs in rela-
tion with care management during the follow-up. Resource use
data were retrieved from two sources: hospital-based administra-
tive databases and patients’ self-reported measures. Resource use
for each collaborative hospital was retrieved for all randomized
patients. To measure ambulatory and community care, and at the
margin to supplement hospital registry data, interview grids were
built to assist participants in reporting individual and prospective
resource use, based on previous studies of service use among peo-
ple with mental illness (Latimer et al., 2017; Loubière et al., 2020).
Resource utilization included those relating to the intervention,
including training of peer workers and time spent for PAD sup-
port, visits to the emergency department (ED), psychiatric hospital
admissions and total number of days at hospital, as well as outpa-
tient care. ese latter costs were assessed through consultations
with general practitioners, referring psychiatrists and other spe-
cialists. Indirect costs were investigated based on the number of
days of work absenteeism and compared between groups where
relevant.
Unit costs for hospital resources were estimated using data from
the French National Hospital Database (https://www.atih.sante.
fr/). e training of peer workers consisted in two half-day ses-
sions per week for 6 weeks, with an estimated cost of €1,215. We
observed the real cost of recruiting a peer worker at each site, either
a full-time or part-time contract over the duration of the study. e
average salary scales at the participating facilities were used to esti-
mate the monthly salary of the peer workers. No overhead costs
were charged to the intervention: the oce room and equipment
(i.e., computer and telephone) were already present and shared
with the care/administrative team. In the city of Paris (i.e., for
two health facilities), transportation costs were taken into account
(intra-city transportation card). For outpatient and community
care costs, national taris were used (Source: National databases
for medical and paramedical acts). All resources were valued in
2019 euros (see Supplementary Table S1) and discounting was not
applied.
Statistical analyses
An intention-to-treat analysis was conducted in the present study
and included all participants randomized in each group whether
or not they received the intervention or were lost to follow-up. To
detect a reduction of 30% in the rate of compulsory admissions to
psychiatric hospitals during the 12-month follow-up, the planned
sample was 200 per group, i.e., 400 in total. is number of sub-
jects allowed for a minimum cost dierence of €320 with a standard
deviation of 1,000, at a statistical power of 90%.
Missing data were addressed using multiple imputations
(van Buuren, 2007), under the assumption of missing at random
(Ware et al., 2012). Markov chain Monte Carlo multiple imputation
was used, which creates multiple ‘complete’ datasets by predictions
for each missing value. Fiy imputed datasets were implemented
using Multivariate Imputation by Chained Equations and mitools
R packages. We ran a sensitivity analysis for only the observed data
(excluding missing data).
A mixed model for longitudinal utility values was used to con-
trol for potential bias due to intra-patient correlated data and
the existence of co-variates inuencing quality of life. Between-
group dierences in service use and total costs were estimated
using generalized estimating equations (GENLIN function) apply-
ing a Poisson distribution with a link log due to skewness. Mean
dierences and beta coecients with 95% CI were provided.
We used non-parametric bootstrapping (with 5,000
replications) to resample observations. e bootstrap results
were combined to calculate the mean values for costs and utilities,
and the SEs for the imputed values were used to calculate 95%
CIs. e incremental cost per QALY gained was then calculated
as mean incremental costs divided by mean incremental QALYs
and reported where relevant. Uncertainty in the cost-eectiveness
results was analysed using both univariate deterministic and
probabilistic sensitivity analyses.
Statistical analyses were performed using R Studio (R version
4.0.2, RStudio, Inc., Massachusetts, USA) and TreeAgePro 2019.
Results
A total of 394 participants were included in the analysis: 196 were
assigned to the intervention group and 198 to the control group.
Interviews at the 12-month follow-up were completed for 127
(65%) in the PW-PAD group (four participants withdrew) and 139
(70%) in the control group (including ve withdrawals) (see details
in Supplementary Figure S2). Baseline characteristics and primary
clinical outcomes are reported in Table 1. Gender, age, comor-
bidities and experience of previous hospital admissions were well
balanced between the two groups. PW-PAD participants showed
signicantly improved self-reported mental symptoms, recovery
and empowerment scores over the 12-month follow-up compared
with the control group.
Eectiveness
e mean utility at baseline (standard error [SE]) for the PW-PAD
and control groups were, respectively, 0.798 (0.015) and 0.757
https://doi.org/10.1017/S2045796023000197 Published online by Cambridge University Press
4 S. Loubière et al.
(0.020) (P=0.101) (Table 2). Participants in the PW-PAD group
reported higher health utilities at month 12 (0.814 vs. 0.755;
P=0.017). From baseline to 12 months of follow-up, health util-
ities improved more in the PW-PAD group than in the control
group (mean dierence, 0.040; 95% CI, 0.003–0.077; P=0.032).
Based on complete data, the mean dierence in health utilities
Table 1. Socio-demographic and clinical characteristics of participants
(N=394)
Characteristics
PW-PAD goup
(n=196)
Control group
(n=198)
Gender, N(%)
Men 127 (64.8) 112 (56.6)
Age mean (SD), y 37.4 (11.7) 41.0 (12.7)
Median (IQR) 36 (28–44) 40 (31–49)
Nationality, N(%)
French 184 (93.9) 180 (91.8)
Education, N(%)
Less than high school
(<bac)
57 (29.2) 75 (37.9)
Completed or postsec-
ondary school
138 (70.8) 123 (62.1)
Marital status, N(%)
Single 132 (67.3) 128 (64.6)
Married/partnered 38 (19.4) 35 (17.7)
Divorced/separated/widow 26 (13.3) 35 (17.7)
Work activity, N(%)
Yes 33 (18.8) 37 (19.9)
EPICES score mean (SD) 40.6 (19.9) 42.8(20.9)
Median (IQR) 40.8 (24–57) 44.6 (26–59)
DSM-5 diagnosis, N(%)
Bipolar I disorder 66 (33.8) 73 (36.9)
Schizophrenia 86 (44.1) 92 (46.5)
Schizoaective disorders 43 (22.1) 33 (16.7)
Alcohol dependence, N(%)
Yes 6 (3.4) 6 (3.5)
Substance dependence, N(%)
Yes 22 (12.6) 24 (13.6)
With at least one somatic
comorbidity, N(%)
Yes 120 (61.2) 137 (69.2)
CGI score mean (SD) 4.1 (1.2) 4.3 (1.1)
Median (IQR) 4.0 (3–5) 4.0 (4–5)
Number of admissions in the
previous year, mean (SD)
1.5 (0.9) 1.4 (0.8)
1 132 (67.3) 148 (75.5)
2 45 (23.0) 37 (18.9)
≥3 17 (8.7) 11 (5.6)
(Continued)
Table 1. (Continued.)
Characteristics
PW-PAD goup
(n=196)
Control group
(n=198)
Clinical outcomes
MCSI score, mean (SD) 11.49 (11.91) 13.87 (10.99)
ES score, mean (SD) 16.80 (26.32) 10.20 (16.04)
RAS score, mean (SD) 72.60 (14.13) 65.55 (13.92)
PW-PAD: peer worker–facilitated psychiatric advance directive; SD: standard deviation; IQR:
interquartile range; DSM-5: Diagnostic and Statistical Manual of Mental Disorders, fih
edition; CGI: Clinical Global Impression scale.
Recovery was assessed using the Recovery Assessment Scale (RAS) (Corrigan et al., 2004),
which measures various aspects of recovery from the perspective of the consumer, with
a particular emphasis on hope and self-determination. This self-administered instrument
comprises 24 items, exploring five domains: personal confidence and hope, willingness
to ask for help, goal and success orientation, reliance on others and no domination by
symptoms. A higher score indicates better recovery.
Mental health symptomatology was assessed using the self-report modified Colorado
Symptom Index (MCSI; Conrad et al., 2001). The MCSI contains 14 items, which evaluate
how oen in the past month an individual has experienced a variety of mental health
symptoms, including loneliness, depression, anxiety and paranoia. Higher scores indicate a
greater likelihood of mental health problems.
Empowerment was assessed using the Empowerment Scale (ES) (Rogers et al., 1997). The
ES comprises 28 items, split into five dimensions: community activism and autonomy, self-
esteem and eicacy, optimism and control over the future, righteous anger and power
and powerlessness. The index score is 0 to 100, where higher scores correspond to higher
empowerment.
was 0.045 (95% CI, 0.002–0.088; P=0.039) (see Supplementary
Table S2).
ree (0.76%) patients died during the 12-month follow-up:
two (1.01%) patients in the TAU group and one (0.51%) patient
in the PW-PAD group.
Service use and costs
Details of service resource use and mean costs are provided in
Table 3. Table S2 (see Supplementary material) reports total psy-
chiatric hospital admissions and length of stay in both compulsory
and voluntary settings. e PW-PAD group experienced less psy-
chiatric hospital days over 1 year compared to the control group
(45.4 vs. 57.1; P=0.026). No signicant dierences were found
in the mean number of ED visits or consultations (P=0.056 and
P=0.309; respectively). Similarly, the rates of patients having
working activity at the end of follow-up did not show signicant
dierence between the PW-PAD and control groups (36.6% vs.
28.2%; P=0.184). No signicant dierences were found in the
mean number of days o work between the PW-PAD and control
groups (3.75 vs. 2.54 days; P=0.383).
Compared with the control group, PW-PAD exhibited a sta-
tistically signicant cost dierence in total costs (€22,094.27 vs. €
26,382.39; P=0.001) (Table 4).
Cost-eectiveness
e incremental benet of PW-PAD versus control group was
0.045 QALY (95% CI, 0.040 to 0.046); the incremental cost
was −€4,286 (95% CI, −4,711 to −4,020) (Table 4). e cost-
eectiveness analysis showed that the PW-PAD intervention was
strictly dominant, that is, less expensive and more eective com-
pared to the usual care.
e bootstrap distribution of the ICER showed that 100% of the
5,000 replicates of ICER were located in the lower-right quadrant
of the scatterplot plan (Fig. 1).
https://doi.org/10.1017/S2045796023000197 Published online by Cambridge University Press
Epidemiology and Psychiatric Sciences 5
Table 2. Change in health utilities score (EQ5D-3L) during follow-up in the DAiP trial
Baseline Month 6 Month 12
PW-PAD group Control group PW-PAD group Control group PW-PAD group Control group Mean dierenceb(95% CI)
Mean utility 0.798 0.757 0.774 0.753 0.814 0.755 0.040 (0.003–0.077)
Standard error 0.015 0.020 0.019 0.021 0.017 0.018
Pvaluea0.101 0.459 0.017 0.032b
PW-PAD: peer worker–facilitated psychiatric advance directive; SE: standard error; 95% CI: 95% confidence interval.
aP-values at each time were provided by t-test analysis (independent samples test) based on imputed data.
bMixed linear models (MIXED) for repeated-measure analyses were applied, using a restricted maximum likelihood approach for variance estimation, with a repeated variable ln(t+1),
where tis the time from baseline. An unstructured covariance matrix for repeated measures was used. The interaction between group and time was tested and was not kept because none
achieved statistical significance.
Values in italic indicate a statistically significant dierence from the group variable (PW-PAD vs. control groups)
Table 3. Mean healthcare utilization and costs at 12 months of follow-up
Utilization Costs
PW-PAD
group
(n=196)
Control
group
(n=198)
PW-PAD
group
(n=196)
Control
group
(n=198)
Outcomes Mean (SE) Mean (SE)
Mean dierence
(95% CI) Mean (SE) Mean (SE)
Relative risk
(95% CI)a
No. of days
in psychiatric
hospitalization
45.41 (3.28) 57.08 (4.09) −11.67 (−21.95
to –1.39)
Total
inpatient
costs
20,127.25 (867.27) 25,062.54 (800.41) −0.22 (−0.33
to –0.11)
No. of emergency
department visits
1.61 (0.36) 0.90 (0.23) 0.71 (−0.02 to 1.44) Emergency
department
costs
278.75 (58.67) 158.21 (36.32) 0.59 (0.11
to 1.08)
No. of ambulatory
visits
13.79 (1.05) 12.35 (0.93) 1.44 (−1.33 to 4.20) Ambulatory
costs
1,290.95 (28.48) 1,168.66 (19.29) 0.10 (0.05
to 0.16)
Intervention
costs
399.66 (3.63) – –
Total costs 22,094.27 (873.36) 26,382.39 (836.15) −0.18 (−0.28
to –0.08)
PW-PAD: peer worker–facilitated psychiatric advance directive; SE: standard error; 95% CI: 95% confidence interval.
aA Poisson distribution with a link log was used.
Generalized linear models were used to address mean dierence and 95% CI for the groups.
Values in italic indicate a statistically significant dierence in pooled imputation dataset from the group variable (PW-PAD vs. control groups), P<0.05.
Table 4. Mean and incremental costs and QALYs for patients receiving PW-PAD
versus usual care
PW-PAD group,
mean (95% CI)
Control group,
mean (95% CI)
Incremental,
mean (95% CI)
QALYs 0.789 (0.783–0.794) 0.746 (0.740–0.752) 0.045 (0.040–0.046)
Costs 22,095
(21,498–24,436)
26,382
(25,695–27,981)
−4,286 (−4,711
to –4,020)
ICER −99,977 (−113,378
to −91,035)
PW-PAD: peer worker–facilitated psychiatric advance directive; QALY: quality-adjusted life-
years; CI: confidence interval.
e graph contains axes that represent the incremental cost
(y-axis) and incremental eectiveness (x-axis). Each point in the
graph represents the incremental cost and incremental eective-
ness values (PW-PAD vs. control) from a single recalculation from
the database.
e willingness to pay (WTP), or ICER threshold, is used
as the slope of a line intersecting the origin of the plot. e
WTP line in the graph intersects points having the specied
ICER value, and the region below and to the right of the
line includes points where the intervention/PW-PAD is more
cost-eective than the usual care/control. e ellipsis shows the
95% CI.
Sensitivity analyses
Sensitivity analyses with complete cases yielded similar results to
imputed datasets (Table S3). e tornado diagram (Fig. 2) indi-
cated that cost-eectiveness was most strongly aected by utility
values altered over their 95% CI and inpatient costs varying from
±30%. is was followed by the mortality rates in the PW-PAD
and control groups. Regardless of the change in these parame-
ters, the higher eectiveness and cost savings were maintained for
PW-PAD. In addition, increasing or decreasing the costs of the
intervention by 50% had less eects on the ICER.
Taking sampling uncertainty into consideration, the cost-
eectiveness acceptability curve for the base-case analysis
(Figure S3) shows a 100% probability that PW-PAD was cost-
eective by comparison with usual care at the threshold of €1,000
WTP per QALY gained.
Discussion
Our study is the largest prospective study on PADs including peo-
ple living with mental illness who were compulsorily admitted
https://doi.org/10.1017/S2045796023000197 Published online by Cambridge University Press
6 S. Loubière et al.
Fig. 1. Cost-eectiveness plane associated with the 12-month cost per QALY analysis.
PW-PAD: peer worker–facilitated psychiatric advance directive; QALY: quality-adjusted life year; WTP: willingness to pay.
Fig. 2. One-way sensitivity analysis of cost-eectiveness associated with PW-PAD intervention.
PW-PAD: peer worker–facilitated psychiatric advance directive; QALY: quality-adjusted life year; CI: confidence interval; ICER: incremental cost-eectiveness ratio.
during the past year. It is also the rst to evaluate the costs and
benets of peer worker–facilitated PADs for the management of a
patient’s mental disorder. In our study, for the overall population,
PW-PAD was strictly dominant. e intervention was associated
with a higher number of QALYs at a lower cost compared to usual
care.
PW-PAD was associated with a signicant decrease in total
psychiatric inpatient days/nights and with a signicant improve-
ment in compulsory admissions, with no signicant reduction in
the rate of overall psychiatric admissions (Tinland et al., 2022).
Our hypothesis is that, rather than preventing psychiatric hospital
admissions, PADs may reduce compulsory admissions by making
participants more willing to consider voluntary admission when a
crisis occurs. e economic analysis goes further and shows, based
on a detailed measure of health service utilization and cost analysis,
that the use of PW-PAD reduces the overall length of stay and costs
compared to usual care. Our ndings are consistent with nonran-
domized studies showing improvement in length of stay and costs
associated with a shi from coercive measures (Dimitri et al., 2018,
Kallert et al., 2008; McLaughlin et al., 2016, Salias and Fenton,
2000).
Few multicentre and randomized controlled studies have eval-
uated the eectiveness of advance directives in psychiatry, none
incorporating facilitation by peer workers (Henderson et al., 2004;
Molyneaux et al., 2019; Papageorgiou et al., 2002; Ruchlewska et al.,
2014; Swanson et al., 2006; ornicro et al., 2013). Of these,
only two had assessed their economic outcomes (Barrett et al.,
2013; Flood et al., 2006). Flood et al. could not nd any dier-
ences in admissions or total costs per participant over a 15-month
follow-up. Similarly, Barrett et al. could not nd any dierences
https://doi.org/10.1017/S2045796023000197 Published online by Cambridge University Press
Epidemiology and Psychiatric Sciences 7
in compulsory admissions or total societal costs per participant
over an 18-month follow-up. Neither of these studies found signi-
cant intervention eects on inpatient stays (i.e., number of nights).
Our results show that the French participants spent considerable
amount of time in hospital, regardless of the group, compared
with those from UK studies with the same inclusion criteria. As
an example, over 1 year, participants in both French study groups
spent, on average, 2.5 times more days in hospital than partici-
pants in Barret’s UK-b ased study. e initial lengthy hospitalization
duration, combined with its sharp fall over 1 year, explains the
majority of our observed cost savings. is is consistent with the
conclusion of Burns’ meta-analysis, which showed that interven-
tions that reduce hospital admissions (in his paper. it is case man-
agement) are more eective when participants are high consumers
of hospital care at baseline (Burns et al., 2007). Dierences between
the interventions could also explain these results: PW-PAD dier
from Flood and Barret’s intervention in that they are PADs, and
peer workers facilitate them. On the one hand, PADs are unique
tools to promote self-determination (Elbogen et al., 2007), and part
of the results could be due to this law-oriented form; on the other
hand, peer workers’ involvement in healthcare has been shown to
be associated with improvements in quality of life, self-ecacy,
hope and empowerment (e.g., Fuhr et al., 2014; Lloyd-Evans et al.,
2014; Mahlke et al., 2017; Vayshenker et al., 2016), which is con-
sistent with our results. Unfortunately, our study does not allow
us to identify the relative role of each ingredient in achieving good
clinical or economic outcomes. PW-PAD is a complex intervention
where the meeting with a peer worker and crisis reection likely
interact with each other.
Over the follow-up period, participants in the PW-PAD group
gained more health utilities than their counterparts in the con-
trol group, with similar EQ5D-3L scores at baseline and 6 months.
is nding could suggest a learning period with PW-PAD, with
the eects in terms of quality of life appearing only aer the rst
6 months following the initiation of the advance directives, sug-
gesting that this could be maintained in the long term. Further
investigations are needed to understand the factors that inu-
ence the early benets of the intervention and to help inform
decision-making.
Even when the PW-PAD intervention was considered to cost
double, the intervention still remained dominant compared to
usual care. In fact, regardless of the variations considered in cost
parameters, the PW-PAD group remained dominant, presenting a
100% chance of being cost-eective at small WTP thresholds for
mental health programmes.
Such results are important because they inform decision mak-
ers and, perhaps more critically, they contribute to international
guidelines on the economic eciency of programmes in men-
tal health. Indeed, cost analyses oer a perspective for policy
changes. e PW-PAD is an intervention, which can be imple-
mented in a fairly straightforward manner and can be very quickly
cost-eective for the healthcare system.
Strengths and limitations
e strengths of our study lie in our design and methodol-
ogy. First, this research was deployed through seven psychiatric
facilities, each with several services. e wide range of practices
across these services reects the diversity of practice in France.
Between and within countries and regions, large variations have
been found between services in rates of inpatient admissions and
compulsion (Gandré et al., 2018b,2018a; Hofstad et al., 2021;
Weich et al., 2017), and these variations are not yet fully under-
stood by the scientic literature (Rugkåsa, 2017). e diversity
of participating centres, associated with broad inclusion criteria
and limited exclusion criteria, enhances our condence in the
generalizability of our results. Second, we captured a measure of
benet expressed in terms of participant’s quality of life rather
than simply a measure of resource use. Our approach is in accor-
dance with international guidelines and avoids the double counting
of resource use in both the denominator and numerator of the
ICER (Neumann et al., 2016). ird, resource use was based on
the data collected from hospital-based databases that captured the
entire hospital pathway associated with the management of mental
illness.
is trial had several limitations. First among them is the short
follow-up period. Given the short-term nature of PAD fullment,
the uncertainty surrounding long-term use and long-term conse-
quences on health utility of PADs and the limited impact of PADs
on survival, the 12-month randomized controlled trial timeline
was deemed appropriate to capture most of the relevant costs and
benets associated with the intervention. Most of the incremental
cost of PW-PAD is spent in the rst 3 months, so this intervention
might become even more cost-eective aer 1 year if participants
receiving the intervention continue to improve more than those
receiving usual care without additional costs.
Second, we assumed the same costs for compulsory and stan-
dard hospitalizations. We may have underestimated the costs in
the control group, although this reinforces the conservative costing
approach. e accounting costs provided by French health agen-
cies for hospitalization are not dierentiated between compulsory
or freely admitted patients, whereas compulsory hospitalization
appears more expensive, especially at the beginning, oen requir-
ing more sta and special facilities such as isolation rooms (Flood
et al., 2008; LeBel and Goldstein 2005). With 32% more compul-
sory admissions in the control group, total costs would be even
higher if the associated containment and seclusion costs had been
valued. At the same time, we did not account for the impact of dis-
charge to the community on people other than the study patient
(e.g., family or partners providing care to the patient) or measure
costs of pharmaceuticals, the latter being assumed to be higher at
entry in compulsory admission (Brown et al., 2010).
Finally, discontinuation rates were around one-third in both
groups. Unfortunately, the COVID-19 pandemic generated a crisis
that suddenly widened the gap in access to healthcare, especially
for vulnerable populations. Our attrition rates were mainly due to
lost to follow-up in periods of lockdown.
Conclusions
Among people living with mental illness, the elaboration of PADs
with the support of a peer worker was associated with a signi-
cant improvement in health utility in parallel to cost-savings over a
12-month period. ese ndings support a national-scale promul-
gation of this type of intervention and an implementation study of
these tools to measure the level of adoption in common practice
and to identify barriers and facilitators.
Supplementary material. e supplementary material for this article can
be found at https://doi.org/10.1017/S2045796023000197.
Data availability statement. Not all the data are freely accessible because
no informed consent was given by the participants for open data sharing, but
we can provide the data used in this study to researchers who want to use them,
following approval by the ethics committee of the Aix Marseille University.
https://doi.org/10.1017/S2045796023000197 Published online by Cambridge University Press
8 S. Loubière et al.
Acknowledgements. Jean Dhorne provided help to monitor this clinical
trial, with the assistance of Marika Larose. Sophie Tardosky and Richard
Malkoun contributed to the data management of the study. Magali Pontier
had the idea to develop DAP in Marseille and provided preliminary results
about the limits of facilitation by professionals. Nicolas Ordener, Julien Grard,
Celine Letailleur, EUTOPIA, MARSS team and the CoFoR Recovery College
provided help and advices to develop the practice of facilitation by peer work-
ers. Lee Antoine, Camille Niard, Nicholas Armstrong, Elsa Castot and Iannis
Mc Cluskey contributed as peer workers. Frederic Mougeot, Emmanuelle
Jouet, Magali Pontier, Nicolas Franck, Juliette Robert, Antoine Simon, Aurélien
Troisoeufs, Julien Grard and Karine Baumstarck were member of the steering
committee. Owen Taylor provided English proofreading. We thank them for
their grateful help.
Financial support. is work was supported by an institutional grant from
the French 2017 National Program of Health Services Research (Programme de
Recherche sur la Performance du système de Soins, PREPS-17-0575, FINESS
number 130786049).
Competing interest. None.
Ethical standards. e authors assert that all procedures contributing to this
work complywith the et hical standardsof the relevant national and institutional
committees on human experimentation and with the Helsinki Declaration of
1975, as revised in 2000.
References
Ambrosini D and Crocker A (2010) Les directives psychiatriques anticipées
(DPA) et le rôle de l’autonomie. Santé Mentale au Québec 34(2), 51–74.
Appelbaum PS and Grisso T (1995) e MacArthur treatment competence
study. I: Mental illness and competence to consent to treatment. Law and
Human Behavior 19(2), 105–126.
Atkinson JM, Garner HC, Patrick H and Stuart S (2003) Issues in the devel-
opment of advance directives in mental health care. Journal of Mental Health
12(5), 463–474.
Barbui C, Purgato M, Abdulmalik J, Caldas-de-almeida JM, Eaton J,
Gureje O, Hanlon C, Nosè M, Ostuzzi G, Saraceno B, Saxena S, Tedesch F
and ornicro G (2021) Ecacy of interventions to reduce coercive treat-
ment in mental health services: Umbrella review of randomised evidence.
British Journal of Psychiatry 218(4), 185–195.
Barrett B, Waheed W, Farrelly S, Birchwood M, Dunn G, Flach C,
Henderson C, Leese M, Lester H, Marshall M and Rose D (2013)
Randomised controlled trial of joint crisis plans to reduce compulsory
treatment for people with psychosis: Economic outcomes. PLoS One 8(11),
e74210.
Bone JK, McCloud T,Scott HR, Machin K, Markham S, Persaud K, Johnson S
and Lloyd-Evans B (2019) Psychosocial interventions to reduce compul-
sory psychiatric admissions: A rapid evidence synthesis. eClinialMedicine10,
58–67.
Brooks R (1996) EuroQol: e current state of play. Health Policy 37(1), 53–72.
Brown S, Chhina N and Dye S (2010) Use of psychotropic medication in seven
English psychiatric intensive care units. e Psychiatrist 34(4), 130–135.
Burns T, Catty J, Dash M, Roberts C, Lockwood A and Marshall M (2007)
Use of intensive case management to reduce time in hospital in people
with severe mental illness: Systematic review and meta-regression. BMJ
335(7615), 336.
Chevalier J and de Pouvourville G (2013) Valuing EQ-5D using time trade-o
in France. e European Journal of Health Economics 14(1), 57–66.
Conrad KJ, Yagelka JR, Matters MD, Rich AR, Williams V and Buchanan M
(2001) Reliability and validity of a modied Colorado Symptom Index in a
national homeless sample. Mental Health Services Research 3(3), 141–153.
Corrigan PW, Salzer M, Ralph RO, Sangster Y and Keck L (2004) Examining
the factor structure of the recovery assessment scale. Schizophrenia Bulletin
30(4), 1035–1041.
DeJong MH, Kamperman AM and Mulder CL (2016) Open wards in psychi-
atric clinics and compulsory psychiatric admissions-reply. JAMA Psychiatry
73(12), 1293–1294.
Dimitri G, Giacco D, Bauer M, Bird VJ, Greenberg L, Lasalvia A, Lorant V,
Moskalewicz J, Nicaise P, Pfennig A, Ruggieri M, Marta W and Priebe S
(2018) Predictors of length of stay in psychiatric inpatient units: Does their
eect vary across countries? European Psychiatry 48(1), 6–12.
Elbogen EB, VanDorn R, Swanson JW, Swartz MS, Ferron J,Wagner RH, and
Wilder C (2007) Eectively implementing psychiatric advance directives to
promote self-determination of treatment among people with mental illness.
Psychology, Public Policy and Law 13, 1–15.
Flood C, Bowers L and Parkin D (2008) Estimating the costs of conict and
containment on adult acute inpatient psychiatric wards. Nursing Economics
26(5), 325–330, 324.
Flood C, Byford S, Henderson C, Leese M, ornicro G, Sutherby K
and Szmukler G (2006) Joint crisis plans for people with psy-
chosis: Economic evaluation of a randomised controlled trial. BMJ
333(7571), 729.
Fuhr DC, Salisbury TT, De Silva MJ, AtifN, van Ginneken N, Rahman A and
Patel V (2014) Eectiveness of peer-delivered interventions for severe men-
tal illness and depression on clinical and psychosocial outcomes: A system-
atic review and meta-analysis. Social Psychiatry and Psychiatric Epidemiology
49(11), 1691–1702.
Gandré C, Gervaix J, illard J, Macé J-M, Roelandt J-L and Chevreul K
(2018a) Geographic variations in involuntary care and associations with the
supply of health and social care: Results from a nationwide study. BMC
Health Services Research 18(1), 253.
Gandré C, Gervaix J, illard J, Macé J-M, Roelandt J-L and Chevreul K
(2018b) Understanding geographic variations in psychiatric inpatient
admission rates: Width of the variations and associations with the supply
of health and social care in France. BMC Psychiatry 18(1), 174.
HAS (2020) Choices in methods for economic evaluation. Haute Autorité de
Santé. https://www.has-sante.fr/jcms/r_1499251/en/choices-in-methods-
for-economic-evaluation (accessed 20 June 2022).
Henderson C, Flood C, Leese M, ornicro G, Sutherby K and Szmukler G
(2004) Eect of joint crisis plans on use of compulsory treatment in psychi-
atry: Single blind randomised controlled trial. BMJ 329(7458), 136.
Henderson C, Swanson JW, Szmuk ler G, ornicroG and Zin kler M (2008)
A typology of advance statements in mental health care. Psychiatric Services
59(1), 63–71.
Hofmann AB, Schmid HM, Hofmann LA, Noboa V, Seifritz E, Vetter S
and Egger ST (2022) Impact of compulsory admission on treatment
and outcome: A propensity score matched analysis. European Psychiatry
65(1), e6.
Hofstad T, Rugkåsa J, Osborg Ose S, Nyttingnes O, Helene Høymork Kjus S
and Lossius Husum T (2021) Service characteristics and geographical
variation in compulsory hospitalisation: An exploratory random eects
within–between analysis of Norwegian municipalities, 2015–2018. Frontiers
in Psychiatry 12(December), 737698.
Hunter RM, Baio G, Butt T, Morris S, Round J and Freemantle N (2015) An
educational review of the statistical issues in analysing utility data for cost-
utility analysis. PharmacoEconomics 33(4), 355–366.
Husereau D, Drummond M, Augustovski F, de Bekker-grob E, Briggs AH,
Carswell C, Caulley L, Chaiyakunapruk N, Greenberg D, Loder E,
Josephine Mauskopf CDM, Petrou S, Pwu R-F, Staniszewska S and
CHEERS 2022 ISPOR Good Research Practices Task Force (2022)
Consolidated Health Economic Evaluation Reporting Standards 2022
(CHEERS 2022) statement: Updated reporting guidance for health economic
evaluations. Value in Health 25(1), 3–9.
Kallert TW, Gl
ockner M and Schützwohl M (2008) Involuntary vs. volun-
tary hospital admission: A systematic literature review on outcome diver-
sity. European Archives of Psychiatry and Clinical Neuroscience 258(4),
195–209.
Latimer EA, Rabouin D, Cao Z, Angela L, Powell G, Aubry T, Distasio J,
Hwang SW, Somers JM, Stergiopoulos V, Veldhuizen S, Moodie EEM,
Lesage A, Goering PN and At Home/Chez Soi Investigators (2017) Costs
of services for homeless people with mental illness in 5 Canadian cities: A
large prospective follow-up study. CMAJ Open 5(3), E576–E585.
LeBel J and Goldstein R (2005) Special section on seclusion and restraint: e
economic cost of using restraint and the value added by restraint reduction
or elimination. Psychiatric Services 56(9), 1109–1114.
https://doi.org/10.1017/S2045796023000197 Published online by Cambridge University Press
Epidemiology and Psychiatric Sciences 9
Lloyd-Evans B, Mayo-Wilson E, Harrison B, Istead H, Brown E, Pilling S,
Johnson S and Kendall T (2014) A systematic review and meta-analysis of
randomised controlled trials of peer support for people with severe mental
illness. BMC Psychiatry 14, 39.
Loubière S, Tinland A, Taylor O, Loundou A, Girard V, Boyer L and
Auquier P (2020) Determinants of healthcare use by homeless people with
schizophrenia or bipolar disorder: Results from the French housing rst
study. Public Health 185(Août), 224–231.
Mahlke CI, Priebe S, Heumann K, Daubmann A, Wegscheider K and Bock T
(2017) Eectiveness of one-to-one peer support for patients with severe
mental illness – A randomised controlled trial. European Psychiatry 42,
103–110.
McLaughlin P, Giacco D and Priebe S (2016) Use of coercive measures dur-
ing involuntary psychiatric admission and treatment outcomes: Data from a
prospective study across 10 European countries. PLoS One 11(12), e0168720.
Molyneaux E, Turner A, Candy B, Landau S, Johnson S and Lloyd-Evans B
(2019) Crisis-planning interventions for people with psychotic illness or
bipolar disorder: Systematic review and meta-analyses. BJPsych Open 5(4),
e53.
Murray H and Wortzel HS (2019) Psychiatric advance directives:Or igins,b en-
ets, challenges and futures directions. Journal of Psychiatric Practice 25(4),
303–307.
Neumann PJ, Sanders GD, Russell LB, Siegel JE, and eodore GG
(2016) Cost-Eectiveness in Health and Medicine, 2nd edn. Oxford: Oxford
University Press.
Nicaise P, Lorant V and Dubois V (2013) Psychiatric advance directives as a
complex and multistage intervention: A realist systematic review. Health &
Social Care in the Community 21(1), 1–14.
Nyttingnes O, Ruud T and Rugkåsa J (2016) ‘It’s unbelievably humiliating’ –
Patients’ expressions of negative eects of coercion in mental health care.
International Journal of Law and Psychiatry 49(November), 147–153.
Papageorgiou A, King M, Janmohamed A, Davidson O and Dawson J (2002)
Advance directives for patients compulsorily admitted to hospital with seri-
ous mental illness: Randomised controlled trial. British Journal of Psychiatry
181(6), 513–519.
Rogers E, Chamberlin J, Ellison ML and Crean T (1997) A consumer-
constructed scale to measure empowerment among users of mental health
services. Psychiatric Services 48(8), 1042–1047.
Ruchlewska A, Wierdsma AI, Kamperman AM, van der Gaag M,
Smulders R, Roosenschoon B-J and Mulder CL (2014) Eect of cri-
sis plans on admissions and emergency visits: A randomized controlled
trial. PLoS One 9(3), e91882.
Rugkåsa J (2017) Why we need to understand service variation in compulsion.
e Lancet Psychiatry 4(8), 577–578.
Sailas E, and Fenton M (2000) Seclusion and restraint for people with seri-
ous mental illnesses. e Cochrane Database of Systematic Reviews 2000(1),
CD001163.
Sibitz I, Scheutz A, Lakeman R, Schrank B, Schaer M and Amering M
(2011) Impact of coercive measures on life stories: Qualitative study. British
Journal of Psychiatry 199(3), 239–244.
Swanson JW, Swartz MS, Elbogen Eric B, Van Dorn RA, Joelle Ferron HRW,
McCauley BJ and Kim M (2006) Facilitated psychiatric advance directives:
A randomized trial of an intervention to foster advance treatment planning
among persons with severe mental illness. American Journal of Psychiatry
163(11), 1943–1951.
Swanson JW, Swartz MS, Elbogen EB, Wagner HR and Burns BJ (2003)
Eects of involuntary outpatient commitment on subjective quality of life
in persons with severe mental illness. Behavioral Sciences & the Law 21(4),
473–491.
ornicro G, Farrelly S, Szmukler G, Birchwood M, Waheed W, Flach C,
Barrett B, Byford S, HendersonC, Sutherby K, Lester H, Rose D, Dunn G,
Leese M and Marshall M (2013) Clinical outcomes of joint crisis plans
to reduce compulsory treatment for people with psychosis: A randomised
controlled trial. e Lancet 381(9878), 1634–1641.
Tinland A, Loubière S, Mougeot F, Jouet E, Pontier M, Baumstarck K,
Loundou A, Franck N, Lançon C, Auquier P and Group DAiP (2022) Eect
of psychiatric advance directives facilitated by peer workers on compulsory
admission among people with mental illness: A randomized clinical trial.
JAMA Psychiatry 79(8), 752.
van Buuren S (2007) Multiple imputation of discrete and continuous data by
fully conditional specication. Statistical Methods in Medical Research 16(3),
219–242.
Vayshenker B, Mulay AL, Gonzales L, West ML, Brown I and
Yanos PT (2016) Participation in peer support services and out-
comes related to recovery. Psychiatric Rehabilitation Journal 39(3),
274–281.
Venturini F, de Moura EC, Bastos PA, Martins LC and Fragoso YD
(2017) Prole and costs involved in long-term compulsory hospital-
ization of psychiatric patients. Revista Brasileira de Psiquiatria 40(3),
306–308.
Ware JH, Harrington D, Hunter DJ and D’Agostino RB (2012) Missing Data.
New England Journal of Medicine 367(14), 1353–1354.
Wasserman D, Apter G, Baeken C, Bailey S, Balazs J, Bec C, Bienkowski P,
Bobes J, Bravo Ortiz MF, Brunn H, Bôke
O, Camilleri N, Carpiniello B,
Chihai J, Chkonia E, Courtet P, Cozman D, David M, Dom G,
Esanu A, Falkai P, Flannery W, Gasparyan K, Gerlinger G, Gorwood P,
Gudmundsson O, Hanon C, Heinz A, Heitor Dos Santos MJ, Hedlund A,
Ismayilov F, Ismayilov N, Isometsä ET, Izakova L, Kleinberg A,
Kurimay T, Klæbo Reitan S, Lecic-Tosevski D, Lehmets A, Lindberg N,
Lundblad KA, Lynch G, Maddock C, Malt UF, Martin L, Martynikhin I,
Maruta NO, Matthys F, Mazaliauskiene R, Mihajlovic G, Mihaljevic
Peles A, Miklavic V, Mohr P, Munarriz Ferrandis M, Musalek M,
Neznanov N, Ostorharics-Horvath G, Pajevi
c I, Popova A, Pregelj P,
Prinsen E, Rados C, Roig A, Rojnic KM, Samochowiec J, Sartorius N,
Savenko Y, Skugarevsky O, Slodecki E, Soghoyan A, Stone DS, Taylor-
East R, Terauds E, Tsopelas C, Tudose C, Tyano S, Vallon P, Van der Gaag
RJ, Varandas P,Vavrusova L, Voloshyn P, Wancata J, WiseJ, Zemishlany Z,
Oncü F and VahipS (2020) Compulsory admissions of patients with mental
disorders: State of the art on ethical and legislative aspects in 40 European
Countries. European Psychiatry 63(1), e82.
Weich S, McBride O, Twigg L, Duncan C, Keown P, Crepaz-Keay D,
Cyhlarova E, Parsons H, Scott J and Bhui K (2017) Variationin compulsory
psychiatric inpatient admission in England: A cross-classied, multilevel
analysis. e Lancet Psychiatry 4(8), 619–626.
https://doi.org/10.1017/S2045796023000197 Published online by Cambridge University Press
